363
| ID | 363 |
|---|---|
| Original Title | The Contextual Process of Bystander Intervention in Bias-Motivated Violent Victimization: An Experimental Approach |
| Sanitized Title | thecontextualprocessofbystanderinterventioninbiasmotivatedviolentvictimizationanexperimentalapproach |
| Clean Title | The Contextual Process Of Bystander Intervention In Bias-Motivated Violent Victimization: An Experimental Approach |
| Source ID | 2 |
| Article Id01 | 617768683 |
| Article Id02 | oai:scholarsarchive.library.albany.edu:etd-1017 |
| Corpus ID | (not set) |
| Dup | (not set) |
| Dup ID | (not set) |
| Url | https://core.ac.uk/outputs/617768683 |
| Publication Url | (not set) |
| Download Url | https://core.ac.uk/download/617768683.pdf |
| Original Abstract | Hate crimes against Asian Americans and LGBT individuals have been on the rise within the U.S. in the past few years (UCR, 2024). Although these offenses tend to occur in public, numerous stories in the media highlight that few bystanders choose to intervene. To unveil why bystanders choose not to intervene, this dissertation explored the extent to which 1,001 U.S. adults engaged with the five-step situational model for bystander intervention proposed by Latané and Darley (1970). Respondents were recruited through Qualtrics and randomly assigned to a hate scenario involving one of six victims: an Asian American man, an Asian American woman, a gay man, a lesbian, a trans man, and a trans woman. The hate scenario escalates such that it begins with a microaggression that then leads to a hate incident before culminating in a hate crime. Study 1 examined whether there was a relationship between increased incident severity as the hate scenario progresses and increased respondent engagement with the situational model. Bystander heterogeneity was also explored through a latent class analysis of bystander progression through each phase of the hate scenario. The latent classes of bystanders then served as one of the dependent variables for the next two studies. Study 2 examined whether bystander progress through the situational model differed based upon the victim’s identity. Study three tested whether bystander characteristics such as empathy, bystander efficacy, and regard for others predict bystander latent class. Several key findings emerged from these three studies. First, bystander progress through the situational model indeed increases as the hate event escalates. This relationship was observed through examining whether the bystander saw a problem, whether the bystander intervened, and the number of steps the bystander proceeded through in the situational model. Second, three latent classes of bystanders were found in this sample: always interveners, those who progressed further through the model as the event escalated, and those who never intervened. Third, situational model progress significantly differed based on the identity of the victim. This most prominently occurred with subjects being significantly more likely to see a problem and intervene for the Asian woman victim than the sexual minority and gender minority victims. Lastly, the bystander’s favorability towards the victim’s group was consistently the strongest predictor of whether the bystander saw a problem or intervened followed by empathy, bystander behavior, and bystander efficacy. These results together will help to inform the applicability of the situational model to hate as well as the cognitive mechanisms that can be leveraged to enhance the probability of intervention |
| Clean Abstract | (not set) |
| Tags | (not set) |
| Original Full Text | University at Albany, State University of New York Scholars Archive Electronic Theses & Dissertations (2024 - present) The Graduate School Summer 2024 The Contextual Process of Bystander Intervention in Bias-Motivated Violent Victimization: An Experimental Approach James Hubbell University at Albany, State University of New York, jhubbell@albany.edu The University at Albany community has made this article openly available. Please share how this access benefits you. Follow this and additional works at: https://scholarsarchive.library.albany.edu/etd Part of the Criminology and Criminal Justice Commons Recommended Citation Hubbell, James, "The Contextual Process of Bystander Intervention in Bias-Motivated Violent Victimization: An Experimental Approach" (2024). Electronic Theses & Dissertations (2024 - present). 8. https://scholarsarchive.library.albany.edu/etd/8 This work is licensed under the University at Albany Standard Author Agreement. This Dissertation is brought to you for free and open access by the The Graduate School at Scholars Archive. It has been accepted for inclusion in Electronic Theses & Dissertations (2024 - present) by an authorized administrator of Scholars Archive. Please see Terms of Use. For more information, please contact scholarsarchive@albany.edu. The Contextual Process of Bystander Intervention in Bias-Motivated Violent Victimization: An Experimental Approach by James T. Hubbell A Dissertation Submitted to the University at Albany, State University of New York In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Nelson A. Rockefeller College of Public Affairs and Policy School of Criminal Justice Summer 2024 ii ABSTRACT Hate crimes against Asian Americans and LGBT individuals have been on the rise within the U.S. in the past few years (UCR, 2024). Although these offenses tend to occur in public, numerous stories in the media highlight that few bystanders choose to intervene. To unveil why bystanders choose not to intervene, this dissertation explored the extent to which 1,001 U.S. adults engaged with the five-step situational model for bystander intervention proposed by Latané and Darley (1970). Respondents were recruited through Qualtrics and randomly assigned to a hate scenario involving one of six victims: an Asian American man, an Asian American woman, a gay man, a lesbian, a trans man, and a trans woman. The hate scenario escalates such that it begins with a microaggression that then leads to a hate incident before culminating in a hate crime. Study 1 examined whether there was a relationship between increased incident severity as the hate scenario progresses and increased respondent engagement with the situational model. Bystander heterogeneity was also explored through a latent class analysis of bystander progression through each phase of the hate scenario. The latent classes of bystanders then served as one of the dependent variables for the next two studies. Study 2 examined whether bystander progress through the situational model differed based upon the victim’s identity. Study three tested whether bystander characteristics such as empathy, bystander efficacy, and regard for others predict bystander latent class. Several key findings emerged from these three studies. First, bystander progress through the situational model indeed increases as the hate event escalates. This relationship was observed through examining whether the bystander saw a problem, whether the bystander intervened, and the number of steps the bystander proceeded through in the situational model. Second, three iii latent classes of bystanders were found in this sample: always interveners, those who progressed further through the model as the event escalated, and those who never intervened. Third, situational model progress significantly differed based on the identity of the victim. This most prominently occurred with subjects being significantly more likely to see a problem and intervene for the Asian woman victim than the sexual minority and gender minority victims. Lastly, the bystander’s favorability towards the victim’s group was consistently the strongest predictor of whether the bystander saw a problem or intervened followed by empathy, bystander behavior, and bystander efficacy. These results together will help to inform the applicability of the situational model to hate as well as the cognitive mechanisms that can be leveraged to enhance the probability of intervention. iv DEDICATION I dedicate this dissertation to my beloved Grandma Judith (Judy) Renwick-Smith (05/27/1941-03/31/2024) and Grandpa Thomas (Tom) Wesley Hubbell II (11/01/1938-04/19/2024), both of whom passed while I was writing this. Although they can’t be here to see me cross the finish line, completing this dissertation serves as a testament to the indelible marks they both had on my personal growth and grit. Grandma and Grandpa, your love endures and reminds me I’m never alone. v ACKNOWLEDGMENTS This dissertation and my progress through the PhD program here would not have been possible without the support of so many people in my life. I think it’s only right to begin with the people who have been there since the beginning. Mom, Dad, and Kristin, I love you all so much. Mom, you taught me the importance of being kind to myself and taking time to enjoy life, whether it be a night in texting each other while watching Drag Race or a night with you at the Straz seeing Jimbo live. Dad, you taught me to live with integrity and how to be strong. I’ll always be grateful for your help with my moves to Albany and D.C. as well as the fatherly advice you’ve dispensed over the years. Kristin, you taught me how to endure the wild things life can throw all while still going out and having fun. I am deeply thankful to you all for being my buoys whenever I’ve felt adrift. To my friends, old and new, thank you for being there for me as the occasional shoulder to cry on and more often as the smiling face reminding me of the joys to be had. To Bernadette, Cassie, and Daniela, thank you for being sympathetic ears when I needed someone who could commiserate with the PhD experience. I’m so proud to count you as my closest friends. I can’t wait for the adventures in store for us when you start at George Mason, Daniela! To my friends I made at Pride Sports Albany Kickball (Emilio, other James, Jim, Jimmy, Maya, Michael, Kira, Ron, and Tim), I cannot express enough how much your friendships mean to me. You’re all the first friends I made outside of school. I can’t describe how affirming it has been to have people who know the real me and not just James the academic. I got to experience whole new parts of myself through that liberation of being able to cut up and be a fool from time to time. I love you all. Thank you for all the fond memories, past and yet to be made. Go Resting Pitch Face! vi To my dissertation committee, thank you for your support and insights that made this document something I can be proud of. Teddy, your mentorship as my dissertation chair has been invaluable. You provided me with pushes in the right direction whenever I needed them, especially when I was grappling with the passing of two of my grandparents and arrival into D.C. Brandon, thank you for your financial support by paying for the Qualtrics survey. This dissertation could not have occurred without your generosity. Dana, thank you for your thoughtful critiques throughout all stages of this process. Your astute questions and comments forced me to engage more deeply with the subject matter, helping me to feel the passage from student to expert. Last but certainly not least, to the OG mentor, Dr. Kathleen M. Heide, you deserve your own paragraph. You are the Athena to my Odysseus. I still remember being so enchanted with your research presentation as a freshman in the USF Honors College that I went back to my dorm room after and began looking into the degree in criminology immediately. While I knew that I would flourish under you academically, I did not know the impact you would have on my growth as a human being. You ignited a passion in helping the downtrodden and vulnerable that continues to burn over 10 years after having met and worked with you. I vividly recall being the brazen, judgmental 19-year-old who rarely believed in giving others second chances. Your gentle questioning of my judgments and insistence that there are few among us who are incapable of redemption completely transformed my worldview. I am truly a better person for having known you. I feel so honored that my dissertation was your last before your retirement. It is astonishing to think of the number of students’ lives you must have affected in your 43 years as a professor. I feel remarkably privileged to have been one. vii TABLE OF CONTENTS Chapter 1: Introduction ................................................................................................................1 Chapter 2: Hate ............................................................................................................................12 The Continuum of Hate .....................................................................................................12 Microaggressions ...............................................................................................................13 Hate Incidents ....................................................................................................................16 Hate Crimes .......................................................................................................................18 Summary ............................................................................................................................21 Chapter 3: Efforts to Prevent Hate Crime ................................................................................22 Legislative Actions to Address Hate Crime .......................................................................22 Federal Law ...........................................................................................................22 State Law ...............................................................................................................24 Police Response to Hate Crime..........................................................................................26 Community Programs ........................................................................................................28 Summary ............................................................................................................................30 Chapter 4: Bystander Intervention ............................................................................................31 Bystander Intervention to Microaggressions, Hate Incidents, and Hate Crimes ...............32 Factors Conducive to Intervening ......................................................................................35 Victim Variables ....................................................................................................35 Bystander Variables ...............................................................................................38 Summary ............................................................................................................................41 Chapter 5: Trends in Public Opinions of Minority Groups ....................................................42 Public Opinions of Asian Americans .................................................................................43 Public Opinions of Sexual Minority Individuals ...............................................................45 Public Opinions of Gender Minority Individuals ..............................................................46 Summary ............................................................................................................................47 Chapter 6: Theoretical Overview ...............................................................................................49 Bystander Intervention Models ..........................................................................................49 Overview of Latané and Darley’s Situational Model ........................................................51 viii Implicit Bias .......................................................................................................................54 Heuristics ...........................................................................................................................56 Social Categorization Theory ............................................................................................58 Chapter 7: Methods .....................................................................................................................62 Sample Characteristics .......................................................................................................62 The Vignette.......................................................................................................................63 External Validity of the Vignettes .........................................................................65 Social Desirability Bias..........................................................................................66 Dependent Variables ..........................................................................................................67 Independent Variables .......................................................................................................68 Analytic Strategy ...............................................................................................................71 Research Question 1 ..............................................................................................71 Research Question 2 ..............................................................................................73 Research Question 3 ..............................................................................................74 Chapter 8: Study One Results ....................................................................................................78 Review of Hate ..................................................................................................................78 Theoretical Background .....................................................................................................80 Progression through the Situational Model........................................................................80 Latent Class Analysis .........................................................................................................83 Discussion ..........................................................................................................................88 Chapter 9: Study Two Results ..................................................................................................104 Review of Victim Traits Affecting Bystander Intervention ............................................104 Theoretical Background ...................................................................................................106 Results ..............................................................................................................................108 Seeing a Problem for Each Phase ........................................................................108 Intervening for Each Phase ..................................................................................111 Latent Class Distribution by Victim ....................................................................113 Discussion ........................................................................................................................115 Chapter 10: Study Three Results .............................................................................................123 Review of Bystander Traits Affecting Bystander Intervention .......................................123 Theoretical Background ...................................................................................................126 Results ..............................................................................................................................128 ix Seeing a Problem for Each Phase ........................................................................128 Intervening for Each Phase ..................................................................................130 Latent Class Membership ....................................................................................133 Sensitivity Analyses .............................................................................................135 Stratification Analyses .........................................................................................136 Discussion ........................................................................................................................141 Chapter 11: General Discussion and Conclusion ....................................................................156 Summary of Main Findings .............................................................................................157 Study 1 .................................................................................................................157 Study 2 .................................................................................................................159 Study 3 .................................................................................................................161 Collective Findings ..............................................................................................163 Theoretical Implications ..................................................................................................166 Situational Model .................................................................................................166 Social Categorization Theory ..............................................................................169 Policy Implications ..........................................................................................................171 Limitations and Directions for Future Research ..............................................................174 Conclusion .......................................................................................................................178 References ...................................................................................................................................179 Appendix .....................................................................................................................................217 x LIST OF FIGURES AND TABLES Table 1. Sample Demographic Characteristics (n=1,001) .............................................................76 Table 2. Progression through the Situational Model by Hate Event Phase (n=1,001) ..................92 Table 3a. Hate Incident Progress through the Situational Model by Microaggression Progress (n=1,001) ............................................................................................................................93 Table 3b. Hate Crime Progress through the Situational Model by Microaggression Progress (n=1,001) ............................................................................................................................94 Table 3c. Hate Crime Progress through the Situational Model by Hate Incident Progress (n=1,001) ............................................................................................................................95 Table 1a. McNemar's Test of Seeing a Problem for Microaggression and Hate Incident (n=1,001)............................................................................................................................................96 Table 4b. McNemar's Test of Seeing a Problem for Microaggression and Hate Crime (n=1,001)............................................................................................................................................96 Table 4c. McNemar’s Test of Seeing a Problem for Hate Incident and Hate Crime (n=1,001)....96 Table 2a. McNemar's Test of Intervening for Microaggression and Hate Incident (n=1,001) .....97 Table 5b. McNemar’s Test of Intervening for Microaggression and Hate Crime (n=1,001) ........97 Table 5c. McNemar’s Test of Intervening for Hate Incident and Hate Crime (n=1,001) .............97 Table 3a. Wilcoxon Signed-Rank Test for Microaggression and Hate Incident Situational Model Progress (n=1,001) .............................................................................................................98 Table 6b. Wilcoxon Signed-Rank Test for Microaggression and Hate Crime Situational Model Progress (n=1,001) .............................................................................................................98 Table 6c. Wilcoxon Signed-Rank Test for Hate Incident and Hate Crime Situational Model Progress ..............................................................................................................................99 Table 4. Ten Most Frequent Combinations of Situational Model Progression by Hate Event Phase (n=1,001) ...............................................................................................................100 Table 5. Latent Class Analysis Selection Criteria .......................................................................100 xi Table 6. Three-Class Solution Latent Class Marginal Means .....................................................101 Table 7. Five-Class Solution Latent Class Marginal Means........................................................102 Table 8. Chi-Square Test of Independence for Three- and Five-Class Membership (n=1,001) .103 Table 9. Chi-Square Tests of Victim Identity and Seeing a Problem ..........................................120 Table 10. Chi-Square Tests of Victim Identity and Intervening ..................................................121 Table 11. Chi-Squares of Victim Identity and Bystander Latent Class Membership ..................122 Table 12. Logistic Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) ..........................................................................................................................147 Table 13. Logistic Regressions of Intervening for Each Phase on Bystander Characteristics (n=1,001) ..........................................................................................................................148 Table 14. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics (n=1,001) .................................................................................................149 Table 18. Stratification Analyses for Seeing a Problem During Each Phase ..............................150 Table 19. Stratification Analyses for Intervening During Each Phase ........................................152 Table 20. Stratification Analyses for Latent Class Membership .................................................154 Table A1. Hate Crime Changes Over Time by Bias Type 2018-2022 ........................................217 Table A2. Hate Crime Changes Over Time by Bias Category Type 2018-2022 ........................220 Table A3. Two-Class Solution Latent Class Marginal Means ....................................................221 Table A4. Omnibus Chi-Square of Victim Identity and Latent Class for the Five-Class Solution..........................................................................................................................................222 Table A5. OLS Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) ..........................................................................................................................223 Table A6. OLS Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) ..........................................................................................................................224 Table A7. Logistic Regressions for Seeing a Problem for Each Hate Event Phase Using Main Effects ..............................................................................................................................225 Table A8. Logistic Regressions for Intervening for Each Hate Event Phase Using Main Effects..........................................................................................................................................226 xii Table A9. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics Using Main Effects .................................................................................227 Table A10. Logistic Regressions for Seeing a Problem for Microaggression by Victim............228 Table A11. Logistic Regressions for Seeing a Problem for Hate Incident by Victim .................229 Table A12. Logistic Regressions for Seeing a Problem with Hate Crime by Victim .................230 Table A13. Logistic Regressions for Intervening for Microaggression by Victim .....................231 Table A14. Logistic Regressions for Intervening for Hate Incident by Victim...........................232 Table A15. Logistic Regressions for Intervening for Hate Crime by Victim ..............................233 Table A16. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Asian Man Victim .............................................................................234 Table A17. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Asian Woman Victim ........................................................................235 Table A18. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Gay Man Victim ................................................................................236 Table A19. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Lesbian Victim ..................................................................................237 Table A20. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Trans Man Victim..............................................................................238 Table A21. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Trans Woman Victim ........................................................................239 Table A22. Logistic Regressions for Seeing a Problem for Each Hate Event Phase by 2020 Election Results ...............................................................................................................240 Table A23. Logistic Regressions for Intervening for Each Hate Event Phase by 2020 Election Results ..............................................................................................................................241 Table A24. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for 2020 Trump States .............................................................................242 Table A25. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for 2020 Biden States ..............................................................................243 Table A26. Logistic Regressions for Seeing a Problem for Each Hate Event Phase by White or non-White Respondents ...................................................................................................244 xiii Table A27. Logistic Regressions for Intervening for Each Hate Event Phase by White or non-White Respondents ..........................................................................................................245 Table A28. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for White Respondents ............................................................................246 Table A29. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for non-White Respondents .....................................................................247 Table A30. Stratifications by Randomly Assigned Victim Identity, 2020 Election State Residency, and White versus non-White Respondents....................................................248 Figure A1. Year-on-Year Percent Changes in Hate Crime by Bias Category Type 2018-2022 .250 1 Chapter 1: Introduction What hurts the victim most is not the cruelty of the oppressor but the silence of the bystander. —Elie Wiesel Hate crimes, defined by the Federal Bureau of Investigation (2022, para. 6) as any “criminal offense against a person or property motivated in whole or in part by an offender’s bias against race, religion, disability, sexual orientation, ethnicity, gender, or gender identity1” are on the rise in the U.S. These overwhelmingly (70%) public offenses have increased by seven percent alone from 2021 to 2022 with consistent increases since 2018 averaging 13.1% annually (see Table A1 in the Appendix) (UCR, 2024). Increases have been observed across a number of vulnerable communities (see Table A2 and Figure A1 in the Appendix). Crimes targeting transgender individuals are up 33% from 2021 to 2022 (UCR, 2024), while those targeting Asian American individuals have also rapidly increased (Kim et al., 2023; Lantz & Wenger, 2020). Hate crimes against other groups, such as Black or African Americans and individuals with disabilities, have remained stable or experienced fluctuations while those targeting LGBT individuals have demonstrated sustained increases over the past five years (UCR, 2024). Most efforts to date on preventing hate crimes have focused on the creation of laws at both the federal and state level designed to enhance the punishment for hate crime, formal criminal justice system responses, such as law enforcement officer training, and community-based programs. The focus on formal law enforcement response among previous policies may 1 According to the American Psychological Association (APA) (2012), sexual orientation is defined by the sex of those to whom someone has sexual, romantic, and/or emotional attractions. Sexual orientations most commonly include being heterosexual, bisexual, gay, or lesbian though other sexual orientations exist. Per the APA (2009), gender identity refers to a person’s own sense of belonging to the male, female, or other sex. 2 inherently exacerbate other inequities within police response. Previous research has found that police response time is slower in cases with non-White victims (Howerton, 2006), increasing the likelihood of prejudicial statements from the perpetrator metamorphosizing into violence against the victim as time passes within hate crime incidents. Improving police response time alone is not the answer as Bayley (1996) noted; police must respond within one minute of the commission of a crime for any effects on clearance rate to be detected. Such a speedy response time is unlikely given the progression of multiple phases that occurs within hate crimes, coupled with the ambiguity of the earliest phases. Further, reporting to the police may not be the panacea for hate crimes given that many racial/ethnic, sexual, and religious minority persons and hate crime victims prefer not to have contact with the police (Lantz & Wenger, 2022; Rennison et al., 2011). Victims from minority groups often choose not to report to the police because of dissatisfaction with them and general mistrust (Xie & Baumer, 2019). Additionally, community programs have been developed to prevent hate crime (Freilich & Chermak, 2013). These programs foster intergroup bond formation to reduce the likelihood of hate crime perpetration. Such programs are premised on intergroup contact theory, which posits that knowing and having positive interactions with members of outgroups helps to improve attitudes towards those outgroups (Blau, 1977; Pettigrew, 1998). Other community programs educate teachers and other school staff members on identifying the risk factors for hate crime perpetration since many hate crime offenders are teenagers (McDevitt et al., 2021). Some of these programs emphasize the role of onlookers in stopping hate crime through a more promising ameliorant for halting them as they occur: bystander intervention. Given that so many hate crimes occur in public, bystander intervention appears to be an auspicious method to stop hate crime. Bystander intervention refers to the actions taken by 3 witnesses to potentially hazardous events that can either prevent or mitigate harm (Darley & Latané, 1968). Emerging shortly after the stabbing death of Kitty Genovese in 1964, where dozens of individuals failed to assist, the five-step model proposed by Latané and Darley (1970) served as a mechanism to improve intervention efforts. Several decades later, bystander intervention remains relatively uncommon. For example, only 45% of individuals who reported witnessing anti-Asian hate crimes during the early stages of the COVID-19 pandemic in the United States ultimately intervened (Lui et al., 2021). Often transitioning from verbal harassment to physical violence or threat, the processual nature through which most hate crimes occur also provides multiple opportunities for bystander intervention. Data from the National Crime Victimization Survey indicate that 99% of hate crime victims knew they were being targeted because of their identities due to hate language used during the attacks (Masucci & Langton, 2017). Victims of hate crime have also shared that their personal violent experiences were preceded by microaggressions, which are brief statements that, intentionally or not, convey hostility towards groups of individuals based on identity (Sue et al., 2007), or by hate speech (Meyer, 2012). The verbal beginnings of many hate crimes provide onlookers with time to engage in decision-making processes to ultimately assist victims, potentially preventing the hate incident from escalating into hate-based violence. Several documented encounters support this potential pathway to hate crime prevention. In one notable instance, a group of people began shouting at a same-sex female couple for kissing in a hotel pool (Barmann, 2021). Bystanders intervened on the couple’s behalf by shouting back at their aggressors who were then removed by hotel security. In another example, a group of over twenty men beat and hurled prejudicial insults at Iyanna Dior, a Black transgender woman, outside of a convenience store (Burns, 2020). Witnesses rushed to her aid, 4 placing themselves between her and her attackers. Their quick thinking enabled Iyanna to escape through the store’s back exit, leaving her only with bruising and swelling. By contrast, in another case, two gay men were hospitalized after being battered at a bodega in New York City (Quinn, 2021). Their assailants began by using homophobic slurs which then swiftly escalated into violence as they stabbed one victim and repeatedly struck the other with a liquor bottle. No intervention by witnesses to the attack was reported. Likewise, one hate incident against an Asian American woman, Vilma Kari, escalated from Anti-Asian slurs into violence while bystanders watched (Sisak, 2021). Surveillance videos showed two workers in a nearby apartment complex who witnessed the attack while it was occurring and did nothing to assist. One person on the street yelled at the assailant only after the situation became violent. By then, the attacker had already kicked Ms. Kari in the stomach and stomped on her head repeatedly. The two workers checked on her and contacted police only after the assailant had left the scene. These cases illustrate that effective bystander intervention can be potentially lifesaving, especially compared against no intervention. Bystander intervention holds much promise for reducing the incidence and severity of hate crimes, yet current research remains limited. The majority of bystander intervention research to date has focused on sexual assault rather than hate crime (Kettrey & Marx, 2021). Scholars have found that bystander intervention programs significantly decrease the prevalence of sexual assault on college campuses (Labhardt et al., 2017). These results also extend to intimate partner violence (Park & Kim, 2023). Preliminary evidence from one program aimed at combatting hate crime in communities suggests bystander intervention may be an effective means of halting hate crimes (Zempi et al., 2021). Specifically, community members reported a 5 significantly higher likelihood of intervening when witnessing hate crimes after receiving bystander intervention training. One factor that impacts bystander decision-making is incident severity. Incident severity has been shown to impact bystander decision-making in the contexts of sexual violence (Bennett & Banyard, 2016), cyberbullying (Huang et al., 2023), and intimate partner violence (Chabot et al., 2009). Across all three studies, perceived higher severity of the offense significantly increased bystander willingness to intervene. Huang et al. (2023) also discovered that incident severity mediated the relationship between feeling obligated to assist and ultimately assisting. Hate offenses notably exist on a continuum from microaggressions to overt prejudicial statements to physical violence (Schweppe & Perry, 2022). As noted, verbal attacks often precede physical attacks in hate crimes, providing onlookers with opportunities for early intervention that can halt the progression to physical violence (Masucci & Langton, 2017). This continuum of hate thus enables a unique look into how hate severity might impact bystander decision-making. The escalation that occurs during hate events merits a closer examination as studies in other contexts look at separate incidents of differing severity rather than one continuously worsening event. Prior studies have explored typologies of bystanders by using latent class analysis. Latent class analysis categorizes subjects into distinct classes in which subjects are similar to members of the same class but distinct from subjects in other classes (McCutcheon, 1987). These studies have explored latent classes of bystanders for bullying (Jenkins et al., 2021), cyberbullying (Jia et al., 2022; Schultze-Krumbholz et al., 2018) and sexual violence (McMahon et al., 2019) but not for hate crime. These studies used perceptions of the incidents as indicator variables as opposed to bystanders’ cognitive processing of the events. This dissertation builds on existing 6 research by investigating decision-making at three points in time and examining whether distinct classes of bystanders exist based upon their decision-making at these specific points. Three specific questions are asked. Research Question 1: Does bystander progression through the situational model increase concomitantly with increasing severity of the hate scenario? This dissertation seeks to answer the first research question using a series of pairwise comparisons between the steps via dependent-samples tests: McNemar’s tests and Wilcoxon signed-rank tests. A latent class analysis to determine if distinct latent classes exist based upon bystander decision-making at each point then follows. Victim group membership also influences bystander behavior. Minority group members frequently receive less aid from bystanders (Seaton et al., 2018). Part of this is due to the nature of prejudice. Some ingroup members appoint themselves to be morally superior compared to outgroup members, often observing them with disgust (Brewer, 1999). This disgust culminates in the decision not to intervene as that passivity expresses the ingroup member’s disapproval of the victim. While there are also other considerations that bystanders make such as safety, these have received little empirical examination to date (Mainwaring et al., 2023). The decreased propensity for bystanders to help minority group members has appeared across several domains. Bystanders to sexual assault on college campuses report being less likely to help Black female victims than white female victims (Katz et al., 2017). Bystanders to sexually prejudiced hate crimes likewise report being less willing to help gay victims than straight victims (Owuamalam & Matos, 2020). Men who observe violence against women are less willing to intercede if slurs identifying the women as transgender were used (Deacon et al., 2020). Few studies have examined how various categories of minority group membership 7 differentially impact bystander decision-making. Vera et al. (2023) noted college students were more likely to intervene in bullying cases with racial minority victims than with sexual or gender minority victims. It remains unknown whether these results replicate in hate crime scenarios and whether a disparity in bystander decision-making exists across various minority groups; therefore, research question 2 of this dissertation is: Research Question 2: Does the decision-making of bystanders to hate crimes differ based upon whether the target is a racial, sexual, or gender minority victim? Characteristics of the bystanders themselves also affect their willingness to intervene. These traits primarily include empathy (Nickerson et al., 2015), bystander efficacy (Fischer et al., 2011), and decisional balance (Banyard & Moynihan, 2011). Individuals with higher levels of empathy report being more likely to intervene on behalf of victims because of their heightened ability to empathize with victims’ distress (Nickerson et al., 2015). Bystanders who report greater confidence in their abilities to intervene also are more likely to intervene because that confidence emboldens them to act (Fischer et al., 2011). Decisional balance has been conceptualized in the form of a cost-benefit analysis performed by the bystander (Banyard & Moynihan, 2011). If that cost-benefit analysis favors intervention, then the bystander will intervene. Additionally, demographic characteristics of bystanders including race, sexual orientation, and gender identity have been shown to influence intervention both alone and in interaction terms representing intersectionality (Brown et al., 2014; Burns et al., 2019). White men, for example, appear to be less willing to intervene when witnessing sexual assault scenarios than white women, Black men, and Black women (Burns et al., 2019). Studies of bystander characteristics tend to focus on one specific trait as an explanation for bystander behavior. Several studies, however, have examined how several bystander-level 8 variables coalesce to impact bystander willingness to intervene (Jenkins et al., 2021; Jin et al., 2022; McMahon et al., 2021). These studies too focused on one trait, typically empathy, as opposed to several, and did not examine bystander intervention in hate events. Given the unique qualities of hate crime, as well as the importance of personality traits on intervention, research question three of this dissertation is: Research Question 3: Do bystander characteristics influence their bystander decision-making processes while witnessing hate? To address these gaps, this dissertation project will examine how bystanders observe, engage with, and eventually do or do not make their decisions to intervene. To do so, a framework of the five-step situational model developed by Latané and Darley (1970) will be used. They proposed that a bystander must witness the incident, consider it a problem, feel obligated to assist, and be able to generate options to help prior to completing the fifth step of implementing the selected intervention. Although researchers have developed other bystander intervention models that will be discussed in the theoretical overview chapter, these have been subjected to little empirical investigation and possess constraints that do not make them ideal for investigating bystander intervention to hate crime. A novel scenario-driven sequential decision-making survey instrument allows for individuals to intervene at each stage as the experience progresses from microaggression to hate incident to hate crime involving the use of violence, and thereby prevent the incident from even requiring law enforcement involvement, where necessary. Thus, this dissertation ultimately aims to explore the nuances of bystander intervention within the context of hate crime. The decision-making process of bystanders through hate crime scenarios will be explored based upon victim race, sexual orientation, and gender identity including transgender victims. 9 This dissertation will specifically study bystander intervention in hate events against Asian American, gay, lesbian, and transgender individuals. These groups are being selected for a few reasons. Hate crimes against Asian Americans have considerably increased since the COVID-19 pandemic due to Sinophobic rhetoric blaming Chinese people for the pandemic (Kim et al., 2023). Anti-gay and anti-transgender sentiments have also increased sharply within the past two years, incited by the boom in anti-LGBT legislation (Jones, 2023a, 2023b). Hate crimes against LGBT individuals have also risen dramatically (UCR, 2024). Furthermore, more research needs to make visible the dynamics of gay, lesbian, and transgender victims of hate crimes as they have been largely ignored within the extant literature, with researchers also decrying their absence from bystander intervention education curricula (Kirk-Provencher et al., 2023). The unprecedented growth in recent years of LGBTQ+ communities further merits increased attention towards hate against sexual minority individuals and gender minority individuals. Gallup estimates show that over 11% of millennials and nearly 20% of Generation Z self-identify as being members of the LGBT community (Jones, 2023c). Lastly from 2018 to 2022, hate crimes based on gender identity and sexual orientation have increased an average of 29.3% and 13.1% annually, respectively (see Table A2 in the Appendix). Hate crimes based on the victim’s race/ethnicity/ancestry additionally increased by 14.3% annually on average, but that appears to be largely driven by increases in anti-Asian hate crime as they increased by 49.1% on average annually from 2018 to 2022 (see Table A1 in the Appendix). Given that these groups experienced the largest average increases in hate crime annually, they in particular merit special attention given the pressing timeliness of these issues. The potential impacts of these studies are considerable. First, these research results will inform existing training efforts to help prevent hate crime offending via improved bystander 10 intervention and de-escalation. These efforts echo recent calls to establish parallel and alternative solutions to combatting hate crime that are not solely dependent upon a police and carceral response (Brennan Center for Justice, 2021). Second, creating effective bystanders promotes the reporting of hate crimes to the police, as law enforcement notification is one important method of bystander intervention. Estimates place hate crime reporting at less than 40% (Davis & O’Neill, 2016), and even the International Association of Chiefs of Police (IACP) acknowledges the need for community-based interventions outside formal law enforcement response (IACP, 2019). Although law enforcement can de-escalate hate incidents prior to violence, it may already be too late by the time they arrive on scene. Further compounding this, onlookers may not contact the police when witnessing microaggressions or hate speech if they know these do not rise to criminal levels. Answers to this dissertation’s research questions will also help to galvanize the efforts of researchers, community members, and practitioners. Knowing the extent to which hate severity influences progression through the situational model will not only validate the basic premise that increased severity positively influences bystander decision-making, but also that bystanders perceive microaggressions, hate incidents, and hate crimes differently. These results will highlight the need for hate-related bystander trainings to delineate the negative consequences of all these acts to reinforce the need for intervention regardless of their initially perceived severity. Further, noting possible disparities in bystander decision-making based upon race versus sexual orientation versus gender identity of the victim will inform practitioners who provide these trainings that hate against each of these groups needs to be considered distinctly from the others and not reified into one large construct of hate. If would-be bystanders perceive hate events differently solely because of the identity of the victim, then trainings must account for that by 11 adequately educating subjects on how hate may appear differently against various groups. Moreover, trainings must address the harms associated with hate that may be exacerbated against certain minority groups to create equitable bystander response. Lastly, if bystanders to hate crime fall into distinct latent classes, the indicator variables such as empathy, bystander efficacy, and other-regarding will show practitioners what traits should be fostered and primed in bystanders to maximize the likelihood of intervention. This dissertation will now proceed as follows. An overview of hate will be presented with attention given to the continuum of hate, microaggressions, hate incidents, and hate crimes. Next, the history of the legislative response to hate crime in the U.S. will be summarized along with synopses of traditional efforts to prevent hate crime. The bystander intervention literature will then be reviewed broadly in its conceptualizations and successes, then specifically within microaggressions, hate incidents, and hate crimes. Factors associated with bystander intervention such as victim traits and bystander traits will then be summarized. Public opinion trends pertaining to Asian Americans, sexual minority individuals, and transgender individuals will then be provided to further bolster this dissertation’s emphasis on these select groups. The theoretical underpinnings of both Latané and Darley’s situational model will be explored along with social categorization theory. The methods proposed to answer the research questions will be given in detail for each study in addition to a summary of the vignette instrument and the included variables. Findings are then presented for the three studies. Lastly, an overall discussion with a summary of the findings, theoretical and policy implications of the dissertation, limitations, and directions for future research. 12 Chapter 2: Hate In this chapter, the phenomenon of hate will be described in detail. First, the continuum of hate will be explored to demonstrate that hate encompasses a variety of beliefs and actions. Individual aspects of the continuum most relevant to hate crime will then be examined including their definitions, how they are perpetrated against the groups being studied in this dissertation, their prevalence, and the consequences for victims and communities. The Continuum of Hate Hate occurs across a continuum of severity (Schweppe & Perry, 2022). This continuum, from least to greatest severity, consists of microaggressions, hate speech, terrorism, and genocide. Although microaggressions are commonly considered the least serious form of hate, their frequency results in significant harm against individuals and communities (Waldron, 2012). Moreover, hate speech may begin as insults and prejudicial slurs directed towards an individual, but they all too often escalate into violent hate crime assaults (Schweppe & Perry, 2022). Terrorism exists on this continuum as the goal within the context of hate to intimidate entire communities of often marginalized individuals. Lastly, genocide is the pinnacle of hate as it involves the extreme dehumanization of a group that endorses all forms of violence including sexual assault. The Anti-Defamation League (2003) published the “Pyramid of Hate” as a visual representation of this continuum. The pyramid’s base is composed of prejudiced attitudes evidenced by condoning prejudicial jokes. Although listeners are not making these jokes, they are still complicit by not stopping them. Failure to interrupt them tacitly signals to others their agreement. The next level of the pyramid consists of discrimination and acts of prejudice that are often individual-specific and entail the stereotyping of individuals by their identities as an Other. 13 Violence constitutes the middle of the pyramid and includes threats, assaults, murder, and terrorism. Genocide acts as the top of the pyramid as the ultimate expression of hate through the extermination of a group of people. The Anti-Defamation League (2003) noted these levels build on one another in the sense that ignoring minor acts of hate establishes environments conducive to increasingly serious hate. Hate crimes themselves reflect this continuum of hate in how they unfold. For example, Herek and colleagues (2002) found from interviews with sexual minority victims of hate crime that their assaults often began with homophobic epithets that progressed to physical violence. Furthermore, Masucci and Langton’s (2017) analysis of NCVS hate crime data revealed that 99% of hate crime victims knew they were being targeted because of their identities due to hate language used during the attacks. Thus, hate exists on a continuum of which lesser verbal transgressions escalate to greater physical crimes. Microaggressions As noted, microaggressions are brief statements that, intentionally or not, convey hostility towards groups of individuals based on identity (Sue et al., 2007). Microaggressions may occur as microassaults, microinsults, and microinvalidations (Sue et al., 2008). Microassaults are overt messages designed to harm the victim through intentional discrimination, epithets, or purposeful avoidance (e.g., refusing to serve customers because they are transgender). Microinsults disparage someone’s identity through rude or insensitive comments (e.g., implying that someone was hired because of affirmative action rather than their merits). Microinvalidations undermine an individual’s experiences as a minority group member (e.g., telling a minority group member who has disclosed being the victim of discrimination that they are being overly sensitive). These are discussed below in the context of the broader term of microaggressions. 14 Sue and colleagues (2007) noted nine microaggressions frequently suffered by Asian Americans. These range from positive stereotypes (e.g., the view of Asian Americans as being good at math) to negative stereotypes such as the perception of Asian Americans as poor drivers. Asian Americans may also be treated as outsiders despite being born in the U.S., have their experiences minimized for being a “model minority”, be fetishized by being treated as sexual objects rather than people, have their ethnicities used interchangeably, be treated as second class citizens, and be left out of racial discussions. Microaggressions against Asian Americans, for example, are ubiquitous. Authors of an epidemiologic study discovered that 72% of Asian Americans in the sample faced microaggressions just within a two-week period (Ong et al., 2013). Although microaggressions are most commonly associated with race, they can also occur in the context of sexual orientation and gender identity (Nadal, 2011). Several types of microaggressions based on sexual orientation have been found (Platt & Lenzen, 2013; Sue, 2010). These include oversexualization through the conflation of sexual minority membership with sexual promiscuity, homophobia through the avoidance of gay men and lesbians, use of heterosexist language and terminology that assume heterosexuality, viewing sexual minority individuals as sinful, treating non-heterosexual orientations as pathological, denying one’s own heterosexism, and the endorsement of heteronormativity. Platt and Lenzen (2013) discovered two additional forms of microaggressions against sexual minority individuals: under-sexualization by closeting sexual minority individuals, and cracking jokes intended to demean the victim. In a nationally representative survey conducted by GLSEN, roughly two-thirds of LGBTQ+ middle and high school students reported hearing homophobic microaggressions while at school in 2021 (Kosciw et al., 2022). 15 The various forms of homophobic microaggressions extend to transgender individuals as well (Nadal et al., 2010; Sue & Capodilupo, 2008). Specifically, these are misgendering, reification of gender non-conforming experiences, fetishization, disapproval, endorsement of cisnormativity, denial of transphobic experiences, and pathologizing gender non-conforming identities. Nadal et al. (2012) later expanded this taxonomy to include harassment, denial of one’s own transphobia, invading the bodily privacy of trans individuals (e.g., asking about a trans person’s genitals), familial rejection due to a member’s gender identity, and systemic or environmental microaggressions such as public bathroom usage and difficulties changing gender on government documents. Approximately 56.2% of LGBTQ+ middle and high school students in the aforementioned GLSEN survey indicated they heard microaggressions against someone’s gender expression (Kosciw et al., 2022). Despite the prefix micro indicating small, microaggressions have been shown to have large adverse effects due to the cumulative effects of their frequency. Moreover, victims often fixate on them given their nature as seemingly minor or innocuous insults that nonetheless convey harmful messages (Sue et al., 2007). Microaggressions have been associated with harm to both physical and mental health (Costa et al., 2023). Regarding the former, microaggressions have been linked to binge drinking (Blume et al., 2012), fatigue (Nadal et al., 2017) and general physical health symptoms such as headaches and pain (Huynh, 2012). Regarding the latter, researchers have demonstrated microaggressions lead to depression and anxiety symptoms (Nadal et al., 2014) as well as negative emotionality (Wang et al., 2011) and higher stress levels (Smith et al., 2011). Microaggressions can also impact nearby witnesses, even if they are not members of the targeted group (Sue, 2010). Witnesses report emotional reactions to the incident by becoming 16 angry or sad on behalf of the victim (Halvorson, 2021). Witnesses who belong to the target’s group may exhibit reduced self-esteem, particularly related to that part of their identity, and reduced connectiveness with others (Sue, 2010). These consequences have been reported for witnesses who are not members of the target’s group (Sue, 2010). Moreover, such bystanders may then start to view microaggressions as normal, especially when they occur frequently and without reproach (Sue et al., 2007). Hate Incidents Although the extant literature contains various definitions for hate incidents (see Vergani et al., 2022 for a pre-registered Campbell systematic review of the topic), the proposed study will use the Department of Justice’s definition as that ensures the reviewed literature and proposed studies conform to the criminal justice system’s view of hate. Per the United States Department of Justice (n.d.), hate incidents, also called bias incidents, encompass prejudicial conduct that does not rise to the level of criminality. Although hate incidents can include microaggressions, they also include more overt forms of prejudice such as the use of hate speech. Parekh (2012) defined hate speech as speech that “expresses, encourages, stirs up, or incites hatred against a group of individuals distinguished by a particular feature or set of features such as race, ethnicity, gender, religion, nationality, and sexual orientation” (p. 40).2 Hate speech can be identified through the subject of the speech, the stigmatization of some aspect of the subject, and typically by the hostile tone of the speaker. Anti-Asian hate incidents have evolved following the COVID-19 pandemic. Prior to 2020, hate incidents against Asian Americans typically revolved around the “model minority” 2 No laws exist in the United States that limit hate speech as the Supreme Court has ruled such laws violate the first amendment’s right to freedom of speech in Matal v. Tam (2017). As such, no legal definition has been provided for hate speech. 17 myth (Kim, 2007). For example, many hate incidents against Asian Americans resulted from offenders viewing them as competitive threats in workplaces and schools. Hate incidents against Asian Americans during and subsequent to the pandemic commonly characterize Asian Americans as being responsible for the genesis and spread of the virus (Gover et al., 2020). Hate speech against Asian Americans rose considerably during the COVID-19 pandemic due to the spread of misinformation on the virus (Kim & Kesari, 2021). Homophobic hate incidents and hate speech have contained the same prejudicial beliefs over time. Historically, hate incidents against lesbian, gay, and bisexual individuals involved the stigmatization of the “homosexual lifestyle” by ascribing such qualities as drug use, prostitution, and the indoctrination of others into same-sex sexual behaviors (Mutz, 2006). Homophobic hate speech has recently risen in the U.S. One study found a resurgence in the false equivalence of LGB people as “groomers” with a 406% increase in social media posts characterizing LGB people as “groomers” after the Parental Rights in Education bill passed in Florida in 2022 (Center for Countering Digital Hate and the Human Rights Campaign, 2022). Transphobic hate incidents tend to convey that transgender individuals deny “biological reality” by not identifying with a gender consistent with their sex assigned at birth (Blyth & McRae, 2018). For example, transphobic hate has occurred in response to the passage of “bathroom bills” that force individuals to use bathrooms of their sex assigned at birth regardless of their gender identity (Councilor, 2021). In these cases, the focus of hate is on the invasion of male or female spaces by “deceivers” (Bettcher, 2007). Additionally, polemic rhetoric that falsely claims gender affirming care for youths involves genital mutilation further fuels transphobia (Libby, 2022; Schipper, 2022). The participation of transgender athletes in sports has also sparked transphobia recently. Media accounts typically use cisheteronormative framing to 18 imply that the inclusion of transgender athletes is unfair to their cisgender competitors (Bailey & Jones, 2023). As with microaggressions, victims of hate speech report symptoms consistent with post-traumatic stress disorder and depression (Wypych & Bilewicz, 2022). In fact, the processing of hate speech resembles the processing of traumatic events including initial distress, internalization and self-blame, and the development of maladaptive coping mechanisms (Leets, 2002). Moreover, the prevalence of hate speech against ethnic minorities is positively correlated with increases in those groups’ suicide rates (Mullen & Smyth, 2004). Thus, hate speech has significant physical and mental health effects on members of targeted minority groups. Hate Crimes Per the Federal Bureau of Investigation (2022), the current definition of a hate crime is any “criminal offense against a person or property motivated in whole or in part by an offender’s bias against race, religion, disability, sexual orientation, ethnicity, gender, or gender identity.” Several characteristics distinguish hate crimes from other forms of violent crime. Hate crimes more frequently have multiple offenders compared to other violent crimes (23% vs. 13%). Per the NCVS, hate crime perpetrators and victims are typically strangers rather than acquaintances, friends or intimates (Kena & Thompson, 2021). Violent hate crime offenders tend to be male (72.4%), white (45.3%), and above 30 years old (55.3%). Offenders rarely use weapons and most often target their victims in public spaces (Pezzella & Fetzer, 2017). Hate crime offenders commit their crimes for a variety of reasons. McDevitt and colleagues (2010) noted there are four types of hate crime offenders. The most frequent type of hate crime offender is the thrill offender. Thrill offenders are often adolescents or young adults who act in groups to attack those perceived to be different from themselves. They hunt for their 19 excitement by travelling to places they know minority group members frequent such as gay bars, to search for potential victims. Importantly, thrill offenders act on the belief that other community members support their actions. Next, defensive hate crime offenders target people who they consider to be invading their spaces. Defensive hate crime offenders also tend to be teenagers and young adults who act by themselves or with likeminded peers. Retaliatory hate crime offenders assault members of a group they believe to have wronged them. These individuals tend to act alone, be young, and have no history with the specific individuals they victimize. Lastly, mission hate crime offenders deliberately target minority groups in an effort to “cleanse” them. These zealots plan their attacks alone and are older than the other types of hate crime offenders. McDevitt et al. (2010) identified thrill offenders, the most frequent type, as the most deterrable followed in descending order by defensive, retaliatory, and mission hate offenders. As will be discussed further in this dissertation, effective bystander intervention to hate signals societal disapproval towards the offenders, potentially preventing the hate from escalating into violence. Hate crimes against Asian Americans significantly rose the start of the COVID-19 pandemic (Han et al., 2023; Kim et al., 2023; Lantz & Wenger, 2020). Specifically, the number hate crime victimization against Asian Americans spiked 146% in 2020 compared to 2019 in the 26 largest cities in the U.S. from 55 to 135 victimizations (Center for the Study of Hate and Extremism, 2021). From 2018 to 2022, hate crimes against Asian Americans have increased by 49.1% on average annually ranging from 148 offenses in 2018 to 753 in 2021 (UCR, 2024). Researchers believe that growing anti-Asian sentiments that blamed China for the pandemic to be responsible for this uptick. Over 30% of respondents in one survey reported having observed someone blaming Asian populations for the creation and spread of COVID-19 (Ipsos, 2020). 20 Sexual orientation-based hate crimes have likewise been increasing. Per the UCR (2024), hate crimes against gay men increased by 24.5% from 2020 to 2021 (763 offenses to 950 offenses), while those against lesbians increased by 47.7% (128 offenses to 189 offenses) and those against bisexual individuals increased by 94.7% (19 offenses to 37 offenses) in the same period. These increases are likely larger as some states still do not consider sexual minorities to be a protected class of individuals, thereby decreasing the number of hate crimes reported to the FBI by local law enforcement agencies. The political environment in the U.S. appears to account for this increase with the passage of “Don’t Say Gay” bills that prohibit the discussion of LGBTQ+ topics in public schools. Legislation such as these bills act as dog whistles, i.e., coded language used by politicians that appears innocuous but conveys negative meanings to subsets of their audiences, that fuel homophobia and biphobia (Hartman, 2023; Oakley, 2022). Hate crimes against trans individuals spiked by 12.6% in 2021 from 2020 (222 offenses to 250 offenses) (UCR, 2024). The extent of hate against trans individuals is likely much more expansive given victims’ underreporting and the failure of law enforcement to correctly classify transphobic hate crimes (Mallory et al., 2015). Propelling this trend, media often frames transgender victims of violence negatively, blaming their “lifestyle” for their demise (Osborn, 2022). The consequences of hate crime victimization are typically more pronounced than other forms of victimization. Research has found greater physical violence to occur within hate crimes, leading to more severe injuries to the victims (Lantz & Kim, 2019). Hate crime victims also suffer greater psychologically, reporting significantly higher levels of mental distress than non-hate crime victims (Herek et al., 1997). Specifically, they report significantly more symptoms associated with post-traumatic stress disorder and depression (Botcherby et al., 2011). 21 Hate crimes not only impact their victims, but also the communities to which the victims belong. Members of the same group as the victim report higher levels of fear, anger, inferiority (Perry & Alvi, 2012). In fact, Wenger and colleagues (2022) discovered that vicarious victimization of hate crime leads to depressive symptomatology. These feelings are known as “waves of harm” that advance from the victim to their surrounding community, and result in strained intergroup relationships (Iganski, 2001). Summary In sum, this chapter explored the various facets of hate, how they can be identified, and what their impacts are. Hate often begins with microaggressions that then escalate into more overt hate language and prejudice that can morph into physical violence. Per the Anti-Defamation League, microaggressions and stereotypical jokes often go uncontested by bystanders. This passivity signals acceptance of prejudicial views, encouraging those communicating these views to become even more offensive and derogatory. These forms of hate all carry consequences for both the intended victim and the communities to which they belong. In particular, Asian Americans, sexual minority individuals, and gender minority individuals have faced increasingly frequent forms of hate in recent years. These trends merit increased attention to how hate can be stopped through bystander intervention as in this dissertation. 22 Chapter 3: Efforts to Prevent Hate Crime Given all the negative sequelae of hate in all its forms, the development of effective prevention and intervention programs is imperative. Previous efforts have begun with the enactment of hate crime legislation at the federal and state levels. The passages of these laws then led to changes in how law enforcement responded to hate crime. Community programs aimed at changing social norms and enhancing intergroup relationships also arose as a result of the increased attention to hate crimes. In this chapter, each of these responses to hate crime will be explored in terms of their origins and evaluations of their abilities to prevent hate crime. Legislative Actions to Address Hate Crime Federal Law Although the term had not yet been coined, hate crime legislation began with Title I of the Civil Rights Act of 1968 primarily in an effort to protect Black Americans during racial integration. Under this law, a person who "willingly injures, intimidates or interferes with another person, or attempts to do so, by force because of the other person's race, color, religion or national origin" can be prosecuted at the federal level (CRA, 1968). The victim must have been performing federally protected actions such as going to school or voting for the offender to be subjected to federal prosecution. U.S. Congress later expanded both the scope of what constituted hate crime, as well as the various protected groups, across several pieces of legislation. The first, the Hate Crimes Statistics Act (HCSA, 1990), required the FBI to begin data collection on hate crime statistics from the UCR. It also encouraged local law enforcement agencies to assist with gathering these data for transmission to the FBI. The HCSA defined hate crimes as instances of premeditated murder, voluntary manslaughter, aggravated assault, forcible rape, simple assault, intimidation, 23 arson, and vandalism that contained prejudicial elements as to the victim’s race/ethnicity, sexual orientation, or religious affiliation (HCSA, 1990). Both violent crimes and property crimes can therefore be considered hate crimes if evidence of prejudice exists. Congress passed an additional bill in 1994. The Violent Crime Control and Law Enforcement Act (VCCLA, 1994) added disability status as a protected group. President Barack Obama in 2009 signed the Matthew Shepard and James Byrd, Jr. Hate Crimes Prevention Act (HCPA). This law further expanded the protected classes under hate crime laws through the inclusion of gender and gender identity. Later, President Joseph Biden signed the COVID-19 Hate Crimes Act (CHCA, 2021) to combat the rise in Anti-Asian hate crimes during the COVID-19 pandemic. This law also included provisions to fund hate crime reporting hotlines in each state, hasten the Department of Justice’s response to hate crime cases, and provide guidance on investigating hate crimes to law enforcement agencies (CHCA, 2021). In addition to making hate crime a federal offense, the penalties for hate crimes have also been increased. President Clinton signed the Hate Crimes Sentencing Enhancement Act in 1994. This bill provided federal sentencing enhancements for offenders found guilty of having committed hate crimes (HCSEA, 1994). It also made hate crime a federal offense with a definition that extended to any crime as opposed to only federally protected activities. Defenders of hate crime enhancements argue that hate crimes represent an offense unique from the co-occurring offenses and deserving of additional punishment. First, the moral repugnancy of the offender’s primary motive being hate merits greater punishment (Kim, 2006). Second, as noted in Chapter 2, hate crimes affect not just the individual victim, but also the communities of which they are a part (Iganski, 2001). Third, criminal law without hate crime enhancements had not effectively deterred hate crime offenders (Trout, 2015). Fourth, these laws 24 act as moral signals that convey the value of all persons regardless of their identities (Woods, 2008). Federal hate crime laws have their limitations in responding to hate crime. Despite the numerous pieces of legislation enacted against hate crime at the federal level, few hate crimes have been pursued at the federal level. For example, only 27 cases in 1996 had enhanced sentences after the Hate Crimes Sentencing Enhancement Act in 1994 became law (White House, n.d.). A Department of Justice report revealed the continued limitations of federal hate crime laws (Motivans, 2021). From 2005 to 2019, federal prosecutors declined 82% of all hate crime cases referred to them. Only 310 defendants were prosecuted across the 15-year period, though 92% of them were convicted. Attorneys cited lack of evidence (55%), prioritization of federal resources due to existing policies (15%), the federal government lacking jurisdiction over the suspect (13%), referral to another jurisdiction (13%), and alternatives to federal prosecution (4%) as reasons for not charging the other 1,548 hate crime suspects. Thus, the evidence suggests that federal hate crime law lacks the ability to deter hate crime offenders given that the vast majority of cases referred to federal prosecutors are denied. State Law In 1978, the first state hate crime law passed in California (California Penal Code, 1978). This law only provided for penalty enhancement in murders committed based upon the religion, race, or nationality of the victim. Oregon then followed suit in 1981 with its own expansions on the number of protected classes and the inclusion of crimes other than murder (Grattet et al., 1998). Washington also instituted its own hate crime law later that year. Twenty-five states enacted hate crime laws between 1982 and 1995 (Grattet et al., 1998). Currently 46 states and Washington, D.C. have enumerated hate crime laws (Bills & 25 Vaughn, 2023). These states all vary, however, in regard to the classes protected by their laws and in their provisions. Per Bills and Vaughn’s (2023) review, 46 states along with Washington, D.C. possess hate crime laws that protect religion, nationality, and race/ethnicity. Thirty-three states also protect individuals based on sexual orientation. Only 12 states have provisions in their hate crime laws for gender identity; that is, laws that protect transgender and gender non-conforming individuals. The laws in 31 states specify that hate crime data need to be collected. Twenty-seven states allow victims to pursue civil cases against hate crime offenders. Five states mandate that individuals convicted of hate crime charges receive educational programs designed to reduce prejudicial beliefs. Just 13 states’ hate crime laws contain provisions requiring law enforcement officers to receive training on hate crimes. States also vary as to the evidentiary criteria required to prove that a crime was motivated by bias (Pezzella & Fetzer, 2021). The two primary models used by states are the discriminatory selection model and the racial animus model, which is also applied to protected groups other than racial groups. The former only requires that the offender intentionally selected a victim who happens to be from a protected group. The latter requires evidence of malicious conduct against the victim because of the victim’s identity as a member of a protected group. Most states utilize a synthesis of the two standards in determining whether a hate crime occurred as the discriminatory selection model applies a broad standard that can capture crimes that were not motivated by bias. For example, a robber who chooses to target trans women because he believes they are less likely to report to the police than other potential victims would be guilty of a hate crime under the discriminatory selection model but not the animus model. In that instance, the robber is not motivated by hate but rather by stereotypes conveying that trans women will be more suitable targets than members of other groups. The racial animus model, in contrast, is 26 stringent and difficult to prove. Overall, state hate crime laws suffer from several notable weaknesses as responses to hate crime. First, several states still have yet to enumerate hate crime laws, which precludes the genesis of any formal mechanism for reducing hate crime (Bills & Vaughn, 2023). Second, states vary widely regarding which groups constitute protected classes under their hate crime statutes, preventing some hate crime offenders from being subject to enhanced penalties or diversion programs needed to reduce prejudice (Pezzella & Fetzer, 2021). Third, as with federal prosecutors, rarely do state and local prosecutors seek hate crime charges (Eisenberg, 2014). Prosecutors advise hate crimes are notoriously difficult to prove in court; therefore, they often do not pursue hate crime charges in court because they suspect it will distract juries from more easily secured convictions. Moreover, hate crime perpetrators who commit severe offenses such as murder will receive harsh sentences such as life without parole or the death penalty when convicted, further reducing prosecutors’ perceived benefit of seeking hate crime penalties. In the same vein, some states have strict definitions of what constitutes hate crime, limiting the use of court-ordered interventions to rehabilitate offenders (Bills & Vaughn, 2023). Lastly, contributing to the difficulty of pursuing hate crime charges, few states provide law enforcement with training on responding to hate crime (Bills & Vaughn, 2023). Police Response to Hate Crime Law enforcement officers have numerous responsibilities when they respond to hate crimes as they do with other crimes (IACP, 2019). They must secure the scene of the crime, contact emergency medical services for the victim if needed, ensure everyone on scene is safe, collect evidence, complete a preliminary investigation by interviewing those on scene, and make an arrest if probable cause exists. The myriad duties completed by law enforcement officers 27 present myriad problems. The use of law enforcement as a means to combat hate crime has been widely criticized by hate crime scholars for several reasons (Pezzella & Fetzer, 2021). Key inhibitors of an effective law enforcement response include departmental issues such as the lack of effective hate crime investigation training, the absence of hate crime policies, and the complex relationships between marginalized groups and the police. As shown in the previous section, not all state hate crime statutes include provisions for law enforcement training on responding to hate crimes (Bills & Vaughn, 2023). Even when officers do receive such training, it often occurs once and only lasts between one to two hours (Ruback et al., 2015). In addition to training deficits, many agencies fail to include any hate crime policies to help guide officers in their investigations (Pezzella & Fetzer, 2021). The consequences of these missteps cannot be overstated. Often the motives for a crime are ambiguous, with competing motives being present that create confusion as to the true reason for the offender committing the crime (Nolan et al., 2004). Lantz and colleagues (2019) found hate crimes compared to non-bias crimes receive less attention from law enforcement officers with such cases not being investigated to the same extent as non-bias crimes. Only hate crimes matching preconceived notions, such as those containing severe violence, perpetrators affiliated with hate groups, and white offenders coupled with Black victims, received equivalent treatment. Importantly, Fetzer and Pezzella (2020) discovered that official reporting of hate crime statistics rely largely on the initial responding officer’s classification of the event as a hate crime. From 1995 to 2015, only 10% of all participating agencies to the UCR hate crime statistics program reported at least one hate crime that occurred within their jurisdictions despite victimization surveys indicating otherwise (Pezzella et al., 2019). However, this deficiency may also be due to the underreporting of hate crimes. Thus, effective training of law enforcement 28 officers regarding hate crimes is vital for officers to appropriately respond to hate crime, but such training is not mandated for many agencies. When police officers fail to accurately classify and treat these events as hate crimes, victims may be unable to receive the ameliorative services they need. Other marginalized members then lose trust in the police, believing that their victimizations are not taken seriously (Pezzella & Fetzer, 2021). Relationships between marginalized groups and the police are already contentious. Many members of minority groups report fearing the police (Graham et al., 2020), viewing them as illegitimate (April et al., 2023), and having more negative interactions than majority group members (Dennison & Finkeldey, 2021; Jackson et al., 2023). These attitudes towards the police and experiences with them ultimately reduce the likelihood of minority group members reporting their victimizations to the police in the future (Slocum, 2018). Community Programs Formal response from the criminal justice system is predicated on victims reporting hate crimes to the police. As noted, victims are often reluctant to report to the police for a variety of reasons, creating a large dark figure of hate crime. In light of this problem, community programs have also been developed to address the issues undergirding hate crimes that may prove more beneficial than the formal criminal justice system. These programs occur at a few levels. Some target communities at large to improve relationships between residents (Freilich & Chermak, 2013). Others target faculty and students in schools to minimize violence given that most perpetrators of hate crimes are adolescents or young adults (McDevitt et al., 2021). Fostering positive relationships between community members represents one pathway to minimizing the incidence of hate crimes. Nolan et al. (2020) found that rural areas with high levels of interdependence between community members had significantly fewer hate crimes than 29 rural areas with conflict and frustration between members. Programs such as Citizens at Heart aim to foster relationships between community members through educating community members on prejudice and hate crime (Zempi et al., 2021). During this program, individuals would discuss their thoughts on prejudice and stereotyping within their communities. Counter-narratives were provided by trained facilitators to challenge any negative stereotypes that were discussed. School programs against hate crimes offer similar training. The No Place for Hate program, sponsored by the Anti-Defamation League, follows a four-step approach to reducing hate in schools (Anti-Defamation League, 2021). In this program, students and faculty first form a committee to get a pulse on any hate or prejudice in the school. Next, the committee holds a school assembly where students take a pledge to make their school no place for hate. Researchers then conduct a climate survey to gain additional input on the presence of any hate. Last, students and faculty complete various training programs on stopping hate. These community programs are still in their infancy. Few outcome evaluations have been completed on their effectiveness in preventing hate crime. The evaluations completed to date have instead focused on shifts in community members’ attitudes towards hate crime. For example, Zempi and colleagues’ (2021) evaluation of the Citizens at Heart program revealed that participants were more likely to recognize hate after attending the program. Results from one dissertation found that only 12% of one school completed the No Place for Hate pledge (Blackwell, 2022). Moreover, members of the student-faculty committee perceived that various stakeholders invited to their meetings rarely participated. No Place for Hate in particular has faced pushback from conservatives. For example, a parent in Broward County Florida filed a complaint against this program under Florida’s Stop W.O.K.E. Law (Travis, 2023a). She claimed the program included elements of critical race 30 theory as well as information on sexual orientation and gender identity that the law prohibited. The Florida Department of Education ultimately chose not to investigate the claim (Travis, 2023b). Many of these programs contain information on effective bystander intervention techniques to stop hate crimes that were rarely evaluated. The following chapter will explain what bystander intervention is, methods for intervention in various parts of the hate continuum, and factors that influence bystander behavior. Summary This chapter examined the various legislative and law enforcement actions that have been taken to stop hate. Although several laws have been passed at the federal level pertaining to hate, state laws vary considerably in defining hate and the actions that can be taken to stop it. State laws, for example, vary in terms of which groups are protected in addition to what level of prejudice must have been displayed for hate crime enhancements to be pursued. Unsurprisingly, law enforcement’s response to hate has been hindered by the confusion stemming from these laws. These laws often do not include law enforcement training in detecting and responding to hate crime. Community-based efforts, however, have shown more promise in halting hate as citizens can be trained to spot and effectively intervene against hate, preventing it from taking root in their communities. 31 Chapter 4: Bystander Intervention The hate literature exhibits a common theme: the importance of combatting hate in all its forms to minimize harm done to individuals and communities. Hate crime scholars have noted bystander intervention as one important avenue for the mitigation and minimization of hate. Within multiple offender scenarios as common with the thrill and defensive hate offenders, McDevitt et al. (2010) described several roles in which offenders can be classified: “leaders, fellow travelers, unwilling participants, and heroes” (p. 134). The leader initiates the hate crime incident and may be the only offender present to physically assault the victim. Fellow travelers act based on the leaders’ actions and are unlikely to perpetrate hate crimes alone. Unwilling participants help neither the offenders nor the victims, often expressing reservations about the offense to their criminal peers after its completion. Heroes attempt to stop the other offenders by attempting to dissuade them or warning the victim to leave. The authors advised people may receive training to be heroes as a viable hate crime prevention method. Additionally, as thrill offenders operate on the belief that the community supports their actions, the presence of bystanders who intervene, also known as upstanders, may dispel these notions and prevent a hate event from escalating. Authors of a recent publication in the Annual Review of Criminology echoed this sentiment by imploring scholars to examine how hate crimes may be prevented outside of formal mechanisms (Farrell & Lockwood, 2023, p. 120). Given this recent call-to-arms by leading scholars, this chapter discusses the importance of bystander intervention in fighting hate. Bystander intervention is described for microaggressions, hate incidents, and hate crimes. Evaluations of bystander intervention programs aimed at each of these intervals of the hate crime continuum are provided. Factors that affect bystander behavior are then presented followed by latent class analyses of bystanders. 32 Bystander Intervention to Microaggressions, Hate Incidents, and Hate Crimes Given their previously described ubiquity, microaggressions represent ideal instances for bystander intervention. Witnesses to microaggressions may react in numerous ways. They can ignore it, harm the microaggressor, educate the microaggressor, stop the microaggression, provide emotional support to the target, find others to intervene, seek authorities, be an ally for the victim, or complete more than one of these possible actions (Sue et al., 2019). According to Sue et al. (2019), these microinterventions ultimately represent four separate goals: educating the offender, defusing the microaggression, rendering the invisible visible, and seeking outside help. Bystander intervention programs targeted against microaggressions have shown promise in a variety of disciplines and settings including health professions (Famouri et al., 2023), faculty in higher education (Haynes-Baratz et al., 2021), and undergraduate students (Banks et al., 2023). However, several barriers inhibit bystander intervention against microaggressions. For example, Xie and Galliher’s (2023) factorial vignette survey revealed witnesses and victims of microaggressions tend to perceive interveners as overreacting to the situation. Bystanders also may not perceive microaggressions as serious despite their deleterious effects and thus choose not to intervene (Mulvey et al., 2016). Verbal harassment during hate incidents may also be halted through bystander intervention. According to the organization Right to Be (n.d.), bystanders can intervene to stop harassment using the 5 Ds: distract, delegate, document, delay, and direct. Bystanders may distract the offender by engaging the victim in an unrelated matter. They can delegate by seeking help from others, especially authority figures like store managers or law enforcement officers. They may document the harassment by recording the events with their smartphone. They can delay intervention by speaking to the victim after the incident to ensure they do not require 33 assistance. Lastly, they can be direct by engaging with the offender and clearly telling them what they are doing is wrong. As with microaggressions, programs aimed at stopping verbal harassment in hate incidents have been successful in training upstanders. These include medical students (York et al., 2021) and public service workers (van Earp et al., 2018). However, survey respondents continue to report several factors that prevent them from intervening. These include their relationship to the victim, the perceived severity of the harassment, presence and actions of other bystanders, low bystander efficacy, perceived costs of intervention, and fear of worsening the incident (Davidovic et al., 2023). Interventions commonly prescribed to halt hate crimes include the same 5 Ds aimed at stopping harassment (Right to Be, n.d.). More specifically, given the criminal element present, contacting law enforcement is also advised. However, contacting the police is not always the ameliorant it is assumed to be. Wolff and Cokeley (2007) found numerous examples of unprofessional law enforcement and dispatcher behavior in response to reported hate crimes such as 911 operators choosing not to send officers to the scene and officers blaming the victim, mocking them, or laughing at the incidents. Moreover, many hate crimes victim choose not to report because of their perceived illegitimacy of the police and legal cynicism due to the over-policing of their communities (Grasso et al., 2023). Thus, other forms of bystander intervention are preferable. Limited research has been conducted on the efficacy of bystander intervention programs targeting hate crime. Zempi and colleagues (2021) evaluated the Citizens at Heart program which was designed to leverage community members against hate through several methods. These included increasing education on hate crimes, increasing hate crime reporting, altering 34 community beliefs and attitudes that enable hate crime, improving responses to hate crimes, and bolstering services for hate crime victims. Overall, citizens who attended the training reported increased feelings of bystander efficacy, indicating some potential for bystander response to hate crime, but the researchers did not test the effect of the program on local hate crime rates. Participants also stated they benefitted from learning what legally constitutes hate crime so they knew when to intervene, a common barrier within hate crime events. Few formal tests of bystander intervention models have been conducted with none existing for hate events. Nickerson and colleagues (2014) surveyed high school students’ reactions to witnessing bullying with items related to each stage of the situational model: noticing an emergency, considering it to be an emergency, feeling morally obligated to assist, generating options to intervene, then selecting an option to intervene and performing it. Structural equation modeling and confirmatory factor analysis of their data supported the model such that progression through one cognitive step increased the likelihood of proceeding through subsequent steps. The authors advised that their results supported the integration of this cognitive process model into bystander trainings. These findings have been supported for interventions with bullying in middle school students (Jenkins & Nickerson, 2019) as well as with teen dating violence (Casey et al., 2017). Moreover, the path that bystanders take through the situational model is racially invariant (Jenkins et al., 2023). However, research must be completed to examine whether these findings extend to hate crimes. Hate crimes occur between strangers usually without law enforcement nearby, whereas bullying in schools happens between peers with proximal authority figures, like school resource officers, who are capable of intervening (Kena & Thompson, 2021). As exemplified, research has demonstrated the power, or potential, for bystander 35 intervention to reduce all forms of hate. However, studies of bystander intervention to hate have tended to focus on each specific point on the hate continuum rather than on all. Additionally, there are distinct barriers for intervening at each point that may differentially impact whether bystanders choose to intervene. Therefore, the studies in this dissertation will examine how bystanders would respond to all forms of hate at the individual level (microaggression, hate incident, and hate crime) in one escalating hate event. These studies also examined single instances of one emergency as opposed to an escalating event such as those found in hate crimes. Therefore, the heterogeneity in bystander intervention was simplified to progression through the situational model only once. To uncover the numerous combinations of bystander progression between microaggression, hate incident, and hate crime, this dissertation will utilize latent class analysis to first classify subjects based upon their progression in the situational model across all three phases of hate. These latent classes will then serve as the dependent variable for the remaining analyses. The bystander intervention model will thus be tested using structural equation modeling and latent class analysis of bystanders’ cognitive processing of the situation. This leads to the first research question: Research Question 1: Does bystander progression through the situational model increase concomitantly with increasing severity of the hate scenario? Factors Conducive to Intervening Victim Variables Victim characteristics have also been found to influence bystander behavior. The extant literature has primarily focused on victim race, but a few studies have also examined the effects of sexual orientation and gender identity on bystanders. These studies often frame bystander 36 intentions as differing based upon victim group membership (i.e., in-group versus out-group). Gaertner and colleagues (1982) examined how long and to what extent subjects would assist a confederate they observed on closed-circuit television from an adjoining room. Subjects could see whether the victim, who had a tower of chairs fall on them, was Black or white. Indeed, subjects responded more quickly and more frequently to white victims than Black victims. The researchers hypothesized the discrepancy may have been due to subjects being more empathic to distressed members of their own group (i.e., white) than members of the outgroup. Such results have been replicated for sexual assault scenarios in college students (Katz et al., 2017) and fall injuries (Kunstman & Plant, 2008). Studies examining the role of target sexual orientation in bystander intervention have found similar results. Per Owuamalam and Matos (2020), both heterosexual men and women are less likely to report upstander behaviors for hate crimes with gay victims. However, bystander intervention by heterosexual men and women is apt to increase when they have been primed for social evaluation by being asked to rate how angry or calm in-group heterosexuals and out-group gay men are. By considering answers to those questions, the judgment of the bystander shifts by first thinking of gay men in either a positive (calm) or negative (angry) light. Moreover, heterosexual men demonstrate less compassion to gay victims of hate crime when given information on the masculinity of the target (Owuamalam & Matos, 2022). Interestingly, heterosexual men displayed more compassion towards feminine gay male victims than masculine gay male victims. Bystanders also intercede more often in cases of opposite-sex intimate partner violence than same-sex intimate partner violence (Graham et al., 2023). Few studies have examined how gender identity, including transgender individuals, influences bystander behavior. Godzisz and Mazurczak (2023) noted that transgender victims of 37 hate crimes receive the least empathy from witnesses compared to both heterosexual and sexual minority individuals. Additionally, cisgender men report less intent to intercede in cases of male violence against women when the use of a trans slur identified the woman as transgender (Deacon et al., 2020). Similarly, teachers are less willing to help trans victims of bullying except when they report strong positive feelings towards trans individuals (Parker et al., 2023). Limited research has examined how victim race, sexual orientation, and gender identity compete with one another when bystanders decide to intervene. Vera and colleagues (2023) found college students were significantly more likely to display upstander behaviors to bullying when the victim was a racial minority. They found no significant effects for when the victim was being bullied for their sexual orientation or gender identity. They postulated that the increased attention to race, especially the emphasis colleges have put on antiracism, caused students to exhibit more upstanding behaviors for racial minority victims. Other researchers have also decried the lack of education on bystander intervention for sexual and gender minorities (Kirk-Provencher et al., 2023). Overall, these studies highlight that bystander willingness to intervene diminishes when victims are from minority groups. This occurs for racial, sexual, and gender minority victims. Moreover, there appears to be a disparity based upon Vera and colleagues’ (2023) study such that bystanders are more inclined to aid racial minority individuals than sexual and gender minority individuals. Their study, however, occurred within the context of bullying in college settings. Therefore, it is still unknown how the general population would react to observing hate crimes. This leads to research question 2: Research Question 2: Does the decision-making of bystanders to hate crimes differ based upon whether the target is a racial, sexual, or gender minority victim? 38 Bystander Variables Bystander intervention researchers have noted several individual-level traits as increasing the likelihood of intervention. These primarily include empathy, bystander efficacy, and decision-making as to the costs and benefits of intervention. Empathy is a multidimensional construct consisting primarily of affective empathy and cognitive empathy, although some researchers have posited it also includes emotional contagion which is when an individual mimics the emotional experience of another(Carré et al., 2013). Affective empathy refers to an individual’s ability to have emotional concern for others whereas cognitive empathy refers to being able to identify another’s perspective (Gini et al., 2007). A meta-analysis conducted by Nickerson and colleagues (2015) revealed that empathy significantly predicted upstanding in response to bullying. They noted specifically that both affective empathy through the recognition of the victim’s suffering, as well as cognitive empathy through considering the victim’s perspective, created the desire to help. However, affective empathy is most responsible for compelling bystanders to act whereas cognitive empathy helps them to recognize and know how to intervene (Secord Fredrick et al., 2020). Bystander efficacy has been shown to positively predict bystander intervention across a host of situations from general emergencies (Latané & Nida, 1981) to sexual violence (Moynihan et al., 2011) and robberies (Huston et al., 1981). Diffusion of responsibility to other bystanders is also less likely to occur when the bystander in question has strong feelings of self-efficacy in their own ability to intervene (Fischer et al., 2011). Several studies by Banyard have also shown that bystander efficacy significantly predicts both intentions to intervene and reported actions in real scenarios (Banyard, 2008; Banyard & Moynihan, 2011; Banyard et al., 2004). McMahon et al. (2015) discovered through longitudinal surveys of college students that there is a reciprocal 39 relationship between bystander efficacy and intentions to intervene, indicating that these two co-evolve over time. Bystander decision-making has been studied in terms of how bystanders weigh the costs and benefits of intervention. The bystander intervention literature refers to this as bystander decisional balance. Upstanders would consider the benefits to outweigh the costs of intervention, thus deciding to intervene. Conversely, non-interveners would consider the costs to outweigh the benefits (Shea et al., 2021). The factors that bystanders weigh include relational and personal costs and benefits of intervening as opposed to doing nothing (Jensen & Raver, 2020). Banyard and Moynihan (2011) found that decisional balance in favor of intervention is an enduring individual trait that promotes intervention. Social psychologists previously fleshed out bystander decisional balance in terms of self-interest (Stroebe & Frey, 1982). Morgan (1978) noted decisional factors included the net benefit to the individual and the net benefit to the group that would come from intervening. Recent criminological research has also noted the impacts of the levels of self-interest versus other-regarding an individual has on decision-making, namely that the self-interested are deterred by both the certainty and severity of punishment while other-regarding individuals are deterred only by the severity of punishment (Paternoster et al., 2017). Thus, the evidence suggests that self-interest and other regarding may also differentiate bystander behavior. Bystander race may also impact bystander behavior, especially in instances where minority status is salient. Ratcliff and colleagues (2023) found that white bystanders were significantly less likely to intervene than Asian American bystanders in instances of anti-Asian prejudice. This difference is partially due to whites being less likely to notice prejudice. Whites often fail to recognize prejudice as they have not been the targets of it (Ashburn-Nardo et al., 40 2008). Even when they do, whites also tend to attribute less prejudice to the event than racial minority individuals (Czopp, 2010). Bystander gender also affects proclivity to intervene. Women are significantly more likely to intervene when witnessing sexual violence scenarios (Mainwaring et al., 2023), workplace incivility (Sinclair, 2021), and non-emergency situations (Cox & Adam, 2018). However, men are significantly more likely to intervene when witnessing violent incidents in public (Liebst et al., 2019). Thus, gender differences have been observed within bystander intervention. Some empirical evidence suggests that bystander sexual orientation and gender identity can influence willingness to intervene. For example, sexual minority individuals were significantly more likely to intervene when listening to audio vignettes of intimate partner violence in both same- and opposite-sex cases of intimate partner violence (Graham et al., 2023). Trans women and gender non-conforming individuals have reported seeing more opportunities to intervene in sexual assault situations, but are significantly less likely to do so than cisgender women (Hoxmeier et al., 2022). Cisgender gay men, who are gay men who continue to identify with their assigned sex at birth (i.e., male) also reported having intervened significantly more frequently in sexual assault situations than cisgender heterosexual men (Hoxmeier et al., 2022). Intersectionality in bystander intervention studies represents a considerable dearth within the bystander intervention literature. Few studies have examined how race and gender interact to impact bystander behavior or willingness to intervene. Burns and colleagues (2019) explored how race and gender impact intent to intervene in college campus sexual assault situations. Before being subjected to a bystander intervention program, white men were least likely to intervene when compared to Latino men, Latina women, Black men, Black women, and white 41 women. White men, however, had the highest gains postintervention with a mean of 2.41 points on the bystander intervention scale. Moreover, the processes impacting bystander intervention also differentially affect groups. Black men, for example, report more instances of intervening when they have peer norms conducive to intervening while this relationship did not hold for white men, white women, or Black women (Brown et al., 2014). Given the above, cisgender, heterosexual, non-Hispanic white men will serve as the reference group when examining how bystander characteristics impact progression through the situation model in these studies. Research Question 3: Do bystander characteristics influence their bystander decision-making processes while witnessing hate? Summary This chapter contained information on bystander intervention as a mechanism for stopping hate. In particular, the effectiveness of programs aimed at stopping microaggressions, hate incidents, and hate crimes were discussed. Factors influencing bystander decision-making such as victim characteristics and traits of the bystanders themselves were examined. Three research questions arose from these examinations of the literature that will be answered in the three studies of this dissertation. 42 Chapter 5: Trends in Public Opinions of Minority Groups This chapter examines public opinion trends towards Asian Americans, sexual minority individuals, and gender minority individuals within the U.S. Public opinion trends factor into hate crime and bystander intervention in a few key ways. First, public sentiment towards perceived outgroups has been correlated with the number of hate crimes perpetrated against outgroup members (Gordon, 2020). Second, public opinion towards outgroups influences the level of blameworthiness attributed to victims from these outgroups (Erentzen et al., 2021). Although victim blameworthiness is beyond the scope of this dissertation, it nevertheless is an important mechanism underlying the decision to intervene (Katz et al., 2015). Thus, it is necessary for readers to acknowledge public opinions towards the groups that will be studied by this dissertation as the current trends engender increased concern for these selected minority groups. As expected based upon the aforementioned hate crime trends, public opinions toward most racial and religious minorities have mostly experienced little change or even become more positive (Brooks & Harmon, 2022). For example, public sentiment towards Black Americans has largely remained stable for the past decade (DeSante & Smith, 2020). Sentiment around immigrants has shown slight fluctuations based upon the politics of immigration but has not demonstrated any significant changes (Sagir & Mockabee, 2023). Public opinion towards Jews has increased in the past several decades such that nearly all groups report a moderate increase of warmth on feeling thermometers, which ask respondents to rate how positively (i.e. warmly) they feel towards a group (Cohen, 2018). However, young adults who endorse far-right political views have shown increases in antisemitic attitudes (Hersh & Royden, 2023). Attitudes towards Muslims have likewise increased since experiencing fluctuations post-9/11 (Griffin et al., 2021). 43 However, public opinion towards Asian Americans, gays and lesbians, and transgender individuals have changed immensely in recent years and typically for the worse as will be explored below. Public Opinions of Asian Americans Opinions of Asian Americans have typically fallen within the view that they constitute a “model minority.” Peterson (1966) generated the myth of the model minority by using the term to praise Japanese Americans while denigrating other racial and ethnic minorities. The stereotypes that emerged from this characterization include views that Asian Americans excel in academics, particularly math and science, experience great deals of upward mobility, hold high status or high-income jobs, and rarely commit crimes (Wong & Halgin, 2006). Analyses of public opinion data have commonly shown that these views persisted in subsequent decades (Marsden et al., 2020). Some scholars have found that the American public tends to view Asian Americans so highly that they blend into their opinions of whites (Kim, 2009; Song, 2021). This has led to some Americans considering Asian Americans, particularly those of East Asian (e.g., Chinese, Japanese, Korean, etc.) and South Asian heritage (e.g., Indian, Nepalese, etc.), to be “honorary whites” who come close to achieving the status of whites in the racial hierarchy (Bonilla-Silva, 2004). These findings are supported by courtroom sentencing studies showing that there are minimal differences in sentencing between Asian Americans and non-Hispanic whites (Franklin & Fearn, 2015; Johnson & Betsinger, 2009). Gans (2012) even predicted that Asian Americans with roots from East Asia would be, for all intents and purposes, deemed white by 2050. However, not all public sentiments towards Asian Americans have been favorable. The positive traits stereotypically applied to Asian Americans sometimes act as the impetus for 44 harassment in work environments or hate crimes (Kim, 2007). Moreover, some studies have highlighted the view that Asian Americans are “forever foreigners” who will never assimilate into American culture (Okihiro, 2014; Xu & Lee, 2013). Media narratives that propagate this view often do so in the context of red scare by emphasizing the threat that communist countries in Asia such as China pose to the U.S. (Del Visco, 2019). Such discourses question the loyalty of Asian Americans, thereby perpetuating the notion that they cannot fully assimilate into American culture. More severe than the “forever foreigners” perception of Asian Americans is the “yellow peril” perspective. This negative stereotype, fostered in the mid-to-late 19th century, portrays Asian Americans as a public health threat in addition to being a foreign threat (Lee, 2007). At its height, the yellow peril stereotype culminated in the 1882 Chinese Exclusion Act that prevented Chinese immigrants from coming to the United States. Although this prejudicial view eventually ebbed with the rise of the model minority stereotype, the “yellow peril” view made a resurgence because of the genesis and spread of the COVID-19 virus in China (Li & Nicholson, 2021; Reny & Barreto, 2022; Wu et al., 2023). For example, one longitudinal study revealed that becoming sick with COVID-19 significantly increased the level of disgust expressed towards Asian Americans in the next month, which then significantly increased anti-Asian prejudice in the following month (Wu et al., 2023). Likewise, Nam et al. (2022) used data from three separate nationally representative surveys that all found anti-Asian attitudes increased after the start of the pandemic while attitudes towards other racial and ethnic minorities remained stable. Part of the uptick of this view rose from then-President Trump and other prominent political figures referring to COVID-19 as the “Chinese virus” and 45 “kung flu” (Rubin & Wilson, 2021). These terms exoticized the pathogen and placed the blame on Chinese people for its existence. Public Opinions of Sexual Minority Individuals Unlike public opinions of Asian Americans, public opinions of sexual minority individuals have long tended to be negative. Yang (1997) noted various trends in U.S. public opinion from the 1970s to the early 1990s regarding sexual minority individuals. During that time period, approximately 70% of U.S. adults claimed that sex between two members of the same sex was morally wrong. Although considered morally wrong, nearly two-thirds of adults nevertheless believed that same-sex sexual relations occurring between adults in private should be legal. Public opinion was evenly divided as to whether sexual minority individuals chose their sexual orientations. Over time, a greater number of respondents reported knowing someone who identified as gay, lesbian, or bisexual. Regardless, feeling thermometer ratings displayed that gay men and lesbians were among the lowest, if not the lowest, social group in the U.S. in terms of whether respondents felt favorably towards them up to 1994. Because sexuality is subject to moral judgment, opinions are less likely to change compared to other types of issues (Olson et al., 2006). Indeed, this observation seems to be supported by studies of public opinion trends over time (Brewer & Wilcox, 2005; Yang, 1997). Newer issues such as same-sex marriage, however, invite the opportunity for the creation of new moral judgments and gay rights that are more susceptible to change as they are not yet ingrained in individuals. As such, there has been growing public support for gay and lesbian equality in the U.S. since the 1990s (Keleher & Smith, 2012). The increase in acceptance of gay men and lesbians has also been attributed to increasing levels of higher education attainment (Budge, 2023). 46 Despite these noted increases in positive perceptions of sexual minority individuals, there has been a recent reversal largely fueled by anti-LGBTQ legislation (Hartman, 2023; Oakley, 2022). An annual Gallup poll showed a seven percent decrease from 2022 to 2023 in the number of respondents who believed homosexuality was morally permissible, dropping from 71% to 64% (Jones, 2023b). Republicans largely drove this decrease as 56% viewed homosexuality as morally acceptable in 2022 compared to just 41% in 2023. Likewise, the percentage of Democrats reporting this view decreased from 85% to 79% in the same period while independents increased from 72% to 73%. Thus, public sentiment towards sexual minority individuals appears to be declining. Public Opinions of Gender Minority Individuals Few studies have been completed on the public’s views towards transgender individuals. That is largely due to the invisibility of transness until the late 1990s with the call for more empirical research regarding trans issues (Stryker, 1998). The absence of attention to trans people signifies the discomfort and other negative feelings most of the public felt towards them, largely fueled by the medicalization and consequent pathologization of transgender people in the late 19th and early 20th centuries (Beemyn, 2013). Per Beemyn (2013), movements within the medical community aimed at accepting transgender people as natural and not abnormal did not occur until the 1960s and 1970s. Even then, opposing medical practitioners continued to espouse anti-trans views that mired the advancement of trans medical rights. Most public opinion studies of views on transgender people and transgender rights have been conducted within the past decade. Respondents to a national probability sample in the U.S. reported an average feeling thermometer rating of transgender people as 32 out of 100 (Norton & Herek, 2013). This rating indicated that respondents viewed transgender people strongly 47 negatively and did not feel comfortable around them. Political conservatism and religiosity significantly predicted low ratings on the feeling thermometer. Later studies have replicated these findings that the U.S. public tends to perceive transgender individuals negatively, even if they agree with pro-trans policies (e.g., the majority supports protection for trans people from workplace discrimination but still views trans people as less trustworthy than cisgender people) (Tadlock et al., 2017; Jones et al., 2018). Attitudinal differences have been observed between situations with transgender men versus those with transgender women (Worthen, 2013; 2016). For example, parents of female-to-male transgender youths are often more accepting of their children than parents of male-to-female transgender youths (Grossman et al., 2005). As with views on sexual orientation, the 2023 Gallup survey revealed decreased support for trans people in several areas from 2021 to 2023 (Jones, 2023a). For instance, approximately 34% of respondents in 2021 believed that athletes should be able to play on teams that align with their gender identity compared to 26% in 2023. Moreover, 46% of all respondents believed undergoing gender confirmation surgery was morally acceptable in 2021, which decreased to 43% in 2023. As with the moral acceptability of sexual minority individuals, Republicans were most responsible for the decline in this view, going from 22% in 2021 to 15% in 2023, whereas Democrats demonstrated an increase, 67% in 2021 to 70% in 2023. Overall, public sentiment towards trans individuals, both in terms of where they can participate with their gender identity, as well as their right to exist as their true selves, has been becoming decreasingly favorable within the U.S. Summary In summary, there have been significant, negative changes in public opinion towards Asian Americans, sexual minority individuals, and transgender individuals just within the past 48 few years. Such changes are important for two reasons. First, the increased animus towards these groups may propel increasing numbers of hate crimes against members of these groups. Second, negative attitudes and feelings of disgust may also inhibit strangers from helping. The following chapter contains the theoretical overview of this dissertation to contextualize how these findings interact with theory and the implications for these studies. 49 Chapter 6: Theoretical Overview Bystander Intervention Models Originating in the late 1960s, the concept of “bystander intervention” first gained national prominence through the work of Latané and Darley (1970), who asserted witnesses must cognitively process an emergency in five steps for intervention to occur. They must 1) observe the situation, 2) perceive it as dangerous, 3) feel obligated to help, 4) be aware of options taken to disrupt it, and finally 5) choose to act. This model has been criticized for being a universal template that explains bystander behavior in isolated incidents involving strangers but not more nuanced emergencies (Banyard, 2015). With the advent of bystander intervention as a process, scholars later developed three other models for bystander intervention: the arousal cost reward model, the model of moral courage, and the Action Coils model. The arousal cost reward model developed by Piliavin and colleagues (1981) explains bystander intervention in terms of the emotional impact the incident has on the bystander. When witnessing another’s plight, an individual becomes psychologically aroused and then intercedes to reduce that arousal. This model specifies that bystanders use a decision-making calculus by weighing the costs and benefits of assisting and then take (in)action based upon the smallest net cost. Batson (1991; 1997) criticized the arousal cost reward model for not specifying what constitutes psychological arousal: fear for one’s own safety, empathic concern for the imperiled, or something else. Moreover, Otten et al. (1998) demonstrated that bystander intervention occurs even when there is no psychological arousal. The model also predicates itself on the notion that individuals will not intervene when there are high costs for both the victim and the bystander 50 (Dovidio, 1984). This tenet of the model, however, does not have empirical support as studies have shown individuals are more likely to intervene the more severe a situation is despite perceived risk to the bystander (Bennett & Banyard, 2016; Chabot et al., 2009; Huang et al., 2023). Similar to the arousal cost reward model, the model of moral courage places emotional arousal on the crux of helping behavior (Osswald et al., 2010). Per this model, bystanders will intervene when they become angered by witnessing injustice that fuels them to be morally courageous. Moral courage requires bystanders to understand that intervening may harm them, the offenders are intimidating, and intervention represents the greater good. Whereas positive affect increases the likelihood of intervening under other models, righteous anger, strong senses of justice, and having previously observed others acting in morally courageous ways are predictors of the moral courage model. These requirements ultimately constrain bystander intervention as bystanders must have strong moral foundations to consider helping after witnessing what they subjectively consider to be a grave injustice. Additionally, situations that require moral courage to intervene tend to occur quickly and typically involve physical danger (Osswald et al., 2010). Enacting informal social control by addressing incivilities (for example, microaggressions) does not constitute moral courage. The model of moral courage therefore cannot be used within the context of these current three studies as it cannot be used to explain bystander responses to either microaggressions or hate incidents. More recently, the Action Coils model created by Banyard (2015) explores bystander intervention specifically in response to sexual assault and intimate partner violence. Unlike the situational model, this frame heavily emphasizes the relationships of the bystander to the victim 51 and the offender. Other considerations such as peer norms, the bystander’s social standing, and the effects that intervention will have on the bystander’s relationships with the victim and the offender consequently gain prominence. Research has validated the use of the Action Coils model within these contexts of sexual assault and intimate partner violence (Banyard et al., 2021; Moschella & Banyard, 2021). The Action Coils model, however, has yet to be used to examine bystander intervention outside of sexual and dating violence contexts. The situational model developed by Latané and Darley (1970) thus appears to be the best model for exploring bystander intervention in hate crimes for a few reasons. First, hate crimes often occur between strangers in one-time incidents, which are the crises that the situational model was designed to explain. Unlike the requirements of Osswald and colleagues’ (2010) model of moral courage, these events do not have to unfold rapidly or necessarily have physical danger present. Second, it does not make any assumptions about the moral character of the bystander and what actions that morality would cause the bystander to take. Third, it emphasizes the most proximal consequences of intervention such as embarrassment, as opposed to more distal outcomes like social standing among peers, which is more relevant to other forms of interpersonal violence such as intimate partner violence and sexual assault. Fourth, the situational model can be used to as a framework for bystander intervention even in situations that are not physically dangerous. Thus, the situational model will be used to explore bystander decision-making in hate situations. Overview of Latané and Darley’s Situational Model Per Latané and Darley (1970), bystander intervention occurs as a cognitive process involving five steps. Bystanders must first detect the problem. Noticing that something is wrong requires onlookers to be attentive to their environment and not overstimulated. They need contextual familiarity (i.e., norms of behaviors and values) to know that a transgression is 52 occurring. Within hate crime contexts, onlookers often fail to detect microaggressions against racial/ethnic, sexual, and religious minorities (Lilienfeld, 2017), which is especially important as insults escalate into violence (Whitehead et al., 2018). Second, they have to perceive the situation as an emergency. Individuals reach this conclusion in ambiguous scenarios based upon cues in the environment, most often in the form of social cues from nearby bystanders (Latané & Darley, 1970). If most other individuals who are present fail to notice the incident or seem unbothered by it (pluralistic ignorance), even actively telling others to ignore what is happening, then the onlooker is significantly less likely to deem it an emergency (Latané & Darley, 1970, p. 48). Sexual (Owuamalam & Matos, 2020) and racial minority victims (Gaertner et al., 1982) have been shown to be most minimized, as their victimizations are deemed less critical than those of majority group victims such as heterosexuals and whites. Individual level traits, such as personality, mood, and the self-persuasiveness of the bystander, also influence perceptions of emergencies. For instance, when witnessing the beginnings of hate crimes, individuals may perceive prejudicial statements as jokes rather than emergencies (Katz et al., 2019a, 2019b, 2022). Moreover, even individuals who consider themselves to be socially liberal fail to recognize microaggressions (McClure, 2020). Third, bystanders must feel morally obligated to assist. The victim’s worthiness of receiving aid, the victim-bystander relationship (for example, stranger versus acquaintance), the bystander’s sense of efficacy, and diffusion of responsibility to other onlookers all coalesce to inform the bystander’s sense of personal responsibility. Importantly, demographic characteristics of the victim (gender identity, race, sexual orientation, etc.), along with victim blameworthiness, signal to the bystander whether the victim deserves assistance. Indeed, majority group members respond less quickly to emergencies with racial minority victims (Gamberini et al., 2015; 53 Kunstman & Plant, 2008) and have less empathy for sexual minority victims, including transgender victims (Mazurczak & Godzisz, 2019). They also consider Jewish victims to have greater blameworthiness (Imhoff & Banse, 2009). Bystanders are most likely to intervene when the victims appear similar to themselves due to in-group identification (Levine et al., 2002). This could be exacerbated in hate crime situations where the public and collective nature of these events makes individual traits more salient. Bystanders also quickly evaluate their defusing effectiveness. If they consider themselves unable to help, they may decide the responsibility belongs to others who are better equipped to manage the emergency (Darley & Latané, 1968). Fourth, bystanders have to generate options to respond to the emergency. To do so, bystanders must develop choice sets based upon all options they know can be undertaken to assist. Per Gettys and colleagues (1987), individuals retrieve various options in response to a situation and then choose from among them to act. For that to occur, people must first 1) know what courses of action are available, and 2) be able to quickly recall those options when the situation requires. Bystander intervention contains considerable choice heterogeneity to two types of interventions: direct or detour interventions (Latané & Darley, 1970). Direct interventions involve bystanders placing themselves into the emergency to halt it, such as shouting at the offender or physically moving in between the bystander and the victim. In contrast, detour interventions happen indirectly, where bystanders report the emergency to a relevant authority who can stop it. The former requires knowledge and skill on the part of the bystander to be effective, whereas the latter frequently requires a 911 phone call or awareness of a capable guardian onsite (Latané & Darley, 1970). As most bias incidents do not initially rise to the level of requiring police intervention (e.g., starting with prejudicial remarks), bystander efficacy 54 becomes especially necessary for involvement to occur (Schafer & Navarro, 2003). Lastly, Latané and Darley (1970) stated the onlooker decides to implement the intervention chosen from the fourth step. This final step is potentially the least difficult, as the above decisions have already been made. However, there are instances wherein the stress of intervening during an emergency may diminish the bystander’s perceived effectiveness (Latané & Darley, 1970). For example, bystanders may worry they are overreacting to the perceived hate crime, especially in the early verbal stages, and thus choose not to act for fear of embarrassing themselves in front of others (Czopp, 2019). Despite the ease with which hate crimes can be integrated within the bystander decision-making model, bystander intervention scholars continue to decry the field for failing to study the barriers to each step more extensively (Mazar, 2019). No research to date has fully explored how bystander progression through the situational model would change within the context of an increasingly dire hate event. Given the aforementioned likelihood of individuals being oblivious to microaggressions, preventing them from initiating bystander decision-making, and the increased likelihood of intervening when physical violence is present, I hypothesize the following: Hypothesis 1: Respondents will go further in the situational model’s decision-making steps as the hate event progresses and becomes increasingly severe. Implicit Bias One concept that may explain bystanders’ decision-making varying by victim group membership is implicit bias. Devine (1989) described implicit bias two decades after Latané and Darley developed their model of bystander intervention. Implicit bias refers to the assumptions an individual makes about members of various groups. These assumptions are learned through 55 socialization and become activated automatically when an individual encounters someone and classifies that person into the groups in which they appear to belong. The individual observing them then projects characteristics onto them based upon learned stereotypes of their groups. Despite being considered an automatic process, implicit bias can be primed by different cues. For example, Dasgupta and Greenwald (2001) found that implicit bias could be diminished when telling individuals to think of positive members of outgroups and negative members of ingroups. Unlike explicit bias, which is more overt, the actions an individual takes when operating under implicit bias are more covert and subtle. More specifically, explicit bias is intentional and easily communicated to others whereas implicit bias is unintentional and requires reflection for someone to become aware of it (Dovidio & Gaertner, 2010). Implicit bias affects a person’s actions despite being unintentional as latent thoughts and feelings are antecedents to perceptions which then inform behaviors (Greenwald & Krieger, 2006). Although Devine (1989) conceived of implicit bias as a mental latent construct that operates automatically in the background of one’s mind, De Houwer (2019) proposed that implicit bias can instead be seen as a behavioral phenomenon. Accordingly, implicit bias then is the “implicit group-based behavior, which is behavior that is influenced in an implicit manner by cues that function as an indicator of the social group to which others belong” (De Houwer, 2019, p. 836). Per De Houwer (2019), feelings and thoughts towards groups may also be considered to be covert forms of behavior. This view reinforces that implicit bias can be changed through not only metacognition, but also awareness of one’s own behavior. One study has directly examined the role of implicit bias in bystander intervention. Carson and Politte (2021) discovered bystanders were significantly more likely to report suspicious activity related to terrorism when the person performing the activity appeared to be 56 Middle Eastern as opposed to white. They explained this behavior was due to the respondents’ implicit bias manifesting as associating Middle Eastern males with terrorism that increased reporting to the police. Thus, implicit bias may operate within the bystander decision-making process to influence the willingness to intervene. Heuristics Other cognitive processes complementary to implicit bias are heuristics. Heuristics are mental shortcuts that people use when they do not have complete information about someone or something based upon into what group the object or person can be classified (Newell et al., 1957). Individuals use heuristics as a form of satisficing, which is when an optimal solution cannot be found and instead individuals estimate or make assumptions based upon incomplete information that provide close approximations to the target goal (Simon, 1956). One particular heuristic that impacts the evaluation and judgment of others is the representativeness heuristic (Kahneman & Tversky, 1972). Per Kahneman and Tversky (1972), the representativeness heuristic occurs when an individual relates a process being observed to its prototypical parent group through evaluating how well the properties of the process reflect the salient features of the parent group. The representativeness heuristic applies not only to events and objects, but to people as well wherein people judge others’ group membership based upon whether those others contain traits similar to those of said groups (Tversky & Kahneman, 1974). The representativeness heuristic can easily lead to prejudicial or discriminatory behavior, explicit or implicit, as individuals apply stereotypes to others based upon their perceived group membership. These effects can be observed in a variety of settings. For example, one experimental study found patients who appeared to be gay were more likely to be suspected of having HIV or AIDS despite reporting the same symptoms as patients who appeared to be 57 straight (Triplet, 1992). Respondents in one survey administered via Mechanical Turk were significantly more likely to attribute mental illness to transgender people than to gays, lesbians, and heterosexual men and women (Locantore & Wasarhaley, 2020). Asian Americans, conversely, were often associated with stability, white collar employment, and scientific careers (Ash et al., 2022). A defendant’s race can impact both judicial and juror decision-making such that individuals relate Blackness with criminality and are therefore more punitive toward Black defendants (Burns, 2016; Guthrie et al., 2001). Importantly, these representativeness heuristics can also impact one’s perceptions of interpersonal risk, which is salient for the act of bystander intervention as risk is intrinsic to helping others during conflict (Waters et al., 2023). As exemplified, the representativeness heuristic impacts people’s judgments of others, which may therefore change perceptions that would lead to intervention. Related to the representativeness heuristic is the availability heuristic. In the availability heuristic, individuals consider an event to be more probable when it is more easily accessed in their minds (Tversky & Kahneman, 1973). This heuristic can operate in a few ways for bystander intervention to hate. First, bystanders may not realize they have witnessed discrimination or prejudice if they themselves have not been subjected to it or trained to see it (Moroz & Campbell, 2018). This lack of accessibility regarding hate can affect how bystanders interpret the situation and their appraisals of why the event is occurring. If they fail to see the undercurrent of animus, they may assume the offender is being rude as opposed to being prejudicial, minimizing the significance of the event and diminishing their likelihood of intervening. Second, both the rhetoric against Asians for the COVID-19 pandemic, as well as the increasing anti-LGBT legislation mired with homophobic and transphobic stereotypes, provide individuals with an influx of negative information on these groups. This information is then easily accessible when 58 encountering members of said groups, thereby influencing perceptions and behavior towards them (Carbado & Roithmayr, 2014). Within bystander intervention scenarios, bystanders may then attribute greater blame to victims from these groups, believing the harassment to be deserved, and choose not to act. While neither implicit bias nor these heuristics are directly tested by the studies of this dissertation, they nevertheless represent mechanisms undergirding how and why bystanders may progress through the situational model differently based on the victim’s identity. Heuristics in particular are the decision-making concepts associated with differential treatment of individuals based on their group membership. A theoretical framework, social categorization theory, follows next that better explores why victim and bystander identity might impact bystander decision-making. Social Categorization Theory Although Latané and Darley’s model can explain bystander intervention overall, it falls short at explaining subtle nuances. One such nuance is why someone might intervene for a victim of one group but not for a victim of a different group. One theory capable of explaining the disparity in bystander intervention between victim groups is social categorization theory, also known as self-categorization theory. Turner and colleagues (1987) developed this theory, positing that there are three levels in which individuals categorize themselves. These levels consist of human identity, social identity, and personal identity. Human identity refers to a person’s self-identification as a human being and member of humanity. In the second level, social identity, individuals consider themselves members of a unique social group opposite of other social groups. Lastly, someone’s personal identity is how that individual identifies as a member of that social group in relation to other 59 members of the same group. Of note, as one of these levels becomes more salient in a given context, the other levels then recede and lose prominence. Accessibility and fit enable categorization of the self and others (Oakes, 1987; Oakes et al., 1991). Categories become more accessible depending on situational factors that prime someone to choose the most relevant categorization. The accessibility of categories depends also on whether the individual has any motivation to use one category over others, as well as the frequency in which the category is routinely activated. Fit may be either comparative or normative. Comparative fit entails the individual choosing a category that maximizes differences with other categories while minimizing differences between members within a category. Normative fit occurs when an individual perceives that someone’s behavior aligns with stereotypes about a certain group. Once self-categorized into a social group, depersonalization then occurs (Turner et al., 1987). In other words, people perceive stereotypic patterns between category members and begin to see members as interchangeable, no longer seeing themselves as unique individuals. These patterns include the behaviors, attitudes, and emotions attributed to the group. Individuals then begin to act in accordance with what they believe other group members would do in a given situation. Such perceptions start in early childhood and continue into adulthood (Rutland et al., 2010). Individuals tend to favor other members of their own group over members of outgroups in a phenomenon known as ingroup favoritism (Mullen et al., 1992). Conversely, individuals also treat members of outgroups negatively, which is known as outgroup derogation (Brewer, 1999). These two concepts have previously been considered to be intertwined, but according to Brewer (1999), these are mutually exclusive attitudes. In other words, people can think positively 60 of their own groups without thinking negatively of outgroups. However, both cognitions inevitably lead to discrimination or prejudice whether the disparity is due to treating the ingroup more favorably or the outgroup more unfavorably. Brewer (1999) further postulated that as ingroups become larger, so does the depersonalization of its members. The values of the bourgeoning ingroup become the moral authority, often absolute and deemed better than the morals of the outgroup. The extent of the difference in morals between the ingroup and outgroup breeds contempt. One of two outcomes typically result in these instances. Ingroup members may segregate themselves from the outgroup and avoid them. Ingroup members may also turn to genocide to remove those they consider immoral. Most individuals choose the former as feelings of contempt and disgust more commonly lead to avoidance as opposed to forms of interpersonal violence. Within the context of hate crime and bystander intervention, one would assume that ingroup members would avoid intervening as a display of their disapproval of the victim’s minority status. Level of intervention would depend upon which minority group to which the victim belongs. As to the three broad categories of victims being examined in this proposal, racial, sexual, and gender minority individuals, these groups all have different perceptions by their relative majority groups of white, heterosexual, and cisgender individuals. Hypothesis 2: Respondents will be most likely to assist Asian Americans, followed by gay men and lesbians, and be least likely to assist transgender individuals. Additionally, based upon the previous literature as well as self-categorization theory, it follows that members of the in-group will be less likely to intervene in hate events. 61 Hypothesis 3: Respondents who are cisgender, heterosexual, non-Hispanic white men will not progress as far through the bystander intervention model as members of out-groups. 62 Chapter 7: Methods The dissertation studies tested the above hypotheses using data from a national sample of 1,001 U.S. adults collected in April 2023. Although Qualtrics’ survey panels are opt-in utilizing stratified quota sampling procedures, research has shown them to reflect the U.S. population in terms of key demographic traits such as race, sex, income, and education (Heen et al., 2014). This panel also reflected the current racial composition of the U.S. Sample Characteristics The demographic characteristics of the Qualtrics sample are listed below in Table 1. Overall, the sample appears to reflect the demographic composition of the United States as specified by the quota sampling used for Qualtrics surveys. The sample was most frequently 65 years and older (24.4%) while the 18–24-year-old age group represented 11.7% of the sample. The majority, 74.5% were White while 13.2% were Black and 5.3% were Asian American or Pacific Islander. The remaining 7% of the sample consisted of multiracial (2.3%), American Indian/Alaskan Native (1.4%), Native Hawaiian (0.5%), and other racial minority group (2.8%) individuals. The majority of the sample, 82.02%, were not of Hispanic, Latino, or Spanish descent. The remaining 17.98% were Mexican/Mexican American/Chicano (9.89%), Puerto Rican (3.40%), Cuban (1.10%), or some other Hispanic, Latino, or Spanish ethnicity (3.60%). Approximately 98.8% of sample identified as cisgender (45.8% cisgender male and 53.0% cisgender female) while 0.7% were non-binary/gender nonconforming, 0.5% were transgender (0.3% transgender male and 0.2% transgender female), and 0.1% were genderfluid. Likewise, 89.6% identified as heterosexual or straight while 3.5% were gay or lesbian, 5.7% were bisexual, and 1.2% were another sexual orientation. When examining the sample across categories, 60.84% were non-Hispanic White cisgender heterosexuals. Specifically, 28.97% were non- 63 Hispanic White cisgender male heterosexuals. Although the sample was represented most by non-Hispanic White cisgender heterosexuals, these are the individuals in the best position to intervene since they are not as vulnerable to being targets of hate as minority group members are. The sample then permits valid testing of the hypotheses, particularly the third study exploring whether majority group members progress as far through the situational model as minority group members. The sample therefore reflects not only U.S. demographics but also the group of the most theoretical interest for bystander intervention to instances of hate. INSERT TABLE 1 HERE Respondents most frequently had some college (35.5%) or high school level educations (25.5%) while 36.2% either had a bachelor’s degree (21.8%) or a graduate/professional degree (14.4%). Only 2.9% had less than a high school level education. Over two-thirds of the sample identified as either liberal (34.8%) or conservative (34.2%). Nearly 34% of respondents are affiliated with the Democratic Party while 28% are affiliated with the Republican Party and 24% are independent voters. Judeo-Christians represented 61.8% of the sample while 31% had no religious affiliation. Approximately 84.5% of the sample claimed to never have been arrested while another 86.3% stated they had never been incarcerated. The Vignette Although some studies have previously used a vignette design for studying intervention to hate crimes, this novel contribution is a dynamic sequential vignette designed to track respondent progression through the five steps of the bystander decision-making model. This is tantamount to a choose-your-own-adventure scenario as the vignette will continue to unfold based upon the participant’s perceptions and choices. Participants had opportunities to progress through the five steps set forth by Latané and Darley (1970) or abstain from intervening when 64 witnessing simulated hate scenarios. Since hate crimes often emerge from earlier statements and threats, I apply this dynamic vignette approach to three escalating phases of hate scenarios: (1) early microaggressions; (2) hate incidents; and (3) hate crimes. Thus, respondents engage with all five steps of the bystander intervention decision-making model for each of the three phases of hate scenario escalation. This design allows for the approximation of real-world decision-making by breaking down the situation into the appropriate steps necessary to intervene. This also allows for the sequential analysis of outcomes dependent on previous steps in the bystander intervention decision-making process, as opposed to solely focusing upon reporting crimes to the police or the final step of intervention. The focus upon the early steps, in particular at earlier stages in hate incident escalation, is key as bystander intervention cannot occur without the bystander first recognizing the problem and forming options or means of intervening. Prior research has indicated that offender race and gender are strongly linked with perceptions of criminality (Brennan & Spohn, 2009; Chiricos et al., 2004; Steffensmeier et al., 2017). To control for this, the scenario will only use the most frequent offender of hate crime: a young white male. This novel approach will build upon the existing measures of bystander intervention in a few ways. First, previous researchers have designed their instruments to tap at general beliefs around bullying (Nickerson et al., 2014) or organ donation (Anker & Feeley, 2011) rather than responses to specific scenarios. Second, another study by Greitemeyer et al. (2006) asked respondents questions about each of the five steps in the contexts of emergencies in which they had already personally intervened or failed to intervene. A few of these scales had poor reliability, with Cronbach’s alphas of .59 for being aware of a situation and .53 for considering various approaches to intervention; this low scale reliability is further compounded by the issue 65 of conditional dependence among items intended to load upon only one of these related steps in bystander intervention decision-making. Retrospective studies like Greitemeyer et al.’s (2006) are subject to respondent recall bias, which can notoriously be worsened when individuals are asked to recall traumatic events (Berntsen et al., 2003). This vignette therefore offers more reliable and valid measures for testing the bystander decision-making process, as it will tap into context-specific reactions and perceptions as subjects progress virtually through hate incidents in a manner reflective of witnessing nonfictional hate incidents. In alignment with methodological research on vignettes, these vignettes are each approximately 50 to 75 words for a total of 150 to 225 words. Brevity is important for vignette construction to ensure that the vignettes hold each respondent’s attention (Hughes & Huby, 2004). Shorter length, especially when conducting studies on violent incidents, can also help to protect respondents by minimizing the possibility of emotional harm from excessive detail (Bradbury-Jones et al., 2014). Moreover, no characteristics were ascribed to either the offender or the victim beyond what was relevant to the hate event. Details as to the offender and victim were kept to a minimum to reduce the likelihood of the dilution effect, which occurs when respondents individuate members of the vignette once too many characteristics are provided (Nisbett et al., 1981). External Validity of the Vignettes In line with methodological best practices for vignette research (see Aguinis & Bradley, 2014; Erfanian et al., 2020), a few strategies were used to maximize the external validity of the vignettes. First, stakeholder groups that offer trainings on hate were identified and asked to provide feedback on the vignettes. These groups included the Center for Anti-Violence Education and the New York City Commission on Human Rights. Representatives of these two 66 organizations examined the vignettes and confirmed they followed how hate events occur in real life. Second, the vignettes provide respondents with enough information to ascertain that from context that the victim is being targeted due to their identity. However, the vignettes are vague enough to allow respondents to provide their own interpretations to the events such that they are not told they are driven by hate. Third, the short length of the vignettes helps to maintain respondents’ interest and lower the reading comprehension skills necessary to understand the events. Combined, these factors enhance the external validity of the vignettes by approximating how hate events naturally unfold, giving respondents enough information to determine the events are motivated by hate without directly stating it, and broadening the number of respondents able to comprehend the vignettes through their short length. Social Desirability Bias Within survey methods, social desirability bias is often a threat to validity in which respondents answer questions based upon what they believe the normative or desirable response is as opposed to the providing the true answer for themselves (Kreuter et al., 2009). Labhardt and colleagues (2017) noted social desirability bias undermines bystander intervention research as individuals may wish to be perceived as heroic by research staff for intervening although they would not do so in a natural event. To reduce this threat, several precautions were taken based on Labhardt and colleagues’ (2017) recommendations. First, the instructions for the vignettes and the survey itself were intentionally vague to prevent respondents from knowing what was being tested. Second, the anonymous nature of the online survey enabled respondents to answer without research staff present as the presence of others elicits socially desirable responses. Third, respondents were asked what they thought of the event without referencing how others would 67 think of it, thereby reducing the risk of them answering based on how they believed others would answer. Dependent Variables The dependent variables are the extent to which the respondent progressed through the situational model for the microaggression, hate incident, and hate crime portions of the vignette. Respondents answered yes or no as to whether they considered what they observed to be a problem, felt obligated to help the victim, and if there was anything they could do to help. Respondents were then asked to list a maximum of four interventions they could use to help the victim in blank text box cells. They then answered yes or no regarding whether they would do something to help the victim, then indicated which of the interventions they would perform to help. A series of binary variables were created for each step of the bystander intervention model and for each part of the escalating hate event. A categorical count variable was coded from zero to four for each part of the vignette, with zero indicating the respondent did not progress at all through the situational model, one indicating the respondent only considered the event to be a problem, two indicating the respondent also felt obligated to assist, three indicating the respondent also thought they could assist, and four indicating the respondent would implement an intervention to help. In addition to this variable of progress through the situational model, three other dependent variables were included after examining the distribution of the data. These were (1) whether the respondent saw the particular phase as a problem and (2) whether the respondent indicated they would intervene for the particular phase. Moreover, as discussed below, a latent class analysis was conducted as a data reduction tool for progress among all three hate event phases as part of study 1. This dependent variable was used for study 2 and study 3. 68 Independent Variables As described below in the analytic strategy, only study 3 required the use of independent variables. The independent variables for study 3 were bystander characteristics. These included bystander efficacy, bystander behavior, empathy, responses to the dictator and ultimatum games, feeling thermometer rating for the relevant minority group, and bystander demographics. Bystander Efficacy Bystander efficacy was evaluated through Slaby et al.’s (1994) nine-item scale to assess the extent to which participants thought violence could be prevented through bystander intervention. Some examples include “people’s violent behavior can be prevented” and “I can learn to do or say the kinds of things that help prevent violence.” These items use a six-point Likert scale with responses ranging from agree completely (=5) to disagree completely (=0). Confirmatory factor analysis retained eight of the nine items. These eight items had a Cronbach’s alpha of 0.9353, indicating high agreement between items (Cortina, 1993). An additive scale was created with respondents having a median of 31 out of 40 points on the scale (M = 30.67, SD = 6.77). Bystander Behavior Bystander behavior was assessed through a four-item scale by Banyard and Moynihan (2011). These items examine whether respondents have confronted prejudicial comments or jokes within the past two months. Specifically, these are in reference to sexist, racist, and homophobic language as well as cat calls. Each item had binary responses of yes (=1) or no (=0). Confirmatory factor analysis retained all four items. These four items had a Cronbach’s alpha of 0.8710, indicating high agreement between the items (Cortina, 1993). An additive scale was 69 created using these four items, on which respondents had a median of 3 on the scale (M = 2.45, SD = 1.64). Empathy The revised 19-item Basic Empathy Scale in Adults (BES-A) developed by Carré et al. (2013) was included to test respondents’ overall level of empathy. Nine of these items are reverse-coded. Each item uses a five-point Likert scale from strongly agree (=4) to strongly disagree (=0). Nine items were retained after conducting a confirmatory factor analysis3. These items had a Cronbach’s alpha of .7072, indicating an acceptable agreement between items (Cortina, 1993). An additive scale was created with the remaining items. Respondents had a median of 22 points on the scale (M = 22.22, SD = 5.46). Dictator and Ultimatum Games The dictator and ultimatum games are derived from Paternoster et al. (2017). Within dictator games, the respondent is provided with various financial sums (here, $100, $500, $5,000 and $10,000) and told to allot a certain amount of money to an imaginary other. The other individual must accept whatever the respondent offers no matter how small. Thus, a selfish respondent should offer nothing to the imaginary second player so that they can keep all the money. Within the ultimatum game, however, the imagined other can choose to reject the offer. If the offer is rejected, then the respondent also gets nothing. Therefore, the initial respondent should offer the imaginary other the minimum amount required to get them to accept the offer. 3 The confirmatory factor analysis revealed the presence of two factors for the empathy scale. This appeared to be a method effect as factor one contained all positively-worded items while factor two contained all negatively-worded items. Such a method effect is common as reverse-coded items represent their own methodological domain (Brown, 2015). Given the high number of reverse-coded items in the empathy scale as well as the 85 respondents who straight-lined empathy scale items, the best explanation for the two-factor solution appears to be the method effect of using reverse-coded items. 70 Two measures were used to capture the respondent’s preferences for other-regarding. First was the average amount the respondent indicated they would give to the second player for the dictator games in thousands of U.S. dollars. Second, a dummy variable indicating if the respondent offered at least half to the other player in all dictator and ultimatum games was used. These measures differ from the empathy scale in that they explore deliberate decision-making processes as opposed to automatic regard for others as with empathy. Respondents gave a median of 1.305 thousand U.S. dollars (M = 1.39, SD = 2.46). Feeling Thermometers Feelings towards various groups were assessed using feeling thermometers as opposed to using several scales measuring prejudice to reduce respondent fatigue (Jeong et al., 2023). These groups included straight men, straight women, gay men, lesbians, transgender men, transgender women, white men, white women, Asian men, and Asian women. Feeling thermometers were adapted from the General Social Survey (Smith et al., 2019). Respondents were advised to report their feelings towards various groups on a scale between 0 and 100 with ratings above 50 indicating favorability and ratings below 50 indicating unfavorability. The respondent’s rating on the feeling thermometer for their assigned victim group was used in a binary variable of whether the respondent felt favorably or unfavorably toward the group. Respondents provided a median of 75 points on the scale (M = 70.43, SD = 27.52). Overall, 83.92% of respondents reported feeling favorably toward their randomly assigned victim’s group. For each individual victim group, respondents assigned to the respective victim gave the following distributions. For the Asian man, respondents reported a median of 80 points on the scale (n=169, M = 76.72, SD = 23.12) with 92.3% indicating warm feelings. For the Asian woman, respondents reported a median of 87 points on the scale (n = 158, M = 81.42, SD = 71 19.16) with 95.57% indicating warm feelings. For the gay man, respondents reported a median of 80 points on the scale (n = 174, M = 72.43, SD = 25.44) with 86.21% indicating warm feelings. For the lesbian, respondents reported a median of 80 points on the scale (n = 167, M = 71.70, SD = 26.99) with 85.63% indicating warm feelings. For the trans man, respondents reported a median of 66 points on the scale (n = 141, M = 63.98, SD = 29.48) with 76.60% indicating warm feelings. For the trans woman, respondents reported a median of 56 points on the scale (n=192, M = 57.65, SD = 31.64) with 68.75% indicating warm feelings. Bystander Demographics Lastly, bystander demographic will be captured with one variable that will indicate whether the bystander is a cisgender, heterosexual, non-Hispanic white male (=0) or not (=1) so that cisgender, heterosexual, non-Hispanic white men will act as the reference group. Respondents were asked to provide their gender identity, sexual orientation, race, and ethnicity. Analytic Strategy Research Question 1: Does bystander progression through the situational model increase concomitantly with increasing severity of the hate scenario? A series of McNemar’s tests comparing differences in whether an individual sees a problem or intervenes for each phase were conducted. Wilcoxon-signed rank tests exploring whether individuals progressed further through escalating phases of the hate event were then used with the count variable for the number of steps in the situational model through which the respondent progressed. A latent class analysis was then completed to reduce the data from three separate dependent variables of the progress for each phase into one dependent variable. These McNemar’s tests and Wilcoxon-signed rank tests were selected over chi-square testing as they are non-parametric tests that account for the dependent nature of the samples for 72 each hate event phase (McDonald, 2014). Specifically, all respondents were presented with all phases of the hate event. Therefore, an individual respondent’s progress through the hate incident is dependent on their progress through the microaggression, for example, as they already have already given their situational model progress for one phase in the sequential series. McNemar’s tests are 2 x 2 contingency tables for paired nominal data. This non-parametric test is suited for matched pair designs in which subjects are treated under two conditions that are then compared to one another (McDonald, 2014). For study 1, these are pairwise comparisons for whether respondents saw a problem between each phase as well as whether they intervened for each phase. Olivier and Bell’s (2013) interpretations of odds ratios were used to determine the effect size for each significant finding. Small, medium, and large effect sizes were represented by odds ratios of 1.22, 1.86, and 3.00, respectively. Wilcoxon signed-rank tests likewise are non-parametric tests designed for dependent samples. Unlike the McNemar’s tests, these use pairwise comparisons of interval or ratio-level data (McDonald, 2014). The test examines whether a respondent progresses further through the situational model from one phase to the next (a positive signed change), progresses less (a negative signed change), or progresses the same (a zero signed change). The Wilcoxon signed-rank test then evaluates the symmetry of the resulting distribution to determine whether respondents were significantly more likely to progress or regress as the hate event escalates. Effect sizes for these tests were calculated by dividing the obtained z-score by the square root of the total sample, which would be 2,002 (Cohen, 1988). Per Cohen (1988), these were interpreted such that effect sizes of .1, .3, and .5 corresponded to small, medium, and large effects, respectively. 73 A latent class analysis was also conducted to unveil heterogeneity in bystander intervention. Latent class analysis is ideal for addressing heterogeneity as it can reveal and categorize respondents into latent subgroups (McCutcheon, 1987). LCA is founded upon there existing a latent categorical variable, X, of which there are subgroups represented by categories of X. These categories originate from observed patterns in a group of manifest variables which are observed categorical variables. In this first study, each manifest variable is a categorical variable that indicates how far the respondent went through the situational model for each phase of the hate event. Such variables must be used as there is local dependence between each step of the model, meaning that including binary variables for each step would violate the statistical assumptions required for LCA. As the number of latent classes must be specified before conducting latent class analysis, the first analysis will specify only one latent class existing. The number of latent classes specified in the models will increase until the Bayesian Information Criterion (BIC) begins to decrease. BIC is one measure of how well the LCA fits the available data (Nagin, 2005). In combination with BIC, the average posterior-probability will also be used to determine the best number of latent classes for bystanders. It is necessary to examine another measure of fit as BIC tends to increase as the number of latent classes increases. Research Question 2: Does the decision-making of bystanders to hate crimes differ based upon whether the target is a racial, sexual, or gender minority victim? A series of chi-square tests were used in study 2 to answer research question 2. Chi-squares for exploring differences in (1) seeing a problem and (2) intervening for each phase were conducted as 2 x 2 contingency tables. Due to the experimental research design, no control variables were necessary to isolate the effects of the victim group on bystander decision-making 74 (York, 2018). Chi-square tests of independence were conducted for each pairwise comparison of victim groups. These tests are equivalent to proportion-z tests as the effect size, 𝜙, provides directionality for 2 x 2 chi-squares (Bradley & Cutcomb, 1977). The two-tailed hypothesis test done by chi-squares allows for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. The Bonferroni correction was applied to adjust for multiple comparisons when doing pairwise comparisons between the victim groups (Armstrong, 2014). A total of 15 pairwise comparisons were made for each dependent variable; therefore, the alpha for significance was p = .0033. The Bonferroni correction was not applied to the omnibus tests; therefore, the alpha for significance was p = .05. Research Question 3: Do bystander characteristics influence their bystander decision-making processes while witnessing hate? A series of logistic regressions were used to answer research question 3. Logistic regression is a statistical test with a binary categorical outcome often defined as the presence or absence of something (Britt & Weisburd, 2010). The coefficient for an independent variable in the model is defined such that a one unit increase in the independent variables is associated with the value of the coefficient’s change in the log of the odds in the dependent variable. These coefficients can be exponentiated to shift the interpretation to an odds ratio as was done for study 3. The first two parts of study 3 consist of (1) whether the respondent saw a problem with the hate event phase and (2) whether the respondent intervened for the hate event phase. These two outcomes were separately regressed upon empathy, bystander behavior, bystander efficacy, 75 average spent on the dictator games, whether the individual was other-regarding in the dictator and ultimatum games, whether the respondent felt favorably or unfavorably towards the victim’s minority group, and whether the respondent was a majority group member. A multinomial logistic regression was then used to answer research question 3 with the dependent variable as the latent class to which each respondent was assigned from study 1. Multinomial logistic regression is an extension of logistic regression in which the outcome variable is a set of three or more categories that cannot be rank-ordered (Hosmer & Lemeshow, 1989). The independent variables included bystander efficacy, bystander behavior, empathy, dictator and ultimatum games, feeling thermometer ratings, and a dummy variable for cisgender, heterosexual, non-Hispanic white males in the sample. The variance inflation factor (VIF) was examined to test for multicollinearity among the independent variables (Thompson et al., 2017). The highest VIF was 1.14, indicating that no multicollinearity was present. The highest VIF for models including direct effects of each of the demographic variables was 1.21, also indicating that no multicollinearity was present. 76 Table 15. Sample Demographic Characteristics (n=1,001) Variables Values % Age Group 18-24 years 11.70% 25-34 years 16.10% 35-44 years 17.80% 45-54 years 15.00% 55-64 years 15.10% 65+ years 24.40% Race White 74.50% Black 13.20% Asian American or Pacific Islander 5.30% Multiracial 2.30% American Indian or Alaskan Native 1.40% Native Hawaiian 0.50% Other 2.80% Gender Identity Cisgender male 45.80% Cisgender female 53.00% Non-binary/gender nonconforming 0.70% Transgender male 0.30% Transgender female 0.20% Genderfluid 0.10% Sexual Orientation Heterosexual 89.60% Gay or Lesbian 3.50% Bisexual 5.70% Other 1.20% Education Less than high school 2.90% High school 25.50% Some college 35.50% Bachelor’s degree 21.80% 77 Graduate or professional 14.40% Political Orientation Liberal 34.80% Neither liberal nor conservative 31.10% Conservative 34.20% Political Party Affiliation Democratic 34.00% Republican 28.00% Independent 24.00% Other 1.20% Not registered 12.90% Religious Affiliation Protestant 34.90% Roman Catholic 24.20% Mormon 0.90% Jewish 1.80% Hindu 0.60% Muslim 2.00% Buddhist 1.90% Other 2.80% No religious affiliation 31.00% Arrest History Yes 15.50% No 84.50% Incarceration History Yes 13.70% No 86.30% 78 Chapter 8: Study One Results In this chapter, the theoretical background informing the study designed to answer research question 1 will be reiterated. This includes a summary of Latané and Darley’s (1970) situational model of bystander intervention and how event severity impacts the model. Descriptive information on the dependent variables then follows, including progress through the situational model for the microaggression, hate incident, and hate crime phases. Contingency tables highlighting differences in situational model progress between each phase are then presented. McNemar’s tests comparing differences in whether an individual sees a problem or intervenes for each phase are shown next followed by Wilcoxon-signed rank tests exploring whether individuals progress further through escalating phases of the hate event. The chapter finishes with the latent class analysis used to generate the combined dependent variable for studies 2 and 3. The most frequent progressions for all three phases are displayed first. Model fitting for the latent class analysis comes after, concluding with a comparison between the three- and five-class solutions as they demonstrated the best model fit. Review of Hate The severity of an incident has been shown to impact bystander decision-making. Researchers have examined severity within the contexts of sexual violence (Bennett & Banyard, 2016), cyberbullying (Huang et al., 2023), and intimate partner violence (Chabot et al., 2009). Within these studies, the researchers all found that bystander willingness to intervene significantly increased the more severe the situation. Moreover, Huang et al. (2023) found incident severity mediated the relationship between feeling obligated to assist and ultimately assisting. No research to date has explored severity in a hate scenario. 79 The relationship between event severity and bystander intervention is paramount when studying hate crime as hate occurs across a continuum of severity (Schweppe & Perry, 2022). This continuum, from least to greatest severity, includes microaggressions, hate speech, terrorism, and genocide. Microaggressions can result in a wide array of psychological and somatic consequences for their targets (Waldron, 2012). Hate speech can often escalate into violent hate crime assaults (Schweppe & Perry, 2022). The Anti-Defamation League (2003)’s Pyramid of Hate represents this continuum. Biased attitudes, which include accepting prejudicial jokes, comprise the foundation of the pyramid. The silence of bystanders to these jokes signals their agreement with the stereotypes presented by the jokes. Discrimination and prejudicial actions towards specific individuals form the next step. Violence including threats, assault, and murder come next. The Anti-Defamation League (2003) observed that these steps of hate build on one another with the tolerance of minor hateful acts being conducive to advancement up the pyramid. Empirical evidence supports this assertion as hate crimes against sexual minority victims often begin with homophobic slurs (Herek et al., 2002). Moreover, 99% of hate crime victims know they were targeted because of their identity based on language used during the offense (Masucci & Langton, 2017). Thus, the study of bystander intervention at minor acts of hate is necessary to uncover mechanisms increasing the likelihood of bystanders helping at these minor stages to prevent escalation to violence. The question being addressed by this first study is then: Research Question 1: Does bystander progression through the situational model increase concomitantly with increasing severity of the hate scenario? 80 Theoretical Background To intervene, Latané and Darley (1970) posited that a passerby must notice the event, consider it an emergency, feel morally obligated to intervene, ponder options to intervene, then implement an intervention. Bystanders often fail to notice microaggressions given their subtle nature (Lilienfeld, 2017). Even when they do, they may fail to consider these problems that merit intervention (McClure, 2020). Some bystanders may perceive microaggressions or hate speech to be jokes, thereby keeping them from progressing through the situational model as they do not consider these incidents to be intervention worthy (Katz et al., 2019a, 2019b, 2022). Thus, the hypothesis being tested for this first study is: Hypothesis 1: Respondents will go further in the situational model’s decision-making steps as the hate event progresses and becomes increasingly severe. Progression through the Situational Model To test hypothesis 1, respondent progress through each hate event phase is examined. Table 2 depicts the furthest step respondents reached for each phase of the hate event in the vignettes. The largest cutoff points for the data appear to be around whether an individual saw the phase as a problem and whether an individual chose to intervene for that phase. Specifically, 23.9% of respondents did not see a problem for the microaggression phase compared to 8.9% and 4.9% of respondents who did not see problems with the hate incident and hate crime phases, respectively. Over one-third of respondents, 37.4%, intervened for the microaggression phase while 65.6% and 80.5% intervened for the hate incident and hate crime phases, respectively. Overall, 83.9% of respondents intervened for at least one of the hate event phases. Although this variation demonstrates that seeing a problem or not and intervening or not could be dependent variables, these represent only the extreme ends of the situational model. Therefore, a method to 81 test the greater complexity of the entire situational model is required in addition to these extremes. INSERT TABLE 2 HERE Tables 3a, 3b, and 3c contain the progress through the situational model respondents made for hate incident based on their progress through the microaggression scenario (Table 3a), for hate crime based on microaggression progress (Table 3b), and for hate crime based on hate incident progress (Table 3c). Importantly, 776 respondents4 either increased progress through the situational model for the hate incident phase compared to the microaggression phase (n = 428, 42.8%) or intervened for both (n = 348, 34.8%). Additionally, 879 respondents either increased progress through the situational model for the hate crime phase compared to the microaggression phase (n = 527, 52.6%) or intervened for both (n = 352, 35.2%). Lastly, 846 respondents either increased progress through the situational model for the hate crime phase compared to the hate incident phase (n = 216, 21.6%) or intervened for both (n = 630, 62.9%). These results highlight that respondents typically do not have a maximum threshold throughout the situational model. Importantly, they suggest that a more nuanced approach may be needed to best combine data across the different escalating hate phases as there are differences in proceeding through the first step and completing the situational model by intervening. Tests of different ways to operationalize data for the dependent variables follow next to determine if there are statistically significant differences in situational model progress as the event escalates. INSERT TABLES 3a-3c HERE 4 Individuals who intervened for both phases are included as the data are truncated with a maximum step of intervention in the situational model. 82 Given that seeing a problem for each phase and intervening for each phase appear to be the biggest respondent breakoff points, several McNemar’s tests were conducted to determine if there were significant differences in the likelihoods of respondents reaching these points of the situational model between each phase. The results for whether a respondent considered the phase of the hate event a problem are summarized in Tables 4a, 4b, and 4c. Respondents were 8.895 times more likely to consider hate incidents to be problems than microaggressions (McNemar’s χ2(1) = 119.68, p < .0001). They were 16.833 times more likely to view hate crimes to be problems than microaggressions (McNemar’s χ2(1) = 168.69, p < .0001). They were also 5.444 times more likely to consider hate crimes to be problems than hate incidents (McNemar’s χ2(1) = 27.59, p < .0001). The magnitudes of these odds ratios indicate that these are large effect sizes for the observed significant differences (Olivier & Bell, 2013). INSERT TABLES 4a-4c HERE The McNemar’s Tests conducted for whether a respondent intervened for the phase of the hate event are summarized in Tables 5a, 5b, and 5c. Respondents were 11.885 times more likely to intervene for hate incidents than microaggressions (McNemar’s χ2(1) = 239.07, p < .0001). They were 20.636 times more likely to intervene for hate crimes than microaggressions (McNemar’s χ2(1) = 392.07, p < .0001). They were also 6.519 times more likely to intervene for hate crimes than hate incidents (McNemar’s χ2(1) = 109.36, p < .0001). The magnitudes of these odds ratios indicate that these are large effect sizes for the observed significant differences (Olivier & Bell, 2013). Overall, the results indicate that respondents are significantly more likely to see the event as a problem and intervene as the event escalates. While these results preliminarily revealed that respondent progress through the situational model increases with 83 increasing severity of what is occurring, they focus on discrete steps rather than the full situational model. INSERT TABLES 5a-5c HERE Next, Wilcoxon signed-rank tests were performed to determine whether respondents progressed further through the situational model as the event escalated while capturing all steps of the situational model. These tests are shown below in Tables 6a, 6b, and 6c. The sign of change indicates whether the respondent went further through the situational model (positive), maintained the same progress through the situational model (zero), or regressed through the situational model (negative) as the event escalated. Respondents progressed significantly further through the situational model for hate incidents than microaggressions (42.8% of respondents increased, z = 17.814, p < .0001), hate crimes than microaggressions (52.6% of respondents increased, z = 20.921, p < .0001) , and hate crimes than hate incidents (21.6% of respondents increased, z = 10.990, p < .0001). The effect sizes for these tests were 0.3981 for microaggression versus hate incident, 0.4676 for microaggression versus hate crime, and 0.2456 for hate incident versus hate crime progression through the situational model, indicating small to medium strength for these effects (Cohen, 1988). These results support hypothesis 1 that individuals progress further through the situational model as the event severity increases. INSERT TABLES 6a-6c HERE Latent Class Analysis The results of the previous tests supported hypothesis 1 but also highlighted the complexity of the data. To further show this, Table 7 contains the ten most frequent combinations of situational model progression in the sample across all three phases of the hate event. One-third of the sample consisted of people who always intervened no matter the phase. 84 Next, 19.4% of the sample saw a problem with the microaggression phase but only intervened during the hate incident and hate crime phases. Even when examining the top ten most frequent patterns of situational model progress, the tenth most frequent pattern reflected only 1.8% of the sample. Given the vast range of possible combinations, multinomial logistic regression cannot be used to assess all possible patterns of situational model progress exhibited by the respondents. An alternative would be to employ a data reduction technique to detect latent patterns of the sample that would allow all respondents to be captured: latent class analysis. The remaining groups possess much heterogeneity as to whether they saw a problem, felt morally obligated to help, or intervened. This heterogeneity in situational model progression highlights the need for a latent class analysis to reduce the data into a more manageable form. INSERT TABLE 7 HERE The results of the latent class model fitting are summarized below in Table 8. As shown, the number of possible latent class solutions were examined until the nine-class solution. Further testing ceased as the lowest posterior probability for each class solution kept decreasing after the six-class solution. Moreover, the Bayesian information criterion (BIC) kept increasing after the five-class solution, indicating that these models were not the best fit for the data (Weller et al., 2020). INSERT TABLE 8 HERE In narrowing down the potential latent class solutions, the lowest average posterior probability for class membership was used as the first exclusion criterion. Posterior probabilities greater than 0.9 are preferred, but ones that are above 0.8 can be considered sufficient so long as the classes make theoretical sense (Weller et al., 2020). Using that rule, the four-, eight-, and 85 nine-class solutions all were excluded as their lowest average posterior probabilities for a given class were 0.5899, 0.7678, and 0.6446, respectively. Next, the sample size of the smallest latent class for each solution was considered such that classes containing less than 5% of the sample (50 respondents) were deemed unfavorable. These were the five-, six-, and seven-class solutions as they possessed minimum class sizes of 23, 13, and 5 respondents, respectively. These class solutions were not removed entirely from consideration as statisticians have noted that small class size can be ignored so long as the class is theoretically sound (Weller et al., 2020). The Akaike Information Criterion (AIC) and the aforementioned BIC were then examined. Lower values on both indicate better model fit. The AIC showed that the three-class solution had a better model fit than the two-class solution (5,446.356 versus 5,471.993). The BIC, however, showed that the two-class solution had a better model fit than the three-class solution (5,579.985 versus 5,593.619). BIC has been found to be a more robust indicator of model fit than AIC, which would therefore give preference to the two-class solution (Nylund et al., 2007). Regardless, the latent class marginal means of each solution were examined to determine which was best. The two-class solution, shown in Table A3 of the appendix, could be summarized as individuals who intervened for every phase in class one, and individuals who only intervened for the hate crime phase in class two. The three-class solution shown in Table 9, conversely, contained classes of individuals who always intervened, those who intervened as the event escalated, and those who never intervened. Given that the two-class solution removed much nuance of the dependent variables, the three-class solution is preferred. Of the three classes with small minimum class memberships, the AIC was lowest for the seven-,six-, then five-class solutions whereas the BIC had the opposite trend such that it was 86 lowest for the five-, six-, then seven-class solutions. Given that the five-class solution had the largest small class size and the highest minimum posterior probability, along with the BIC being a better indicator for model fit than the AIC, the five-class solution remained a possible selection over the six- and seven-class solutions. Thus, two potential class solutions emerged from the analysis: the three-class and five-class solutions. While the three-class solution had the lower minimum average posterior probability (.8545 versus .8974) and a higher BIC (5,593.619 versus 5,520.045) than the five-class solution, it had a much larger smallest class of 240 respondents as opposed to the 23 respondents constituting one class of the five-class solution. The latent class marginal means for the three- and five-class solutions are depicted in Tables 9 and 10. INSERT TABLES 9 AND 10 HERE The classes generated by the three-class solution can best be summarized as individuals who always intervene (38.56% of the sample), individuals who intervene as the hate event escalates (37.46% of the sample), and individuals who never intervene as they either fail to see a problem or fail to feel morally obligated to assist (23.98% of the sample). Within the five-class solution, the classes are comprised of people who always intervened (45.35%), intervened for hate incident and hate crime phases but not microaggression (24.68%), only intervened for hate crime (17.98%), never saw a problem for any phase (9.69%), and progressed further through the model as the event escalated but did not intervene (2.3%). A chi-square test of independence was conducted to test whether there was an association between three-class and five-class membership in the sample. The test was also performed because of the overlap in general class descriptions arising from the latent class marginal means. The first class of both models mimic one another in that the class members tend to intervene for 87 every phase of the hate event. The second class of the three-class solution is similar to classes two, three, and five of the five-class solution as respondents belonging to those classes progress further through the situational model as the event escalates. The third class of the three-class solution and fourth class of the five-class solution similarly contain respondents who never intervene. Table 11 conveys the results of this test. INSERT TABLE 11 HERE Per Table 11, approximately 95.3% of the always interveners in the three-class solution also share membership with the always interveners in the five-class solution. The remaining 4.7% belong to the second class of the five-class solution composed of individuals who intervened for both the hate incident and hate crime phases. For the subjects in the class that intervenes as the situation escalates in the three-class solution, 76.6% belonged to similar classes in the five-class solution: 61.1% were classified as subjects who intervened as the situation escalated beyond a microaggression, and 15.5% were classified as subjects who intervened once physical violence occurred during the hate crime. Over one-fifth, 22.4%, belonged to the five-class solution’s always interveners while 1.1% were the individuals who never saw a problem. Lastly, the three-class solution’s subjects who never intervened were almost evenly split between the five-class solution’s individuals who only intervened once the event became violent (50.8%) and individuals who never saw a problem for any phase (40.4%). The remaining 7.9% were classified as individuals who progressed through the model as the situation escalated but never intervened. There appears to be considerable overlap in the three- and five-class solutions’ memberships. While the five-class solution has a stronger model fit than the three-class solution, the additional classes are not theoretically relevant for the purposes of this dissertation. Classes 88 two, three, and five all follow the theoretical expectation of respondents progressing further through the situational model as the event escalates with different points at which intervention occurs. This point of intervention, in and of itself, is not being tested within this chapter. Rather, the overall pattern of respondents intervening as the situation escalates as encapsulated by class two of the three-class solution is. Mixture modeling techniques such as latent class analysis tend to prefer additional classes even when some of them are unnecessary, requiring researchers to look beyond model fit criteria when selecting the best solution (Collins & Lanza, 2010). The parsimony of the three-class solution and the heightened theoretical relevance of its classes makes it preferable to the five-class solution. Considering the dependence of these class solutions revealed through the chi-square test of independence and the overlap between both class solutions, the three-class solution will be used for the remaining studies due to its parsimony, theoretically relevant classes, and larger class sizes. Discussion Overall, these results indicate that bystanders progress significantly further through the situational model as a hate event escalates. No threshold effect was observed in which respondents would progress only to a point within the situational model and stop progressing across all hate phases. The substantial heterogeneity in situational model progress for all three phases necessitated the use of a data reduction technique, latent class analysis, to uncover hidden patterns and better classify bystanders. Additionally, bystanders in this sample fell into three latent classes in a probabilistic sense for data reduction purposes: always interveners, those who intervene as the situation progresses, and those who never intervene. What remains to be unpacked are how victim and bystander characteristics impact situational model progress and latent class membership. A 89 sizeable proportion of participants could be classified as always interveners. The results from study 2 will exemplify whether the identity of the victim predicts latent class membership. The results from study 3 will help to elucidate what characteristics, if any, help to propel bystanders to intervene. The obtained results support Latané and Darley’s (1970) situational model in that bystanders tended to progress further through the process as the event escalated. For example, 37.4% of bystanders intervened for microaggressions compared to 65.6% who intervened for hate incidents and 80.5% who intervened for hate crimes. These results conform to the prior literature which has found that increasing incident severity significantly predicts increased bystander willingness to intervene (Bennett & Banyard, 2016; Chabot et al., 2009; Huang et al., 2023). However, the results differ from Latané and Darley’s (1970) expectation that subjects can commit themselves to not intervening through failing to intervene at the onset of an emergency. They reasoned that bystanders may continue to talk themselves into believing no emergency is occurring to reinforce their initial decision not to perceive it as such. Thus, the longer an event transpires, the less likely a bystander is to modify their decision-making. Respondents in the current study were subjected to an escalating situation which may have acted as an additional stimulus for them to react. Indeed, the latent class analysis displayed that over three-fourths of subjects tended to either intervene in all situations (38.6%) or intervene with the escalating hate severity (37.5%). Additionally, the results comport with studies showing that many bystanders fail to recognize microaggressions as over one-fourth of respondents failed to consider them problematic (Lilienfeld, 2017). Even when they did recognize the microaggressions as 90 problematic, approximately 45% did not feel morally obligated to intervene. Although bystanders were significantly more likely to find the hate incident to be problematic and intervene, nearly one-third still did not intervene for this more explicit form of hate. The findings of this first study reveal the need for education on bystander intervention and the effects of hate. Given that many individuals either did not see the microaggression as a problem (23.9%) or did not feel morally obligated to assist (34.3%), bystanders must be taught to (1) recognize microaggressions and (2) understand their deleterious effects to see why intervention is so important. As previously noted, microaggression victimization has been associated with depression and anxiety (Nadal et al., 2014), higher stress (Smith et al., 2011), negative emotional states (Wang et al., 2011), and physical health symptomatology such as headaches and body pains (Huynh, 2012). People must learn about these consequences of microaggressions to feel compelled to intervene when witnessing them. This especially rings true when considering that effective bystander intervention that involves aftercare for the victim reduces the likelihood of these sequalae occurring (Hamby et al., 2016). This study possesses several limitations. First, respondents were recruited via quota sampling from Qualtrics rather than through a probability-based sampling method. This means that the respondents do not truly reflect the U.S. population as not every member had an equal chance of being selected, limiting the validity of the results as they may not extend to the true U.S. population. Second, there may be question order effects in which respondents went further through the situational model for the hate incident and hate crime phases given their responses for the prior phase(s). To fully explore how respondents would react to a hate event unfolding before them, it was necessary to include all phases and the situational model progress questions at the end of each phase. Each phase of the hate event ideally would have been randomly 91 assigned to eliminate the possibility of order effects, but that would have made the resultant cell sizes too small for meaningful analysis when combined with the random assignment of victim characteristics. Lastly, the latent class solution ultimately selected did not have the best goodness of fit in comparison to other class solutions, namely the five-class solution. While the three-class solution was selected based on its parsimony, theoretical relevance, and larger class sizes, it is possible that further nuance will remain veiled by using the three-class solution in the future studies of this dissertation rather than the five-class. 92 Table 16. Progression through the Situational Model by Hate Event Phase (n=1,001) Microaggression Hate Incident Hate Crime Maximum Threshold n (%) n (%) n (%) n (%) Did not see a problem 239 (23.9%) 89 (8.9%) 49 (4.9%) 32 (3.2%) Saw a problem 343 (34.3%) 186 (18.6%) 89 (8.9%) 83 (8.3%) Felt morally obligated to assist 29 (2.9%) 41 (4.1%) 40 (4.0%) 23 (2.3%) Considered options to intervene 16 (1.6%) 28 (2.8%) 17 (1.7%) 23 (2.3%) Intervened 374 (37.4%) 657 (65.6%) 806 (80.5%) 840 (83.9%) Note: Percentages do not add to 100% due to rounding. The percentages listed in each cell correspond to the number of respondents who reached the corresponding situational model step for that particular hate event phase out of all 1,001 respondents. 93 Table 17a. Hate Incident Progress through the Situational Model by Microaggression Progress (n=1,001) Hate Incident Intervention Step Did not see a problem Saw a problem Felt morally obligated to assist Considered options to intervene Intervened Increased progress n (%) n (%) n (%) n (%) n (%) n (%) Micro-aggress- ion Did not see a problem 70 (29.3%) 72 (30.1%) 10 (4.2%) 2 (0.8%) 85 (35.6%) 169 (70.7%) Saw a problem 9 (2.6%) 105 (30.6%) 16 (4.7%) 15 (4.4%) 198 (57.7%) 229 (66.8%) Felt morally obligated to assist 1 (3.5%) 3 (10.3%) 2 (6.9%) 4 (13.8%) 19 (65.5%) 23 (79.3%) Considered options to intervene 2 (12.5%) 1 (6.3%) 0 (0%) 6 (37.5%) 7 (43.8%) 7 (43.8%) Intervened 7 (1.9%) 5 (1.3%) 13 (3.5%) 1 (0.3%) 348 (93.1%) Note: The frequencies in this table are presented as joint frequencies between progress in the microaggression phase and progress in the hate incident phase. The percentages in this table are presented as row totals for situational model progress during the hate incident phase conditioned on situational model progress during the microaggression phase. Thus, each row adds to 100% with the percentage of respondents who increased progress in the last column being based on the number of respondents who stopped at that particular step for the microaggression phase. 94 Table 3b. Hate Crime Progress through the Situational Model by Microaggression Progress (n=1,001) Hate Crime Intervention Step Did not see a problem Saw a problem Felt morally obligated to assist Considered options to intervene Intervened Increased Progress n (%) n (%) n (%) n (%) n (%) n (%) Micro-aggress- ion Did not see a problem 37 (15.5%) 40 (16.7%) 9 (3.8%) 6 (2.5%) 147 (61.5%) 202 (84.5%) Saw a problem 8 (2.3%) 41 (12%) 12 (3.5%) 5 (1.5%) 277 (80.8%) 294 (85.7%) Felt morally obligated to assist 2 (6.9%) 1 (3.5%) 6 (20.7%) 1 (3.5%) 19 (65.5%) 20 (69.0%) Considered options to intervene 1 (6.3%) 1 (6.3%) 1 (6.3%) 2 (12.5%) 2 (12.5%) 11 (68.8%) Intervened 1 (0.3%) 6 (1.6%) 12 (3.2%) 3 (0.8%) 352 (94.1%) Note: The frequencies in this table are presented as joint frequencies between progress in the microaggression phase and progress in the hate crime phase. The percentages in this table are presented as row totals for situational model progress during the hate crime phase conditioned on situational model progress during the microaggression phase. Thus, each row adds to 100% with the percentage of respondents who increased progress in the last column being based on the number of respondents who stopped at that particular step for the microaggression phase. 95 Table 3c. Hate Crime Progress through the Situational Model by Hate Incident Progress (n=1,001) Hate Crime Intervention Step Did not see a problem Saw a problem Felt morally obligated to assist Considered options to intervene Intervened Increased Progress n (%) n (%) n (%) n (%) n (%) n (%) Hate Incident Did not see a problem 40 (44.9%) 16 (18.0%) 4 (4.5%) 0 (0%) 29 (32.6%) 49 (55.1%) Saw a problem 4 (2.2%) 60 (32.3%) 12 (6.5%) 8 (4.3%) 102 (54.8%) 122 (65.6%) Felt morally obligated to assist 1 (2.4%) 2 (4.9%) 7 (17.1%) 0 (0%) 31 (75.6%) 31 (75.6%) Considered options to intervene 1 (3.6%) 3 (10.7%) 4 (14.3%) 6 (21.4%) 14 (50.0%) 14 (50.0%) Intervened 3 (0.5%) 8 (1.2%) 13 (2.0%) 3 (0.5%) 630 (95.9%) Note: The frequencies in this table are presented as joint frequencies between progress in the hate incident phase and progress in the hate crime phase. The percentages in this table are presented as row totals for situational model progress during the hate crime phase conditioned on situational model progress during the hate incident phase. Thus, each row adds to 100% with the percentage of respondents who increased progress in the last column being based on the number of respondents who stopped at that particular step for the hate incident phase. 96 Table 18a. McNemar's Test of Seeing a Problem for Microaggression and Hate Incident (n=1,001) Hate Incident Saw a problem Did not see a problem Microaggression Saw a problem 743 19 Did not see a problem 169 70 McNemar’s χ2(1) 119.68 p-value <.0001 Odds ratio 8.895 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. Table 4b. McNemar's Test of Seeing a Problem for Microaggression and Hate Crime (n=1,001) Hate Crime Saw a problem Did not see a problem Microaggression Saw a problem 750 12 Did not see a problem 202 37 McNemar’s χ2(1) 168.69 p-value <.0001 Odds ratio 16.833 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. Table 4c. McNemar’s Test of Seeing a Problem for Hate Incident and Hate Crime (n=1,001) Hate Crime Saw a problem Did not see a problem Hate Incident Saw a problem 903 9 Did not see a problem 49 40 McNemar’s χ2(1) 27.59 p-value <.0001 Odds ratio 5.444 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. 97 Table 19a. McNemar's Test of Intervening for Microaggression and Hate Incident (n=1,001) Hate Incident Intervened Did not intervene Microaggression Intervened 348 26 Did not intervene 309 318 McNemar’s χ2(1) 239.07 p-value <.0001 Odds ratio 11.885 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. Table 5b. McNemar’s Test of Intervening for Microaggression and Hate Crime (n=1,001) Hate Crime Intervened Did not intervene Microaggression Intervened 352 22 Did not intervene 454 173 McNemar’s χ2(1) 392.07 p-value <.0001 Odds ratio 20.636 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. Table 5c. McNemar’s Test of Intervening for Hate Incident and Hate Crime (n=1,001) Hate Crime Intervened Did not intervene Hate Incident Intervened 630 27 Did not intervene 176 168 McNemar’s χ2(1) 109.36 p-value <.0001 Odds ratio 6.519 Note: A two-tailed McNemar’s test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. 98 Table 20a. Wilcoxon Signed-Rank Test for Microaggression and Hate Incident Situational Model Progress (n=1,001) Sign of Change n (%) Sum ranks Expected Positive 428 (42.8%) 330,192 180,127.5 Zero 531 (53.0%) 141,246 141,246 Negative 42 (4.2%) 30,063 180,127.5 z 17.814 p-value <.0001 Effect size 0.3981 Note: A two-tailed Wilcoxon signed-rank test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. Table 6b. Wilcoxon Signed-Rank Test for Microaggression and Hate Crime Situational Model Progress (n=1,001) Sign of Change n (%) Sum ranks Expected Positive 527 (52.6%) 385,179.5 202,680 Zero 438 (43.8%) 96,141 96,141 Negative 36 (3.6%) 20,180.5 202,680 z 20.921 p-value <.0001 Effect size 0.4676 Note: A two-tailed Wilcoxon signed-rank test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. 99 Table 6c. Wilcoxon Signed-Rank Test for Hate Incident and Hate Crime Situational Model Progress Sign of Change n (%) Sum ranks Expected Positive 216 (21.6%) 189,815 112,552.5 Zero 743 (74.2%) 276,396 276,396 Negative 42 (4.2%) 35,290 112,552.5 z 10.990 p-value <.0001 Effect size 0.2456 Note: A two-tailed Wilcoxon signed-rank test was used. The two-tailed hypothesis test was performed for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. 100 Table 21. Ten Most Frequent Combinations of Situational Model Progression by Hate Event Phase (n=1,001) Microaggression Hate Incident Hate Crime n (%) 1 Intervened Intervened Intervened 333 (33.3%) 2 Saw a problem Intervened Intervened 194 (19.4%) 3 Did not see a problem Intervened Intervened 78 (7.8%) 4 Saw a problem Saw a problem Intervened 58 (5.8%) 5 Did not see a problem Saw a problem Intervened 39 (3.9%) 6 Saw a problem Saw a problem Saw a problem 35 (3.5%) 7 Did not see a problem Did not see a problem Did not see a problem 32 (3.2%) 8 Did not see a problem Saw a problem Saw a problem 24 (2.4%) 9 Did not see a problem Did not see a problem Intervened 23 (2.3%) 10 Felt morally obligated to help Intervened Intervened 18 (1.8%) Table 22. Latent Class Analysis Selection Criteria Number of Classes AIC BIC Lowest Average Posterior Probability Smallest Class Size 2 5471.993 5579.985 0.9812 314 3 5446.356 5593.619 0.8545 240 4 5332.866 5524.307 0.5899 93 5 5289.334 5520.045 0.8974 23 6 5282.174 5571.790 0.8815 13 7 5280.420 5579.854 0.8767 5 8 5284.728 5579.253 0.7678 19 9 5273.862 5597.839 0.6446 9 101 Table 23. Three-Class Solution Latent Class Marginal Means (1) (2) (3) Latent Class Description Always interveners Intervene with escalating severity Never interveners Class Membership Size (n=1,001) 386 (38.56%) 375 (37.46%) 240 (23.98%) Average Posterior Probability 98.89% 85.45% 92.11% Minimum Posterior Probability 49.38% 53.95% 50.83% Microaggression Did not see problem 0.0875 0.1815 0.5759 Saw a problem 0.0000 0.7906 0.3077 Felt morally obligated to assist 0.0448 0.0000 0.0421 Considered options to intervene 0.0000 0.0279 0.0267 Intervened 0.8677 0.0000 0.0476 Hate Incident Did not see problem 0.0000 0.0007 0.3632 Saw a problem 0.0054 0.1858 0.4933 Felt morally obligated to assist 0.0355 0.0532 0.0332 Considered options to intervene 0.0000 0.0288 0.0745 Intervened 0.9591 0.7315 0.0359 Hate Crime Did not see problem 0.0000 0.0000 0.2002 Saw a problem 0.0121 0.0081 0.3318 Felt morally obligated to assist 0.0314 0.0151 0.0890 Considered options to intervene 0.0057 0.0000 0.0598 Intervened 0.9508 0.9768 0.3193 102 Table 24. Five-Class Solution Latent Class Marginal Means (1) (2) (3) (4) (5) Latent Class Description Always interveners Intervene as the situation escalates beyond micro-aggression Only intervene once physical violence occurs Never see a problem Progress through the model as expected but rarely intervene Class Membership Size (n=1,001) 454 (45.35%) 247 (24.68%) 180 (17.98%) 97 (9.69%) 23 (2.30%) Average Posterior Probability 89.74% 97.22% 96.6% 98.89% 90.02% Minimum Posterior Probability 57.28% 54.33% 51.12% 39.86% 53.27% Microaggression Did not see problem 0.2101 0.0000 0.3988 0.7808 0.0112 Saw a problem 0.0000 0.7674 0.5886 0.0962 0.4101 Felt morally obligated to assist 0.0000 0.0644 0.0000 0.0300 0.2825 Considered options to intervene 0.0000 0.0255 0.0013 0.0204 0.2359 Intervened 0.7899 0.1427 0.0013 0.0727 0.0603 Hate Incident Did not see problem 0.0000 0.0000 0.0003 0.8564 0.0000 Saw a problem 0.0083 0.0000 0.9612 0.0369 0.1704 Felt morally obligated to assist 0.0353 0.0516 0.0243 0.0367 0.1348 Considered options to intervene 0.0000 0.0190 0.0142 0.0146 0.6948 Intervened 0.9564 0.9293 0.0000 0.0553 0.0000 Hate Crime Did not see problem 0.0000 0.0000 0.0126 0.4505 0.0000 Saw a problem 0.0129 0.0082 0.3364 0.1753 0.0836 Felt morally obligated to assist 0.0321 0.0092 0.0588 0.0476 0.3240 Considered options to intervene 0.0074 0.0000 0.0376 0.0000 0.2649 Intervened 0.9476 0.9826 0.5546 0.3267 0.3275 103 Table 25. Chi-Square Test of Independence for Three- and Five-Class Membership (n=1,001) Five-Class 1 2 3 4 5 Three-Class 1 368 (95.3%) 18 (4.7%) 0 (0%) 0 (0%) 0 (0%) 2 84 (22.4%) 229 (61.1%) 58 (15.5%) 0 (0%) 4 (1.1%) 3 2 (0.8%) 0 (0%) 122 (50.8%) 97 (40.4%) 19 (7.9%) χ2(8) 1,250.863 p-value <.0001 Cramér’s V 0.790 Note: The two-tailed hypothesis test done by chi-squares allows for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. 104 Chapter 9: Study Two Results In this chapter, the literature and theoretical background informing the study designed to answer research question 2 will be reiterated. This includes a summary of the prior literature as to how minority status impacts bystander intervention and an overview of social categorization theory. Chi-square tests of independence are then presented comparing victims in terms of: (1) whether respondents considered the microaggression, hate incident, or hate crime phases to be problems, (2) whether bystanders intervened during the microaggression, hate incident, or hate crime phases, and (3) the distribution of latent classes for the three-class solution. In each case, an omnibus chi-square comparing respondents across all victim types occurs first followed by pairwise comparisons by victim type. Review of Victim Traits Affecting Bystander Intervention Various victim characteristics have been associated with changes in bystander behavior. Prior research has most often emphasized the role of victim race in impacting the willingness to intervene, but a few studies have also examined the roles of victim sexual orientation and gender identity. These studies tend to find that bystander intervention varies by the victim’s identity. Studies of the impact of victim race have found bystanders react more quickly and are more willing to assist White victims compared to Black victims. This phenomenon has been found across several different contexts including accidents (Gaertner et al., 1982), fall injuries (Kunstman & Plant, 2008), and even sexual assault (Katz et al., 2017). Likewise, bystanders assist less frequently for gay victims than straight victims (Owuamalam & Matos, 2020) and have less compassion for them (Owuamalam & Matos, 2022). Bystanders also show less empathy towards transgender victims of hate crime than cisgender victims (Godzisz & Mazurczak, 2023). People are less likely to intervene for transgender victims of violence 105 (Deacon et al., 2020). Researchers have also found teachers to be less likely to assist transgender students who are victims of bullying than cisgender students (Parker et al., 2023). Disparities in bystander intervention have also been found when comparing racial, sexual, and gender minority victims. Vera and colleagues (2023) noted college students were more likely to intervene for racial minority victims of bullying than sexual or gender minority victims. Within the situational model by Latané & Darley (1970), there are several steps at which the victim’s identity impacts progress towards intervention. Generally, onlookers often fail to detect microaggressions against racial/ethnic, sexual, and religious minorities (Lilienfeld, 2017), which is especially important as insults escalate into violence (Whitehead et al., 2018). Second, individuals often fail to recognize events as emergencies with sexual (Owuamalam & Matos, 2020) and racial minority victims (Gaertner et al., 1982) being the most minimized in comparison to heterosexual and white victims. In terms of feeling morally obligated to assist, bystanders often look to the demographic characteristics of the victim to determine victim blameworthiness. Indeed, majority group members respond less quickly to emergencies with racial minority victims (Kunstman & Plant, 2008; Gamberini et al., 2015) and have less empathy for sexual minority victims, including transgender victims (Mazurczak & Godzisz, 2019). Overall, prior studies have exhibited bystander intervention decreases when the victim is a minority. The type of minority appears to make an impact too as discovered by Vera and colleagues (2023), who found racial minority victims received more assistance than sexual and gender minority victims. Little research has examined the extent to which this relationship holds for hate events. Therefore, the research question explored in this chapter is: Research Question 2: Does the decision-making of bystanders to hate crimes differ based upon whether the target is a racial, sexual, or gender minority victim? 106 Theoretical Background Social categorization theory provides a framework for explaining why bystanders respond differently based upon the victim’s identity. Turner and colleagues (1987) proposed there are three levels of self-categorization: human identity, social identity, and personal identity. Human identity is the view of oneself as a member of the human race. Social identity pertains to what social groups an individual belongs, especially ingroup/majority versus outgroup/minority groups. Personal identity is how one views oneself in relation to other members of their social identity groups. This theory extends to how others are defined as well. Situational factors influence how individuals categorizes themselves and others in a given moment. Oakes (1987) and Oakes and colleagues (1981) found individuals categorize themselves and others based on the accessibility of possible groups. Moreover, individuals use these categories in ways that maximize between-group differences while minimizing within-group differences. Depersonalization then occurs in which individuals notice stereotypical patterns among group members. These group members cease to be unique individuals to the observer and are viewed as interchangeable. Stereotypic views of Asian Americans, for instance, have historically fallen into two diametrically opposed points: the model minority and the yellow peril. The former stereotype reinforces the view that Asian Americans commit little crime, excel in academics, are upwardly mobile, and hold high status jobs (Bonilla-Silva, 2004; Peterson, 1966). The latter stereotype views Asian Americans as the bringers of plagues such as SARS and COVID-19 as well as drugs such as opium (Reny & Barreto, 2022). Views of sexual and gender minority individuals have been decidedly more negative. Public sentiment towards sexual minority individuals has grown increasingly more negative related to the passage of “don’t say gay” bills entrenched in 107 homophobic rhetoric (Hartman, 2023; Oakley, 2022). Likewise, views of gender minority individuals have worsened with Republicans driving large increases in transphobia (Jones, 2023a). Within the context of this study of the dissertation, one prominent heuristic may impact how bystanders process the hate event. With the representativeness heuristic, bystanders may apply stereotypes to individuals based upon their minority group membership as they lack information about the individual (Kahneman & Tversky, 1972). For example, doctors perceive gay patients as being more likely to have HIV/AIDS even when provided with symptoms that matched those of heterosexual patients (Triplet, 1992). Individuals often associate being transgender with mental illness and deceitfulness (Locantore & Wasarhaley, 2020). Asian Americans, however, are often associated with professional success and stability (Ash et al., 2022). This stereotypical information is then easily accessible when encountering members of said groups, thereby influencing perceptions and behavior towards them (Carbado & Roithmayr, 2014). Within bystander intervention scenarios, bystanders may then attribute greater blame to victims from these groups, believing the harassment to be deserved. These bystanders may then choose neither to see the event as a problem nor as one worthy of intervention. In particular, this is expected to occur for microaggressions. For example, Xie and Galliher’s (2023) factorial vignette survey revealed witnesses and victims of microaggressions tend to perceive interveners as overreacting to the situation. Bystanders also may not perceive microaggressions as serious despite their deleterious effects and thus choose not to intervene (Mulvey et al., 2016). Given that individuals often use stereotypical views of groups to fill in the gaps about people they do not know, it follows that this information would inform bystander progress 108 through the situational model. Specifically, the contrasting stereotypic views of Asian Americans could lead to increased intervention for them compared to sexual and gender minority victims as those stereotypes are more negative. Sexual minority individuals moreover have more public support than gender minority individuals. Therefore, the hypothesis being tested for study 2 of this dissertation is: Hypothesis 2: Respondents will be most likely to assist Asian Americans, followed by gay men and lesbians, and be least likely to assist transgender individuals. Results Seeing a Problem for Each Phase The beginning of testing this hypothesis is whether respondents differed in their view of microaggression as a problem based on the victim’s identity. Table 12 shows the results of the omnibus chi-square test of independence between victim identity and whether a respondent viewed the hate event phase as a problem. Over 90% (90.51%) of respondents randomly assigned the hate scenario with an Asian woman victim considered the microaggression against her to be a problem. Respondents in the transgender man and Asian man conditions considered the microaggression to be a problem at similar rates: 80.85% for the transgender man and 80.47% for the Asian man. Subjects with the transgender woman hate victim displayed the next highest rate of seeing a problem of 75%. Respondents assigned hate events with a sexual minority victim reported the lowest rate of seeing a problem for microaggression: 68.97% for the gay man victim and 62.87% for the lesbian victim. The omnibus chi-square test showed that these proportions of seeing problems with microaggression for each victim were significantly differently (χ2(5) = 42.644, p < .001). Thus, the null hypothesis of no relationship between victim identity and seeing a problem with microaggression is rejected in favor of the alternative that a 109 bystander seeing a problem for microaggression is dependent on the victim’s identity. This relationship had a small effect size (Cramér’s V = .2064). INSERT TABLE 12 HERE Pairwise chi-square tests of victim identity and seeing a microaggression as a problem follow in Table 12. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Five of the comparisons met this criterion. Bystanders were significantly more likely to view a microaggression as a problem when the victim was an Asian man as opposed to a lesbian (χ2(1)= 12.829, p < .001, φ = .1954). Similarly, bystanders were significantly more likely to view a microaggression as a problem when the victim was an Asian woman rather than a gay man χ2(1)= 23.338, p < .001, φ = .2651), lesbian (χ2(1) = 34.288, p < .001, φ = .3248), or transgender woman (χ2(1) = 14.120, p < .001, φ = .2009). Bystanders were also significantly less likely to view a microaggression as a problem for a lesbian than a transgender man victim (χ2(1) = 12.025, p < .001, φ = -.1976). Per the φ values, these relationships all had small to small-medium effect sizes except for the comparison between an Asian woman victim and a lesbian victim which had a medium effect size. Next, Table 12 also contains the results of the omnibus chi-square test of independence between victim identity and whether a respondent viewed the hate incident as a problem. Over 90% of respondents saw the hate incident in their scenario as a problem except when the victim was a transgender woman. For the vignette involving a transgender woman as the victim, 85.42% of subjects saw the hate incident as a problem. The omnibus chi-square test showed that these proportions of seeing problems with hate incidents for each victim were significantly differently (χ2(5) = 14.302, p = .014). Thus, the null hypothesis of no relationship between victim 110 identity and seeing a problem with a hate incident is rejected in favor of the alternative that a bystander seeing a problem for a hate incident is dependent on the victim’s identity. This relationship had a small effect size (Cramér’s V = .1195). Pairwise chi-square tests of victim identity and seeing a microaggression as a problem follow. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Only one comparison fell within the significance threshold. Subjects assigned the vignette with an Asian woman victim were significantly more likely to view the hate incident as a problem that subjects assigned the vignette with a transgender woman victim (χ2(1) = 13.234, p < .001, φ = .1945). This relationship was small to medium strength with a φ of .1945. Table 12 then shows the results of the omnibus chi-square test of independence between victim identity and whether a respondent viewed the hate crime phase as a problem. Over 90% of respondents saw the hate crime in their scenario as a problem. All but one subject assigned to the Asian woman victim vignette considered the hate crime to be a problem. Conversely 13 respondents (6.77%) given the transgender woman victim vignette did not see the hate crime as a problem. The omnibus chi-square test showed that these proportions of seeing problems with hate crime for each victim did not significantly differ (χ2(5) = 8.782, p = .118). Thus, the null hypothesis of no relationship between victim identity and seeing a problem with hate crime failed to be rejected. Pairwise chi-square tests of victim identity and seeing a hate crime as a problem follow. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. 111 None of the pairwise comparisons achieved statistical significance. Thus, there was no relationship observed between victim identity and seeing a hate crime as a problem. Intervening for Each Phase Table 13 contains the results of the omnibus chi-square test of independence between victim identity and whether a respondent intervened after witnessing the microaggression. Respondents in the Asian woman hate event were most likely to intervene (54.43%), followed by those who had the Asian man (39.64%), transgender man (38.30%), transgender woman (36.46%), and gay man (33.91%) vignettes. Respondents in the lesbian hate event were least likely to intervene to a microaggression (22.75%). The omnibus chi-square test showed that these proportions of intervening for microaggressions for each victim type were significantly differently (χ2(5) = 36.278, p < .001). Thus, the null hypothesis of no relationship between victim identity and intervening for a microaggression is rejected in favor of the alternative that a bystander intervening for a microaggression is dependent on the victim’s identity. This relationship had a small effect size (Cramér’s V = .1904). INSERT TABLE 13 HERE Pairwise chi-square tests of victim identity and intervening for microaggressions follow. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Five comparisons fell within the significance threshold. Subjects assigned the vignette with an Asian man victim were significantly more likely to intervene for the microaggression than were subjects assigned the vignette with a lesbian victim (χ2(1) = 11.154, p = .001, φ = .1822). Respondents assigned the Asian woman victim vignette were significantly more likely to intervene for a microaggression than were respondents assigned the gay man, (χ2(1) = 14.177, p 112 < .001, φ = .2066) lesbian, (χ2 = 34.522, p < .001, φ = .3259) or transgender woman (χ2(1) = 11.332, p = .001, φ = .1799) victim scenarios. Lastly, bystanders to the microaggression against a trans man were significantly more likely to intervene than bystanders to the microaggression against a lesbian (χ2(1) = 8.8173, p = .003, φ = .1692). Per the obtained φ values, these relationships all had small effect sizes save for the comparison between Asian woman victim and lesbian victim bystanders which had a medium effect size. Table 13 further contains the results of the omnibus chi-square test of independence between victim identity and whether a respondent intervened after witnessing the hate incident phase. Respondents in the Asian woman hate event were most likely to intervene (81.65%), followed by those who had the Asian man (68.84%), gay man (67.24%), trans man (63.83%), lesbian (61.08%), and trans woman (53.65%) hate events. The omnibus chi-square test showed that these proportions of intervening for hate incidents for each victim type were significantly differently (χ2(5) = 32.808, p < .001). Thus, the null hypothesis of no relationship between victim identity and intervening for a hate incident is rejected in favor of the alternative that a bystander intervening for a hate incident is dependent on the victim’s identity. This relationship had a small effect size (Cramér’s V = .1810). Pairwise chi-square tests of victim identity and intervening for hate incidents follow. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Four comparisons fell within the significance threshold. All four significant relationships were between whether the victim was an Asian woman or a sexual or gender minority individual. Specifically, bystanders were significantly more likely to intervene during hate incidents involving an Asian woman victim than ones with a gay man (χ2 = 8.951, p = .003, φ = .1642), a 113 lesbian (χ2 = 16.707, p < .001, φ = .2267), a trans man (χ2 = 12.068, p = .001, φ = .2009), or a trans woman (χ2 = 30.406, p < .001, φ = .2947). In order of largest to smallest effect size, bystanders were thus significantly more likely to intervene in hate incidents with an Asian woman victim than those with a trans woman, lesbian, trans man, or gay man. Table 13 also contains the results of the omnibus chi-square test of independence between victim identity and whether a respondent intervened after witnessing the hate crime phase. Respondents in the Asian woman hate crime were most likely to intervene (91.14%), followed by those who had the Asian man (81.66%), lesbian (81.44%), trans man (80.14%), gay man (77.59%), and trans woman (72.92%) hate crimes. The omnibus chi-square test showed that these proportions of intervening for hate incidents for each victim type were significantly differently (χ2(5) = 19.632, p = .001). Thus, the null hypothesis of no relationship between victim identity and intervening for a hate crime is rejected in favor of the alternative that a bystander intervening for a hate crime is dependent on the victim’s identity. This relationship had a small effect size (Cramér’s V = .1400). Pairwise chi-square tests of victim identity and intervening for microaggressions follow. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Two comparisons fell within the significance threshold. Bystanders were significantly more likely to intervene during hate crimes involving an Asian woman victim than ones with a gay man (χ2(1) = 11.338, p = .001, φ = .1848) or a trans woman (χ2(1) = 18.810, p < .001, φ = .2318). Latent Class Distribution by Victim Table 14 contains the results of the omnibus chi-square test of independence between victim identity and which of the three latent classes from study 1 the respondent was best 114 classified as. Results for the same test but of the five-class solution are found in Tables A4 and A5 of the Appendix. The majority of respondents (53.80%) assigned into the Asian woman hate event were best classified as always interveners in class 1. Comparatively, 40.43% of those assigned to the trans man, 40.24% of those assigned to the Asian man, 37.50% of those assigned to the trans woman, 35.63% of those assigned to the gay man, and 25.15% of those assigned to the lesbian could be best classified as the always interveners. Respondents tended to next most frequently be assigned to the second class comprised of individuals whose progress through the situational model increased as the event became more severe (40.24% Asian man, 36.71% Asian woman, 39.66% gay man, 44.31% lesbian, 34.75% trans man, and 29.69% trans woman). However, more respondents assigned the trans woman vignette were classified into the third class of individuals who never intervened (32.81%) whereas respondents for all other victims least frequently were classified into that third class. The omnibus chi-square test showed that these proportions of class membership significantly differed by the victim with a small effect size (χ2(10) = 48.607, p < .001, Cramér’s V = .1558). INSERT TABLE 14 HERE Table 14 also contains the pairwise chi-square tests of victim identity and latent class membership. To correct for multiple comparisons, the Bonferroni correction is applied for the 15 pairwise chi-square tests, making the required alpha to reject the null hypothesis of no relationship 0.0033. Cramér’s V values are shown for the pairwise comparisons that retain statistical significance after the Bonferroni correction. Four comparisons fell within the significance threshold. In all four of these cases, the group being compared to was respondents to the Asian woman hate event. Latent classification of subjects significantly differed between those assigned to the Asian woman hate event and those assigned to the gay man (χ2(3) = 17.338, 115 p < .001, Cramér’s V = .2285), lesbian (χ2(3) = 35.913, p < .001, Cramér’s V = .3324), trans man (χ2(3) = 13.355, p = .001, Cramér’s V = .2113), and trans woman (χ2(3) = 27.581, p < .001, Cramér’s V = .2807) hate events. Discussion The results of study 2 for this dissertation indicate partial support for the hypothesis that respondents would be most likely to assist Asian Americans, followed by gay men and lesbians, and be least likely to assist transgender individuals. Bystanders to the Asian American woman microaggression were significantly more likely to see a problem compared to bystanders to the gay man, lesbian, and trans woman. Bystanders to the Asian man and trans man microaggressions also were significantly more likely to see a problem than those assigned to the lesbian microaggression. These differences began to fade as the hate event escalated. For the hate incident phase, subjects given the Asian woman scenario were significantly more likely to see a problem only compared to those witnessing the trans woman victim scenario. No significant differences were observed in seeing a problem with the hate crime phase among all victims. In terms of intervention, however, some significant differences persisted across all phases of the hate event. For the microaggression, hate incident, and hate crime phases, bystanders to an Asian woman victim were significantly more likely to intervene than bystanders to a gay man or trans woman. Likewise, bystanders to an Asian woman victim were significantly more likely to intervene than bystanders to a lesbian victim during the microaggression and hate incident phases. Subjects assigned to the Asian woman victim condition were also significantly more likely to intervene than bystanders to an Asian man and bystanders to a trans man during the microaggression phase were significantly more likely to intervene than bystanders to a lesbian victim. 116 Lastly, examination of the latent classes into which bystanders can best be categorized highlighted the differences in how bystanders react based upon the victim’s identity. The majority of bystanders to the Asian woman victim could be classified as always interveners. There were significant differences in these classifications between bystanders to the Asian woman victim and bystanders to the gay man, lesbian, trans man, and trans woman victims. Overall, partial support for hypothesis 2 was found given that the Asian woman received significantly more support than the sexual and gender minority victims per the latent classes, intervention during each phase, and seeing a problem for microaggression or hate incident. Moreover, bystanders were more likely to see a problem and intervene for the microaggression phase for an Asian man than a lesbian. No evidence was found to support that gay men or lesbians would receive more bystander progress through the situational model than trans men or trans women. In fact, bystanders to the trans man victim were significantly more likely to see a problem and intervene for a microaggression than were bystanders to the lesbian victim. The results of these analyses conform to theoretical expectations from social categorization theory. Given that little information was provided on the victim aside from their defining minority status, it was assumed that bystanders would infer information based upon stereotypes of the respective minorities. Although public opinion towards Asian Americans has fluctuated between positive and negative stereotypes, there typically remains an overall positive sentiment towards Asian Americans (Bonilla-Silva, 2004). For example, hate crimes against Asian Americans were down 33.7% in 2022 compared to 2021 as shown in Table A1, suggesting that public sentiment towards Asian Americans may be increasing in parallel with the waning severity of COVID-19. Sexual and gender minority individuals, however, have not elicited the same positive public perception that Asian Americans sometimes have had. Homophobic and 117 transphobic rhetoric has been growing, with Gallup surveys showing consequent increases in homophobic and transphobic attitudes by the U.S. public (Jones, 2023a). Thus, as stereotypes towards Asian Americans tend to be positive, the results support the theoretical expectations from social categorization theory as Asian Americans, and the Asian woman in the vignettes in particular, received significantly more bystander progress through the situational model than sexual and gender minority individuals. However, contrary to the hypothesis for this study, trans men were more likely to have subjects view microaggressions against them as problems and ultimately intervene compared to lesbians. Moreover, the results of this study conformed to the expectations set by the situational model. Significant differences were observed between various victim groups in terms of whether bystanders saw a problem or intervened. As previously noted, individuals often fail to recognize events as emergencies with sexual (Owuamalam & Matos, 2020) and racial minority victims (Gaertner et al., 1982). The results showed that sexual and gender minority victims in particular were the least likely to have their victimizations considered problems and the least likely to receive intervention compared to racial minority victims. These significant differences were most often observed between the Asian woman victim and sexual and gender minority victims. Though little research has examined discrepancies in bystander behavior by various victim minority statuses, the results comport with Vera and colleagues (2023) study. Per their study, college students were more likely to intervene for racial minority victims of bullying than sexual or gender minority victims. Likewise, the results of this study found that bystanders most frequently were significantly more likely to intervene for an Asian woman hate victim than gay man, lesbian, trans man, and trans woman hate victims. Moreover, few significant differences emerged between sexual and gender minority victim comparisons. These significant differences 118 were that bystanders for the trans man victim were more likely to see a problem and intervene for a microaggression than were bystanders for the lesbian victim. The results of this study illuminate the need for bystander intervention training programs to discuss intervention for many different types of victims. Bystanders clearly perceive hate events differently based upon victim’s identity. Therefore, these trainings must exhibit how hate impacts different types of minority groups to encourage bystanders to intervene for all regardless of the victim’s specific minority identity (see Kirk-Provencher et al., 2023 for a discussion on the exclusion of sexual and gender minority individuals in bystander trainings). Moreover, these trainings ought to include discussions as to how implicit bias can impact their cognition of events as recent research has shown its inclusion significantly improves the likelihood of bystanders intervening when observing discrimination (Stephens et al., 2023). Doing so can help to counteract the lack of progress through the situational model observed for sexual and gender minority victims compared to racial minority victims. This study possesses several limitations that may impact the validity of the results. First, it was assumed that respondents would use minority stereotypes due to the limited information provided to them about the victim aside from their minority status. However, no measurements were taken of how the respondents actually perceived these victims and to what extent, if at all, they attributed stereotypic attributes to the victims. Doing so may have revealed why the lesbian victim vignettes had the least number of respondents willing to see a problem and ultimately intervene during the microaggression phase. Straight men are more likely to help effeminate gay men than masculine gay men per Owuamalam and Matos (2022). For the current study, it is possible that, in a similar fashion, respondents summoned forth images of a masculine “butch” lesbian rather than a feminine “lipstick” lesbian (Jones, 2015). The perceived masculinity of the 119 former may have thereby reduced respondents’ situational model progress as they deemed her less worthy of help without the activation of chivalry by effeminacy. Lastly, the valence of the microaggression was not consistent for all minority types. The microaggression used in the Asian man and Asian woman vignettes was a positive one based upon the model minority view of Asian Americans as excelling in mathematics. This positive microaggression may have primed respondents with positive views of Asian Americans, thus arousing a greater obligation to help in comparison to the sexual and gender minority victims for whom negative microaggressions were used. What remains unknown is what caused bystanders overall to respond differently to the scenario. In other words, regardless of the victim presented in the scenario, why did some bystanders intervene at all phases whereas some did not even see problems for any of the phases? This will be explored next in study 3 of this dissertation. 120 Table 26. Chi-Square Tests of Victim Identity and Seeing a Problem Microaggression Problem Hate Incident Problem Hate Crime Problem n (%) n (%) n (%) Victim Identity Asian Man 136L (80.47%) 154 (91.12%) 158 (93.49%) Asian Woman 143GM, L, TW (90.51%) 153TW (96.84%) 157 (99.37%) Gay Man 120AW (68.97%) 160 (91.95%) 166 (95.40%) Lesbian 105AM, AW, TM (62.87%) 153 (91.62%) 159 (95.21%) Trans Man 114L (80.85%) 128 (90.78%) 133 (94.33%) Trans Woman 144AW (75.00%) 164AW (85.42%) 179 (93.23%) All Groups 762* (76.12%) 912* (91.11%) 952 (95.10%) Omnibus Tests χ2 42.644 14.302 8.782 Degrees of Freedom 5 5 5 p-value <.001 .014 .118 Cramér’s V .2064 .1195 .0937 Note: Chi-square tests of independence were conducted for each pairwise comparison. These tests are equivalent to proportion-z tests as the effect size, 𝜙, provides directionality for 2 x 2 chi-squares (Bradley & Cutcomb, 1977). The two-tailed hypothesis test done by chi-squares allows for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. The Bonferroni correction was applied to adjust for multiple comparisons when doing pairwise comparisons between the victim groups. A total of 15 pairwise comparisons were made for each dependent variable; therefore, the alpha for significance was p = .0033. The Bonferroni correction was not applied to the omnibus tests; therefore, the alpha for significance was p = .05. Superscripts denote significant differences between groups: *—Omnibus significance p < .05, **—Omnibus significance p < .01, ***—Omnibus significance p < .001, AM—Asian Man, AW—Asian Woman, GM—Gay Man, L—Lesbian, TM—Trans Man, TW—Trans Woman 121 Table 27. Chi-Square Tests of Victim Identity and Intervening Note: Chi-square tests of independence were conducted for each pairwise comparison. These tests are equivalent to proportion-z tests as the effect size, 𝜙, provides directionality for 2 x 2 chi-squares (Bradley & Cutcomb, 1977). The two-tailed hypothesis test done by chi-squares allows for the detection of results in the opposite direction of that specified by the hypothesis. Moreover, the two-tailed test has a more conservative alpha, reducing the likelihood of type I error in the results. The Bonferroni correction was applied to adjust for multiple comparisons when doing pairwise comparisons between the victim groups. A total of 15 pairwise comparisons were made for each dependent variable; therefore, the alpha for significance was p = .0033. The Bonferroni correction was not applied to the omnibus tests; therefore, the alpha for significance was p = .05. Superscripts denote significant differences between groups: *—Omnibus significance p < .05, **—Omnibus significance p < .01, ***—Omnibus significance p < .001, AM—Asian Man, AW—Asian Woman, GM—Gay Man, L—Lesbian, TM—Trans Man, TW—Trans Woman Microaggression Intervene Hate Incident Intervene Hate Crime Intervene n (%) n (%) n (%) Victim Identity Asian Man 67L (39.64%) 116 (68.64%) 138 (81.66%) Asian Woman 86GM, L, TW (54.43%) 129GM, L, TM, TW (81.65%) 144GM, TW (91.14%) Gay Man 59AW (33.91%) 117AW (67.24%) 135AW (77.59%) Lesbian 38AM, AW, TM (22.75%) 102AW (61.08%) 136 (81.44%) Trans Man 54L (38.30%) 90AW (63.83%) 113 (80.14%) Trans Woman 70AW (36.46%) 103AW (53.65%) 140AW (72.92%) All Groups 374* (37.36%) 657* (65.63%) 806* (80.52%) Omnibus Tests χ2 36.278 32.808 19.632 Degrees of Freedom 5 5 5 p-value <.001 <.001 .001 Cramér’s V .1904 .1810 .1400 122 Table 28. Chi-Squares of Victim Identity and Bystander Latent Class Membership Class 1 Class 2 Class 3 Class Description Always interveners Intervene with escalating severity Never interveners n (%) n (%) n (%) Victim Identity Asian Man 68 (40.24%) 68 (40.24%) 33 (19.53%) Asian Woman 85GM, L, TM, TW (53.80%) 58 (36.71%) 15 (9.49%) Gay Man 62AW (35.63%) 69 (39.66%) 43 (24.71%) Lesbian 42AW (25.15%) 74 (44.31%) 51 (30.54%) Trans Man 57AW (40.43%) 49 (34.75%) 35 (24.82%) Trans Woman 72AW (37.50%) 57 (29.69%) 63 (32.81%) All Groups 386* (38.56%) 375 (37.46%) 240 (23.98%) Omnibus Test χ2 48.607 Degrees of Freedom 10 p-value <.001 Cramér’s V .1558 Note: Chi-square tests of independence were conducted for each pairwise comparison of LCA class membership for all three classes and each pair of victim groups. The Bonferroni correction was applied to adjust for multiple comparisons when doing pairwise comparisons between the victim groups. A total of 15 pairwise comparisons were made for each dependent variable; therefore, the alpha for significance was p = .0033. The Bonferroni correction was not applied to the omnibus test; therefore, the alpha for significance was p = .05. Superscripts denote significant differences between groups: *—Omnibus significance p < .05, **—Omnibus significance p < .01, ***—Omnibus significance p < .001, AM—Asian Man, AW—Asian Woman, GM—Gay Man, L—Lesbian, TM—Trans Man, TW—Trans Woman 123 Chapter 10: Study Three Results In this chapter, the theoretical background informing the study designed to answer research question 3 will be reiterated. This includes a summary of bystander traits known to impact intervention, a brief overview of Latané and Darley’s (1970) situational model of bystander intervention, and a reiteration of how mechanisms of social categorization theory explain how bystanders’ self-identification can impact their decision-making. Logistic regression models are then presented, starting with whether bystanders saw a problem for each hate event phase then finishing with whether bystanders intervened for each hate event phase. A multinomial logistic regression is then shown to test how bystander traits impact their best classification into the latent classes of the three-class solution from study 1 of this dissertation. Review of Bystander Traits Affecting Intervention As with victim traits in study 2, bystander traits also impact whether bystanders intervene. These bystander traits most prominently include empathy (Nickerson et al., 2015; Secord Fredrick et al., 2020), bystander efficacy (Banyard, 2008; Banyard & Moynihan, 2011; Banyard et al., 2004), bystander decision-making (Jensen & Raver, 2020; Morgan, 1978; Shea et al., 2021), and demographic traits (Mainwaring et al., 2023; Ratcliff et al., 2023). Nickerson and colleagues (2015) found empathy significantly predicted whether bystanders intervened in response to bullying. They noted specifically that both affective empathy through the recognition of the victim’s suffering, as well as cognitive empathy through considering the victim’s perspective, created the desire to help. However, affective empathy is most responsible for compelling bystanders to act whereas cognitive empathy helps them to recognize and know how to intervene (Secord Fredrick et al., 2020). 124 Bystander efficacy, which refers to the confidence an individual has in their own ability to intervene, has been shown to positively predict bystander intervention across a host of situations from general emergencies (Latané & Nida, 1981) to sexual violence (Moynihan et al., 2011) and robberies (Huston et al., 1981). Several studies by Banyard have also shown that bystander efficacy significantly predicts both intentions to intervene and reported actions in real scenarios (Banyard, 2008; Banyard & Moynihan, 2011; Banyard et al., 2004). As most bias incidents do not initially rise to the level of requiring police intervention (e.g., starting with prejudicial remarks), bystander efficacy becomes especially necessary for intervention to occur (Schafer & Navarro, 2003). Bystander decision-making has been studied in terms of how bystanders weigh the costs and benefits of intervention. Morgan (1978) noted decisional factors included the net benefit to the individual and the net benefit to the group that would come from intervening. Upstanders would consider the benefits to outweigh the costs of intervention, thus deciding to intervene. Conversely, non-interveners would consider the costs to outweigh the benefits (Shea et al., 2021). The factors that bystanders weigh include relational, and personal costs and benefits of intervening as opposed to those costs and benefits attending doing nothing (Jensen & Raver, 2020). Banyard and Moynihan (2011) found that decisional balance in favor of intervention is an enduring individual trait that promotes intervention. Bystander demographics such as race, gender, and sexual orientation can impact bystander behavior as well. Ratcliff and colleagues (2023) found that white bystanders were significantly less likely to intervene than Asian American bystanders in instances of anti-Asian prejudice. Whites often fail to recognize prejudice as they have not been the targets of it (Ashburn-Nardo et al., 2008). Even when they do recognize instances of discrimination, whites 125 also tend to attribute less prejudice to the event than racial minority individuals do (Czopp, 2010). These findings are particularly relevant as to whether bystanders see problems or the need to intervene for microaggressions given they may be unfamiliar with microaggressions. Bystander gender also affects proclivity to intervene. Women are significantly more likely to intervene when witnessing sexual violence scenarios (Mainwaring et al., 2023), workplace incivility (Sinclair, 2021), and non-emergency situations (Cox & Adam, 2018) compared to men. Sexual minority individuals were significantly more likely to intervene when listening to audio vignettes of intimate partner violence in both same- and opposite-sex cases of intimate partner violence than were heterosexual individuals (Graham et al., 2023). Trans women and gender non-conforming individuals have reported seeing more opportunities to intervene in sexual assault situations but are significantly less likely to do so than cisgender women (Hoxmeier et al., 2022). Intersectionality in bystander intervention studies represents a considerable dearth within the bystander intervention literature. Few studies have examined how race and gender interact to impact bystander behavior or willingness to intervene. Burns and colleagues (2019) explored how race and gender impact intent to intervene in college campus sexual assault situations. Before being subjected to a bystander intervention program, white men were least likely to intervene compared to Latino men, Latina women, Black men, Black women, and white women. Moreover, the processes impacting bystander intervention also differentially affect groups. Black men, for example, report more instances of intervening when they have peer norms conducive to intervening while this relationship did not hold for white men, white women, or Black women (Brown et al., 2014). 126 Given the lack of research on intersectionality, despite the vast literature demonstrating how demographic traits impact bystander behavior, the research question explored in the current study is: Research Question 3: Do bystander characteristics influence their bystander decision-making processes while witnessing hate? Theoretical Background Within the situational model, various studies have exemplified how bystander traits impact their progress through the model. Majority group members may fail to even perceive a microaggression has occurred as they need contextual familiarity to know what microaggressions are. Onlookers often fail to detect microaggressions against racial/ethnic, sexual, and religious minorities (Lilienfeld, 2017), which is especially important as insults escalate into violence (Whitehead et al., 2018). Individual level traits, such as personality, mood, and the self-persuasiveness of the bystander, also influence whether the bystander sees the event as an emergency. For instance, when witnessing the beginnings of hate crimes, individuals may perceive prejudicial statements as jokes rather than emergencies (Katz et al., 2019a, 2019b, 2022). Moreover, even individuals who consider themselves to be socially liberal fail to recognize microaggressions (McClure, 2020). Majority group members typically respond less quickly to emergencies with minority group victims (Gamberini et al., 2015; Kunstman & Plant, 2008) and demonstrate less empathy toward them (Mazurczak & Godzisz, 2019). Conversely, bystanders are most likely to intervene when the victims appear similar to themselves due to in-group identification (Levine et al., 2002). 127 Social categorization theory again can be used to explain this phenomenon. Turner and colleagues (1987) developed social categorization theory, positing that there are three levels in which individuals categorize themselves. These levels consist of human identity, social identity, and personal identity. Human identity refers to a person’s self-identification as a human being and member of humanity. In the second level, social identity, individuals consider themselves members of a unique social group opposite of other social groups. Lastly, someone’s personal identity is how that individual identifies as a member of that social group in relation to other members of the same group. Of note, as one of these levels becomes more salient in a given context, the other levels then recede and lose prominence. Once self-categorized into a social group, depersonalization then occurs (Turner et al., 1987). In other words, people perceive stereotypic patterns between category members and begin to see members as interchangeable, no longer seeing themselves as unique individuals. These patterns include the behaviors, attitudes, and emotions attributed to the group. Individuals then begin to act in accordance with what they believe other group members would do in a given situation. Per Mullen et al. (1992) and Brewer (1999), ingroup favoritism and outgroup derogation occur subsequent to social categorization. These attitudes are not mutually exclusive. According to Brewer (1999), these two beliefs can occur together, coalescing into prejudice in which ingroup members are treated favorably whereas outgroup members are treated unfavorably. The larger the ingroup, the stronger the depersonalization toward outgroups. This depersonalization most often leads to avoidance of outgroup members, but it can also lead to contempt and violence. Within the context of hate events and bystander intervention, ingroup members may 128 abstain from intervening. Doing so is their way of signaling disapproval or apathy towards the outgroup victim. The extent of each bystander’s progress through the situational model may therefore vary based upon whether the bystander belongs to an outgroup or not. Thus, the hypothesis for the current study is: Hypothesis 3: Respondents who are cisgender, heterosexual, non-Hispanic white men will not progress as far through the bystander intervention model as members of out-groups. Results Seeing a Problem for Each Phase INSERT TABLE 15 HERE Table 15 contains the results of the logistic regression models5 for seeing a problem during the microaggression, hate incident, and hate crime phases. The model used to predict seeing a problem during the microaggression phase was statistically significantly predictive (χ2(7) = 85.61, p < .0001). Four variables achieved statistical significance: empathy (p = .004) , bystander behavior (p = .001), other-regarding (p = .034), and favorability toward the assigned victim group (p < .001). For every additional point on the empathy scale, respondents were 1.0449 times more likely to see a problem with the microaggression. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.1755 times more likely to see a problem with the microaggression. Respondents who 5 Interaction terms such as the one used to explore how majority group membership impacts bystander intervention are best used in OLS regression models as opposed to logistic regression models (Mize, 2019). Specifically, interaction terms cannot be interpreted directly in logistic regression models and must therefore be assessed using predicted probabilities. Although it was not calculated as an interaction term, it is akin to one since it was formulated as an individual being non-Hispanic and white and heterosexual and cisgender male. Logistic regression models are used throughout this chapter because of the assumed readers’ familiarity with using logistic regression for binary dependent variables. The results of the OLS regression models are located in the appendix as supplementary tables in Table A4 and Table A5. The results did not substantively differ between the logistic regression and OLS regression models in terms of which variables achieved statistical significance. 129 showed regard for others on the dictator and ultimatum games were 1.4683 times more likely to see a problem for the microaggression than those who did not. Subjects who indicated feeling favorably toward their assigned victim group were 2.989 times more likely to see a problem than those who felt unfavorably. These odds ratios signify that empathy and bystander behavior had weak to small effect sizes, other-regarding had a small effect size, and favorability toward the assigned victim group had a large effect size (Olivier & Bell, 2013). The model for seeing a problem during the hate incident phase similarly was significantly predictive (χ2(7) =76.65, p < .0001). Empathy (p < .001), bystander behavior (p = .017), and favorability toward the assigned victim group (p < .001) all significantly predicted whether a respondent saw a problem for this phase. For every additional point on the empathy scale, respondents were 1.0988 times more likely to see a problem with the hate incident. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.1952 times more likely to see a problem with the hate incident. Subjects who indicated feeling favorably toward their assigned victim group were 4.3145 times more likely to see a problem than those who felt unfavorably. These odds ratios signify that empathy and bystander behavior had weak to small effect sizes, and favorability toward the assigned victim group had a large effect size (Olivier & Bell, 2013). Lastly, the model for predicting whether a subject saw a problem with the hate crime phase was statistically significant (χ2(7) = 46.63, p < .0001). Only empathy ( p = .011) and favorability toward the assigned victim group (p < .001) significantly predicted seeing a problem for the hate crime phase. For every additional point on the empathy scale, respondents were 1.0798 times more likely to see a problem with the hate crime. Subjects who indicated feeling favorably toward their assigned victim group were 4.9035 times more likely to see a problem 130 than those who felt unfavorably. These odds ratios signify that empathy had a weak to small effect sizes whereas favorability toward the assigned victim group had a large effect size (Olivier & Bell, 2013). Thus, across all phases of the hate event, empathy and favorability toward the assigned victim group significantly predicted whether a respondent saw a problem. Empathy consistently had a weak to small effect size while favorability toward the assigned victim group consistently had a large effect size. Responses on the bystander behavior scale significantly predicted seeing a problem for the microaggression and hate incident phases. Other-regarding as determined by the dictator and ultimatum games only significantly predicted seeing a problem with the microaggression. Bystander efficacy, the average amount of money in $1,000 respondents would give in the dictator and ultimatum games, and whether the respondent was a non-Hispanic white, heterosexual, cisgender man never reached statistical significance. The hypothesis that majority group members would be less likely to see a problem was therefore not supported. Intervening for Each Phase Given that majority group members were not significantly less likely to see a problem for any of the hate phases but other variables emerged as significant predictors, any differences in terms of intervening for each phase were then tested. Table 16 contains the results of the logistic regression models for intervening during the microaggression, hate incident, and hate crime phases. The model used to predict intervention during the microaggression phase was statistically significant (χ2(7) =144.49, p < .0001). Bystander behavior (p < .001), bystander efficacy (p < .001), and favorability toward the assigned victim group (p < .001) all significantly predicted whether subjects would intervene. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.2887 times more likely 131 to intervene. Each additional point on the bystander efficacy scale was associated with respondents being 1.0690 times more likely to intervene. Subjects who indicated feeling favorably toward their assigned victim group were 4.484 times more likely to intervene when witnessing a microaggression than those who felt unfavorably. The obtained odds ratios indicate that bystander behavior had a small effect size, bystander efficacy had a weak to small effect size, and favorability toward the assigned victim group had a large effect size (Olivier & Bell, 2013). INSERT TABLE 16 HERE The model used to predict intervention during the hate incident phase was statistically significant (χ2(7) =223.82, p < .0001). Empathy (p = .002), bystander behavior (p < .001), bystander efficacy (p < .001), and favorability toward the assigned victim group (p < .001) significantly predicted whether respondents would intervene during the hate incident. For every additional point on the empathy scale, respondents were 1.0470 times more likely to intervene for the hate incident. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.3320 times more likely to intervene for the hate incident. Each additional point on the bystander efficacy scale was associated with respondents being 1.0728 times more likely to intervene. Subjects who indicated feeling favorably toward their assigned victim group were 5.2022 times more likely to intervene than those who felt unfavorably. These odds ratios indicate empathy and bystander efficacy had weak to small effect sizes, bystander behavior had a small effect size, and favorability toward the victim group had a large effect size (Olivier & Bell, 2013). The model used to predict intervention during the hate crime phase was statistically significant (χ2(7) =151.36, p < .0001). Empathy (p < .001), bystander behavior (p < .001), 132 bystander efficacy (p < .001), being other-regarding (p = .023), and favorability toward the assigned victim group (p < .001) significantly predicted whether respondents would intervene when witnessing a hate crime. For every additional point on the empathy scale, respondents were 1.0844 times more likely to intervene for the hate crime. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.1951 times more likely intervene for the hate crime. Each additional point on the bystander efficacy scale was associated with respondents being 1.0566 times more likely to intervene. Respondents who showed regard for others on the dictator and ultimatum games were 1.6126 times more likely to intervene for the hate crime than those who did not. Subjects who indicated feeling favorably toward their assigned victim group were 3.8233 times more likely to intervene than those who felt unfavorably. These odds ratios indicate empathy, bystander behavior, and bystander efficacy had weak to small effect sizes, other-regarding had a small to medium effect size, and favorability toward the victim group had a large effect size (Olivier & Bell, 2013). Bystander behavior, bystander efficacy, and favorability toward the victim’s group significantly predicted intervention for each hate event phase. Empathy significantly predicted intervening during the hate incident and hate crime phases. Other-regarding significantly predicted intervening during the hate crime phase. As with the models for considering a phase to be a problem, the effect sizes for these relationships tended to be consistent between models. Bystander efficacy possessed a weak to small effect size in all models while favorability toward the assigned victim group had large effect sizes. The relationship between empathy and intervention had weak to small effect sizes for both the hate incident and hate crime phases. Bystander behavior changed from having small effect sizes for intervening during the microaggression and hate incident phases to having a weak to small effect size for the hate crime 133 phase. Other-regarding had a medium to large effect size for predicting intervention during the hate crime. The average amount given on the dictator and ultimatum games as well as whether the bystander was a majority group member failed to significantly predict intervening for any phase. Therefore, he hypothesis that majority group members would be less likely to intervene was not supported. Latent Class Membership Given how these bystander variables significantly predicted seeing a problem and intervening, a multinomial logistic regression was then conducted on the three-class solution membership of each respondent to determine the influence of these variables across all hate event phases at once. The results of this multinomial logistic regression are presented in Table 17. The model was significantly predictive of latent class membership in the three-class solution (χ2(7) =270.36, p < .0001). INSERT TABLE 17 HERE When using class 1, the always interveners, as the reference group, significant differences emerged in the bystander variables used in the model. When examining class 2, the individuals who progressed further through the situational model as the event became more severe, four variables significantly predicted membership into that class over the always interveners: empathy (p = .013), bystander behavior (p < .001), bystander efficacy (p < .001), and favorability toward the assigned victim group (p = .002). For every additional point on the empathy scale, respondents were 1.0361 times more likely to be best classified into class 2 as opposed to class 1. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 0.8238 times as likely to be best classified into class 2 as opposed to class 1. Each additional point on the bystander efficacy scale was associated with 134 respondents being 0.9361 times as likely to be best classified into class 2 as opposed to class 1. Subjects who indicated feeling favorably toward their assigned victim group were 0.4163 times as likely to be best classified into class 2 as opposed to class 1. These relative risk ratios represent that empathy and bystander efficacy had weak to small effect sizes, bystander behavior had a small effect size, and favorability to the assigned victim group had a medium to large effect size (Olivier & Bell, 2013). Furthermore, empathy (p = .005), bystander behavior (p < .001), bystander efficacy (p < .001) and favorability to the assigned victim group (p < .001) significantly predicted whether a respondent could be best classified into class 3 as opposed to class 1. For every additional point on the empathy scale, respondents were .9501 times as likely to be best classified into class 3 as opposed to class 1. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 0.6799 times as likely to be best classified into class 3 as opposed to class 1. Each additional point on the bystander efficacy scale was associated with respondents being 0.9169 times as likely to be best classified into class 3 as opposed to class 1. Subjects who indicated feeling favorably toward their assigned victim group were 0.1084 times as likely to be best classified into class 3 as opposed to class 1. These relative risk ratios represent that empathy and bystander efficacy had weak to small effect sizes, bystander behavior had a small effect size, and favorability to the assigned victim group had a large effect size (Olivier & Bell, 2013). Lastly, respondents being best classified as class 2, those who progressed further through the situational model as the event became more severe, were examined against respondents best classified as class 3, those who never intervened. Empathy (p < .001), bystander behavior (p < .001), and favorability to the assigned victim group (p < .001) significantly predicted whether a 135 respondent could be best classified into class 2 as opposed to class 3. For every additional point on the empathy scale, respondents were 1.0906 times more likely to be best classified into class 2 as opposed to class 3. For each additional event on the bystander behavior scale for which respondents reported indicating their distaste, they were 1.2117 times more likely to be best classified into class 2 as opposed to class 3. Subjects who indicated feeling favorably toward their assigned victim group were 3.8388 times more likely to be best classified into class 2 as opposed to class 3. These relative risk ratios represent that empathy had a weak to small effect size, bystander behavior had a small effect size, and favorability to the assigned victim group had a large effect size (Olivier & Bell, 2013). Overall, empathy, bystander behavior, and favorability to the assigned victim group consistently distinguished membership among the three latent classes. These commonly occurred in the expected direction, with the exception of class 2 members scoring higher on the empathy scale than class 1 members. Neither of the decision-making measures, average on the dictator and ultimatum games and other-regarding, distinguished between classes. Moreover, majority group membership did not significantly differentiate between class membership. Thus, there was no evidence to support the hypothesis that non-Hispanic white, heterosexual, cisgender men were significantly less likely to engage with the situational model than minority group members. Sensitivity Analyses Sensitivity analyses were conducted using main effects for each of the demographic traits rather than the combined term. Dummy variables for whether an individual was Hispanic, white, cisgender, heterosexual, or male (which included both cisgender and transgender men) were used. The results of these analyses are contained within Tables A7, A8, and A9 of the Appendix. Bystander demographic traits rarely were statistically significant in any of the models. Males 136 were significantly less likely than females to see a problem with the hate incident phase (p = .043). Heterosexuals were significantly less likely to intervene for the microaggression (p = .002) and hate crime phases (p = .024) compared to sexual minority individuals. Whites as opposed to non-whites were significantly more likely to intervene during the hate incident (p = .045) and hate crime (p = .008) phases. For latent class membership, whites as opposed to non-whites were significantly more likely to belong to class 2 comprised of individuals who progressed further through the situational model than class 3, which contained individuals who never intervened (p = .006). Heterosexual individuals as opposed to sexual minority individuals were significantly more likely to be best classified as class 2 (p = .03) and class 3 (p = .005) compared to class 1 which contained the always interveners. Neither being Hispanic as opposed to non-Hispanic nor cisgender as opposed to transgender differentiated bystanders on any of the outcomes (seeing a problem, intervening, or belonging to a different latent class). Stratification Analyses Several stratifications of the sample were tested for each of the dependent variables. These stratifications were conducted by victim group (Asian man, Asian woman, gay man, lesbian, trans man, and trans woman), 2020 election results for the respondent’s resident state (Trump 2020 state or Biden 2020 state), white respondents, and non-white respondents. The general results of these stratifications are located below in Tables 18 (seeing a problem), 19 (intervening), and 20 (latent class membership). INSERT TABLE 18 HERE In terms of seeing a problem for microaggression, only four models retained significance for empathy (Asian man victim, trans woman victim, Biden state resident, non-whites), three for bystander behavior (trans woman victim, Biden state resident, and whites), one for other- 137 regarding (Biden state resident), and seven for favorability toward the victim group (Asian man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). Moreover, bystander efficacy gained statistical significance for the gay man model. For seeing a problem during the hate incident phase, five models retained significance for empathy (Asian man victim, Trump state resident, Biden state resident, whites, and non-whites), three for bystander behavior (trans woman, Trump state resident, and Biden state resident), and eight for favorability toward the victim group (Asian man victim, gay man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). No other variables gained statistical significance in any of the models. For seeing a problem during the hate crime phase, three models retained significance for empathy (trans woman victim, Trump state resident, and whites), and four for favorability toward the victim group (Asian man victim, trans man victim, Trump state resident, and Biden state resident). Bystander behavior gained statistical significance for respondents to the trans woman victim vignette. The average amount given on the dictator and ultimatum games gained statistical significance for the lesbian and trans man victim vignette respondents. Being a non-Hispanic white, heterosexual, cisgender man gained statistical significance for the model of trans man victim vignette respondents. INSERT TABLE 19 HERE For intervening during the microaggression, seven models retained significance for bystander behavior (gay man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites), eight for bystander efficacy (Asian man victim, Asian woman victim, gay man victim, lesbian victim, trans man victim, Trump state 138 resident, Biden state resident, and whites), and seven for favorability toward the victim group (gay man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). Other-regarding gained statistical significance for non-white respondents. For intervening during the hate incident phase, three models retained significance for empathy (trans man victim, Trump state resident, and whites), seven for bystander behavior (gay man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites), seven for bystander efficacy (Asian man victim, gay man victim, lesbian victim, trans woman victim, Trump state resident, Biden state resident, and whites), and seven for favorability towards the victim group (gay man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). For intervening during the hate crime phase, seven models retained significance for empathy (Asian man victim, trans man victim, trans woman victim, Biden state resident, Trump state resident, whites, and non-whites), six for bystander behavior (gay man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites), five for bystander efficacy (Asian man victim, trans woman victim, Biden state resident, whites, and non-whites), one for other-regarding (trans woman victim), and nine for favorability toward the victim group (Asian man victim, gay man victim, lesbian victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). Being a non-Hispanic white, heterosexual, cisgender man gained statistical significance for the model of trans woman victim vignette respondents. In terms of being best classified into class 2 as opposed to class 1, one model retained significance for empathy (Biden state resident), four for bystander behavior (trans woman victim, 139 Trump state resident, Biden state resident, and whites), eight for bystander efficacy (Asian man victim, Asian woman victim, gay man victim, lesbian victim, trans man victim, Trump state resident, Biden state resident, and whites), and five for favorability toward the victim group (trans man victim, trans woman victim, Trump state resident, Biden state resident, and whites). Other-regarding gained statistical significance for trans woman victim respondents. INSERT TABLE 20 HERE In terms of being best classified into class 3 as opposed to class 1, four models retained significance for empathy (Asian man, trans man, Trump state resident, and non-whites), eight for bystander behavior (gay man victim, lesbian victim, trans woman victim, trans man victim, Trump state resident, Biden state resident, whites, and non-whites), six for bystander efficacy (Asian man victim, gay man victim, lesbian victim, Trump state resident, Biden state resident, and whites), and nine for favorability toward the victim group (Asian man victim, gay man victim, lesbian victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites). Lastly, in terms of being best classified into class 2 as opposed to class 3, seven models retained significance for empathy (Asian man victim, trans man victim, trans woman victim, Trump state resident, Biden state resident, whites, and non-whites), five for bystander behavior (trans man victim, Trump state resident, Biden state resident, whites, and non-whites), and six for favorability toward the victim group (Asian man victim, gay man victim, trans man victim, Trump state resident, Biden state resident, and whites). Bystander efficacy and other-regarding gained significance in the model for trans woman victim vignette respondents. Being a non-Hispanic white, heterosexual, cisgender man gained statistical significance for the model of gay man victim vignette respondents. 140 Overall, variables that retained significance from the main model were in the expected direction. However, some of the variables that gained significance from the main model were not in the expected direction. Specifically, the higher the average given on the dictator and ultimatum games for the lesbian and trans woman victim vignette models, the less likely the respondent was to see a problem with hate crime. Further, other-regarding when comparing class 2 to class 1 membership for the trans woman victim vignette model was not in the expected direction. Moreover, being a non-Hispanic white, heterosexual, cisgender man for the hate incident intervention model for the Asian man victim vignette, the intervention model for hate crime with a trans woman victim, and class 2 as opposed to class 3 membership for the gay man victim scenario were all in the direction opposite of what was expected. In other words, majority group members were significantly more likely to intervene for the Asian man during the hate incident as well as the trans man during the hate crime. Majority group members also were more likely to belong to class 2 over class 3. Regarding the overall meaningfulness of the variables in the models compared to their significance in the main models, empathy retained significance for 34 out of 80 models, bystander behavior retained significance for 43 out of 80 models, bystander efficacy retained significance in 34 out of 50 models, other-regarding retained significance for two out of 20 models, and favorability toward the victim retained significance in 62 out of 90 models. Thus, favorability toward the victim’s minority group appears to be the most consistent indicator of bystander intervention followed by bystander efficacy, bystander behavior, empathy, and other-regarding. Eighteen of the models did not substantively differ from the main results. In terms of the outcome, the distribution was that seeing a problem had six models that did not substantively 141 differ (microaggression-1 model, hate incident-2 models, hate crime-3 models), intervening had six models that did not substantively differ (microaggression-5 models, hate incident-1 model, hate crime-0 models), and latent class membership had six models that did not substantively differ (class 3 vs class 1-1 model, class 3 vs class 1-1 model, and class 2 vs class 3-4 models). Intervening for the microaggression phase was therefore the least likely to change by stratification. Pertaining to the stratification performed, there were a total of nine models per individual stratification. Four models did not substantively differ by victim type (Asian man-0 models, Asian woman-0 models, gay man-1 model, lesbian-0 models, trans man-2 models, trans woman-1 model), 2020 election results by state residency did not substantively differ for 11 models (Trump 2020 state residents-5 models, Biden 2020 state residents-6 models), and three models by race did not substantively differ (whites-3 models, and non-whites-0 models). Therefore, state residency by 2020 election results appeared to least affect the results followed by respondent race and victim group. Discussion In sum, the main results of study 3 for this dissertation showed no support for the hypothesis that respondents who are cisgender, heterosexual, non-Hispanic white men would be less likely to see problems, intervene, and fall within latent classes that intervened. These individuals did not significantly differ from bystanders who belong to minority groups (e.g., women, transgender individuals, sexual minority individuals, racial minority individuals) for any of the analyses conducted. These included whether the respondent saw the phase as a problem, whether the respondent intervened for the phase, and the latent class membership of the respondent. When examining main effects in the sensitivity analysis, race, sexual orientation, and 142 gender significantly predicted a few of the outcomes. Males were significantly less likely than females to see a problem with the hate incident phase. Heterosexuals were significantly less likely to intervene for the microaggression and hate crime phases compared to sexual minority individuals. Heterosexual individuals were also significantly less likely to belong to the class of always interveners compared to their likelihoods of belonging to classes 2 and 3. Whites as opposed to non-whites were significantly more likely to intervene during the hate incident and hate crime phases. Thus, the results of the sensitivity analysis provided conflicting evidence regarding hypothesis 3 as gender and sexual orientation were in the predicted directions whereas race was not. Regarding the other bystander variables, the results typically fell within the expected directions. The higher a respondent’s empathy, the more likely they were to see problems during the microaggression, hate incident, and hate crime phases. Moreover, these individuals were more likely to intervene for hate incidents and hate crimes. They were also more likely to belong to the classes that intervened as opposed to class 3 which did not intervene for any phase. Respondents who reported having previously intervened per the bystander behavior scale were significantly more likely to see problems during the microaggression and hate incident phases, intervene for the hate incident and hate crime phases, and be best classified as always interveners. Bystander efficacy did not predict whether a bystander found any of the phases to be a problem. It did, however, predict whether a bystander intervened during all of the hate event phases and whether they belonged to the classes of always interveners or those who intervened as the situation escalated compared to those who never intervened. Other-regarding as a decision-making process only differentiated whether respondents found microaggressions to be problems or intervened for hate crimes. Most importantly, whether bystanders felt favorably or 143 unfavorably toward the victim’s group significantly predicted whether they would find the hate event phases to be problems, intervened for them, and belonged to the classes of always interveners or those who intervened as the situation escalated compared to those who never intervened. This was the only variable that had a large effect size whereas all the other variables only had weak to small effect sizes. The results typically conformed with the prior literature. While empathy did not always predict intervening, it was predictive of intervening for more overt forms of hate such as hate incidents and hate crimes. These findings are consistent with Nickerson et al.’s (2015) meta-analysis that empathy predicts intervening for bullying. Interestingly, respondents best classified as those who intervened as the situation escalated were more empathic than those who always intervened. However, this finding had a weak effect size and may have been a statistical artifact with a p-value of .045 whereas nearly all other p-values were less than .01 for empathy. Likewise, having previously indicated distaste for a variety of discriminatory events in the bystander behavior scale significantly predicted seeing problems with microaggressions and hate incidents as well as intervening for all phases and being best classified into classes 1 and 2. While it did not predict seeing problems with hate crimes, this may be due to its operationalization. The scale used instances of racist, sexist, and homophobic remarks as opposed to violence. Therefore, it may have a threshold effect in that physical violence supersedes past histories of intervening for verbal comments. In a similar manner, bystander efficacy significantly predicted intervening for all the hate event phases and for belonging to the class that always intervened but not for seeing problems with the phases. This again may be due to the variable’s operationalization in that the items focused on whether one believed they could intervene in a violent situation and whether they 144 believed intervening would be effective for stopping such incidents. Since the scale did not focus on the early stages of the situational model such as seeing incidents as problems, it is unsurprising that it could only differentiate between later stages of the situational model like intervening. The two decision-making variables based upon a similar construct of regarding others most frequently failed to reach statistical significance. The average amount given during the dictator games never significantly distinguished between outcomes on any of the dependent variables. Other-regarding when making decisions only distinguished between seeing a problem with a microaggression and intervening for hate crime. These results suggest that decision-making for bystander intervention operates differently from decision-making in other contexts. Lastly, feeling favorably as opposed to unfavorably toward the assigned victim group consistently predicted seeing a problem, intervening, and belonging to classes that intervened across all possible dependent variables. These results conform to prior findings in the bystander literature as well as the theoretical expectations of social categorization theory in that positive feelings toward a group increase the likelihood of assisting. These findings carry several policy implications. First, the findings that empathy increased intervention for all phases denotes its importance. Empathy can be fostered from an early age through socioemotional learning curricula in early childhood education (Papieska et al., 2019). Lawmakers and politicians, such as Florida’s Governor DeSantis, have recently made socioemotional learning a target because of its link to critical race theory, banning it from the classroom (Wong, 2023). Considering that much of empathy development occurs in early childhood (Decety & Michalska, 2010), the use of socioemotional learning needs to be reinstated and reinforced to increase the likelihood of bystander intervention in hate-based scenarios. 145 In a similar vein, the favorability towards outgroups significantly predicted seeing problems and intervening for all phases in addition to classifications as members of the latent classes that intervened presents an important policy implication. The “Don’t Say Gay” and “Don’t Say Trans” laws recently instituted in some states have led to large decreases in public sentiment towards gay and trans individuals, especially among Republicans (Jones, 2023b). These laws seemingly reinforce negative stereotypes of gay and trans folks. The elimination of these laws should likewise signal acceptance towards them and hopefully reverse this trend in public sentiment. Doing so would be conducive toward bystanders helping gay and trans victims of hate. The effects of bystander efficacy as well as having histories of intervening per the bystander behavior scale show the importance of bystander training. A review of bystander program effectiveness found that nine out of ten programs reviewed significantly improved bystander efficacy (Bell et al., 2019). Moreover, seven out of nine programs significantly improved bystander behavior. The statistical significance of these variables within the various models of this chapter highlights their importance. Thus, additional funding should be given toward bystander intervention programs to improve bystander efficacy and bystander behavior to consequently improve bystander response to hate. As with all studies, this study possessed a few limitations. First, the sample size may not have been large enough to detect the effect of cisgender, heterosexual non-Hispanic white men on any of the outcome variables. Such a variable in the model constitutes an interaction effect between several variables, which exponentially increases the sample size required to detect an effect. Reducing the likelihood of committing a type II error would require thousands of participants for such an interaction effect (Brysbaert, 2019). Second, empathy was unable to be 146 examined in terms of affective versus cognitive empathy. The confirmatory factor analysis noted in the methods section revealed that, while there were two factors for the empathy scale, these distinguished only the straight coded and reverse coded items. Third, the inclusion of the bystander behavior scale was tautological in that respondents indicated if they previously intervened by voicing their distaste at hearing sexist, racist, and homophobic remarks. While past history of intervention is important to note for the willingness of bystanders to intervene, it nevertheless represents a tautology present within the models as they were asked whether they would intervene in similar situations. Lastly, the vignettes did not include any indication of threat towards the bystander for intervening. The failure to include this information has long been noted as a dearth in the bystander literature as that represents a significant barrier to helping. 147 Table 29. Logistic Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) Microaggression Hate Incident Hate Crime Coefficient (SD) Odds Ratio Coefficient (SD) Odds Ratio Coefficient (SD) Odds Ratio Empathy 0.0439** (.0151) 1.0449 .0942*** (.0234) 1.0988 .0768* (.0301) 1.0798 Bystander Behavior 0.1617** (.0494) 1.1755 .1783* (.0747) 1.1952 .1703 (.0981) 1.1856 Bystander Efficacy .0012 (.0118) 1.001 .0043 (.0166) 1.0043 .0027 (.0209) 1.0027 Dictator Game Average -.0673 (.0381) .9350 -.0203 (.0384) .9799 -.0393 (.0360) .9614 Other-Regarding .3841* (.1814) 1.4683 .1860 (.2836) 1.2045 .1141 (.3678) 1.1209 Feeling Toward Victim Group 1.095*** (.1885) 2.989 1.4620*** (.2479) 4.3145 1.590*** (.3183) 4.9035 Majority Group Member -.1588 (.1705) .8532 .2667 (.2621) 1.3056 .0136 (.3284) 1.0137 Constant -1.0593* (.4677) .3467 -1.319* (.6671) .2673 -.1639 (.8256) .8488 Model Fit χ2(7) 85.61 76.65 46.63 p-value <.0001 <.0001 <.0001 Pseudo R2 .0778 .1276 .1192 Note: *p<.05, **p<.01, ***p<.001 The coefficients indicating the difference in logarithmic odds ratios were converted into odds ratios in the likelihood of seeing a problem compared to corresponding reference groups. 148 Table 30. Logistic Regressions of Intervening for Each Phase on Bystander Characteristics (n=1,001) Microaggression Hate Incident Hate Crime Coefficient (SD) Odds Ratio Coefficient (SD) Odds Ratio Coefficient (SD) Odds Ratio Empathy -.0161 (0.134) .9840 .0459** (.0146) 1.0470 .0810*** (.0175) 1.0844 Bystander Behavior .2536*** (.0470) 1.2887 .2867*** (.0466) 1.3320 .1783** (.0548) 1.1951 Bystander Efficacy .0667*** (.0119) 1.0690 .0703*** (.0121) 1.0728 .0550*** (.0131) 1.0566 Dictator Game Average -.0191 (.0284) .9811 .0233 (.0507) 1.0236 -.0266 (.0310) .9737 Other-Regarding .1735 (.1517) 1.1894 .1380 (.1757) 1.1480 .4779* (.2104) 1.6126 Feeling Toward Victim Group 1.501*** (.2594) 4.4840 1.6491*** (.2053) 5.2022 1.341*** (.2005) 3.8233 Majority Group Member -.0903 (.1599) .9137 .1912 (.1710) 1.2107 .3173 (.2001) 1.3734 Constant -4.223*** (.5135) .0147 -4.6362*** (.5219) .0097 -3.579*** (.5523) .0279 Model Fit χ2(7) 144.49 223.82 151.36 p-value <.0001 <.0001 <.0001 Pseudo R2 .1092 .1738 .1533 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 The coefficients indicating the difference in logarithmic-odds were converted into odds ratios in the likelihood of intervening compared to corresponding reference groups. 149 Table 31. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics (n=1,001) Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy .0355* (.0150) 1.0361 -.0512** (.0184) .9501 .0867*** (.0178) 1.0906 Bystander Behavior -.1938*** (.0416) .8238 -.3859*** (.0600) .6799 .1921*** (.0560) 1.2117 Bystander Efficacy -.0661*** (.0120) .9361 -.0867*** (.0152) .9169 .0206 (.0136) 1.0209 Dictator Game Average .0225 (.0322) 1.0227 .0315 (.0377) 1.0320 -.0090 (.0354) .9910 Other-Regarding -.0196 (.1585) .9806 -.3849 (.2149) .6805 .3654 (.2075) 1.4410 Feeling Toward Victim Group -.8764** (.1180) .4163 -2.2215*** (.2784) .1084 1.3452*** (.2170) 3.8388 Majority Group Member .2218 (.2141) 1.2483 .0264 (.2113) 1.0267 .1954 (.1992) 1.2158 Constant 2.4804*** (6.7004) 11.9458 6.0862*** (.6598) 439.7628 -3.6059*** (.5666) .0272 Model Fit χ2(14) 270.36 p-value <.0001 Pseudo R2 .1253 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 The coefficients indicating the difference in logarithmic-odds were converted into relative risk ratios (RRR) in the likelihoods of class membership. From study 1 of this dissertation, class 1 contains the always interveners, class 2 contains those who intervene as the event escalates, and class 3 contains the individuals who never intervene. 150 Table 18. Stratification Analyses for Seeing a Problem During Each Phase Sensitivity Model Stratifications Rationale Tables Microaggression Hate Incident Hate Crime Victim Identity Asian Man Victim Ascertain whether results were specific to randomly assigned victim groups. A10-A12 Only empathy and feelings towards the victim group significantly predicted seeing a problem. Empathy and feelings toward the victim group significantly predicted seeing a problem. Feeling towards the victim group significantly predicted seeing a problem. Asian Woman Victim A10-A12 No variable reached statistical significance. Feelings towards the victim group perfectly predicted seeing a problem. This analysis could not run as four variables perfectly predicted results. Gay Man Victim A10-A12 Bystander efficacy gained significance. No other variable was significant. Feeling towards the victim group significantly predicted seeing a problem. No variable was significant. Lesbian Victim A10-A12 No variable reached statistical significance. No variable significantly predicted seeing a problem. Average spent on the dictator games gained statistical significance. Empathy and feeling toward the victim group lost significance. Trans Man Victim A10-A12 Only feelings towards the victim group significantly predicted seeing a problem. Feeling towards the victim group significantly predicted seeing a problem. Average spent on the dictator games and being a non-Hispanic white, heterosexual, cisgender man gained statistical significance. Feeling towards the victim group stayed significant. 151 Trans Woman Victim A10-A12 Empathy, bystander behavior, and feeling towards the victim group significantly predicted seeing a problem. Bystander behavior and feeling towards the victim group significantly predicted seeing a problem. Results did not substantively differ from the main model. 2020 Political Vote of State Residency Trump 2020 State Ascertain whether results differed on residency in Trump or Biden 2020 states. A22 Feeling toward the victim group significantly predicted seeing a problem. There was no substantive difference with the main model. There was no substantive difference with the main model. Biden 2020 State A22 There was no substantive difference with the main model. There was no substantive difference with the main model. Feeling toward the victim group significantly predicted seeing a problem. Respondent Race White Ascertain whether results differed based on whether respondents’ race. A26 Bystander behavior and feeling toward the victim group significantly predicted seeing a problem. Empathy and feeling toward the victim group significantly predicted seeing a problem. There was no substantive difference with the main model. non-White A26 Empathy and feeling toward the victim group significantly predicted seeing a problem. Empathy and feeling toward the victim group significantly predicted seeing a problem. Feeling toward the victim group significantly predicted seeing a problem. 152 Table 19. Stratification Analyses for Intervening During Each Phase Sensitivity Model Stratifications Rationale Tables Microaggression Hate Incident Hate Crime Victim Identity Asian Man Victim Ascertain whether results were specific to randomly assigned victim groups. A13-A15 Bystander efficacy significantly predicted intervening. Bystander efficacy significantly predicted intervening. Being a non-Hispanic white, heterosexual, cisgender man gained significance. Empathy, bystander efficacy, and favorability toward the victim group significantly predicted intervening. Asian Woman Victim A13-A15 Bystander efficacy significantly predicted intervening. No variable significantly predicted intervening. Average given on the dictator games significantly predicted intervening. Gay Man Victim A13-A15 There was no substantive difference with the main model. Bystander behavior, bystander efficacy, and favorability towards the victim group significantly predicted intervening. Bystander behavior and feeling toward the victim group significantly predicted intervening. Lesbian Victim A13-A15 Bystander behavior significantly predicted intervening. Bystander efficacy significantly predicted intervening. Feeling toward the victim group significantly predicted intervening. Trans Man Victim A13-A15 There was no substantive difference with the main model. Empathy, bystander behavior, and favorability toward the victim group significantly predicted intervening. Empathy and feeling toward the victim group significantly predicted intervening. Trans Woman Victim A13-A15 Bystander behavior and favorability towards the victim Bystander behavior, bystander efficacy, and favorability toward the victim group Being a non-Hispanic white, heterosexual, cisgender man gained statistical significance. 153 group significantly predicted intervening. significantly predicted intervening. There were no other substantive differences. 2020 Political Vote of State Residency Trump 2020 State Ascertain whether results differed on residency in Trump or Biden 2020 States. A23 There was no substantive difference with the main model. Bystander behavior, bystander efficacy, and feeling toward the victim group significantly predicted intervening. Empathy, bystander behavior, and feeling toward the victim group significantly predicted intervening. Biden 2020 State A23 There was no substantive difference with the main model. There was no substantive difference with the main model. Empathy, bystander behavior, bystander efficacy, and feeling toward the victim group significantly predicted intervening. Respondent Race White Ascertain whether results differed based on whether respondents’ race. A27 There was no substantive difference with the main model. Bystander behavior, bystander efficacy, and feeling toward the victim group significantly predicted intervening. Empathy, bystander behavior, bystander efficacy, and feeling toward the victim group significantly predicted intervening. non-White A27 Bystander behavior and feeling toward the victim group significantly predicted intervening. Other-regarding gained statistical significance. Empathy, bystander behavior, and feeling toward the victim group significantly predicted intervening. Empathy, bystander behavior, bystander efficacy, and feeling toward the victim group significantly predicted intervening. 154 Table 20. Stratification Analyses for Latent Class Membership Sensitivity Model Stratifications Rationale Tables Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Victim Identity Asian Man Victim Ascertain whether results were specific to randomly assigned victim groups. A16 Bystander efficacy reduced the likelihood of being in class 2. Empathy, bystander efficacy, and feeling toward the victim group reduced the likelihood of being in class 3. Empathy and feelings toward the victim group significantly increased the likelihood of being in class 2. Asian Woman Victim A17 Bystander efficacy reduced the likelihood of being in class 2. No significant differences were observed on any variable. No significant differences were observed on any variable. Gay Man Victim A18 Bystander efficacy reduced the likelihood of being in class 2. Bystander behavior, bystander efficacy, and feeling toward the victim group reduced the likelihood of belonging to class 3. Feelings toward the victim group and being a non-Hispanic white, heterosexual, cisgender man significantly increased the likelihood of being in class 2. Lesbian Victim A19 Bystander efficacy reduced the likelihood of being in class 2. Bystander behavior, bystander efficacy, and feeling toward the victim group reduced the likelihood of belonging to class 3. No significant differences were observed on any variable. Trans Man Victim A20 Bystander efficacy and feeling toward the victim group reduced the likelihood of being in class 2. Empathy, bystander behavior, and feeling toward the victim group reduced the likelihood of belonging to class 3. There was no substantive difference with the main model. Trans Woman Victim A21 Bystander behavior, other-regarding, and feeling toward the Bystander behavior and feeling toward the victim group reduced the Empathy, bystander efficacy, and being other-regarding 155 victim group reduced the likelihood of being in class 2. likelihood of belonging to class 3. significantly increased the likelihood of being in class 2. 2020 Political Vote of State Residency Trump 2020 State Ascertain whether results differed on residency in Trump or Biden 2020 States. A24 Bystander behavior, bystander efficacy, feeling toward the victim group significantly reduced the likelihood of being in class 2. There was no substantive difference with the main model. There was no substantive difference with the main model. Biden 2020 State A25 There was no substantive difference with the main model. Bystander behavior, bystander efficacy, and feeling toward the victim group significantly reduced the likelihood of being classified into class 3. There was no substantive difference with the main model. Respondent Race White Ascertain whether results differed based on whether respondents’ race. A28 Bystander behavior, bystander efficacy, and feeling toward the victim group significantly reduced the likelihood of being in class 2. Bystander behavior, bystander efficacy, and feeling toward the victim group significantly reduced the likelihood of being in class 3. There was no substantive difference with the main model. non-White A29 No variable was statistically significant. Empathy, bystander behavior, and feeling toward the victim group significantly reduced the likelihood of being in class 3. Empathy and bystander behavior significantly increased the likelihood of being in class 2. 156 Chapter 11: General Discussion and Conclusion Various forms of hate, most prominently hate crimes, against minority groups have been increasing within the United States in recent years (UCR, 2024). All forms of hate victimization have been associated with negative effects on mental health (Nadal et al., 2014). In particular, hate incident and hate crime victimization have each been associated with the development of post-traumatic stress disorder and depression (Botcherby et al., 2011; Leets, 2002). These acts, both criminal and non-criminal, leave marks on victims that generate waves of harm radiating to the communities to which they belong and beyond to other communities (Iganski, 2001; Wenger et al., 2022). The harms of hate cannot be overstated. They represent the importance of halting hate in all its forms to prevent these consequences for both victims and their communities. Given the public nature of these offenses, these consequences of hate can be mitigated through effective bystander intervention (Sue et al., 2019). Many onlookers, however, choose not to assist despite the potential efficacy of bystander intervention (Zempi et al., 2021). Using Latané and Darley’s (1970) situational model of bystander intervention as a framework, this dissertation made several contributions to the literature on bystander intervention to hate as it 1) explored whether bystanders come closer to intervening as a hate event escalates 2) expanded the bystander literature by assessing whether bystanders differentially progress through the situational model depending on the victim’s identity as either a racial, sexual, or gender minority individual, and 3) evaluated whether commonly known bystander traits such as empathy and bystander efficacy predicted bystander intervention in an escalating hate event. Specifically, study 1 examined whether bystanders progressed further through the situational model as a hate event with three escalating phases unfolded. These phases were: a microaggression, a hate incident in which a racist/homophobic/transphobic slur was used, and a 157 hate crime in which the victim was forcefully shoved to the ground and threatened with further physical violence. Situational model progress was defined as seeing a problem for each phase, intervening for each phase, and respondent classification into three classes derived from latent class analysis. Study 2 then assessed whether the victim’s identity impacted the aforementioned operationalizations of situational model progress. Potential victims included an Asian woman, an Asian man, a gay man, a lesbian, a trans man, and a trans woman. Lastly, study 3 examined whether bystander characteristics affected situational model progress with an emphasis on exploring majority group members (i.e., non-Hispanic white, heterosexual, cisgender males) compared to minority group members. All three studies in this dissertation relied on data from a Qualtrics online survey conducted in 2023 of 1,001 U.S. residents using quota sampling. In this chapter, the main findings of each study will be summarized with discussion as to their theoretical and policy implications, limitations, and directions for future study. Summary of Main Findings Study 1 Study 1 reported whether respondents would progress further through the situational model as the hate event became increasingly more dire. The aims of this study were to see if (1) there was empirical support for Latané and Darley’s (1970) assertion that bystanders progress further through their model the more severe the situation and (2) if bystanders fell into distinct latent classes based upon their progress across all three hate event phases: microaggression, hate incident, and hate crime. The findings from study 1 supported the hypothesis for Latané and Darley’s (1970) situational model progress as well as patterns of progress displayed via latent classes. Pertaining 158 to the former, no threshold effect was observed in which respondents would progress only to a point within the situational model and stop progressing across all hate phases. The six McNemar’s tests conducted to determine if individuals were more likely to see problems or more likely to intervene in the escalating phases revealed that, indeed, respondents were significantly more likely to see problems and intervene as the hate event escalated. That was true for all combinations of hate event phase and seeing a problem as well as intervening. The Wilcoxon signed-ranks tests supported the hypothesis of this study as bystanders progressed significantly further through the situational model for hate incidents than microaggressions, hate crimes than microaggressions, and hate crimes than hate incidents. Furthermore, the classes generated by the three-class solution revealed bystanders in the sample could be classified as individuals who always intervene (class 1), individuals who intervene as the hate event escalates (class 2), and individuals who never intervene as they either fail to see a problem or fail to feel morally obligated to assist (class 3). Thus, the individual tests exploring between-step differences in seeing a problem, intervening, and overall situational model progress in conjunction with the latent class analysis highlight that individuals progress further through the situational model as the event becomes more severe. These results demonstrate that bystanders can be compelled to act as a hate event becomes more severe. Given that hate language often precedes violent hate crimes (Masucci & Langton, 2017), bystanders armed with the knowledge that insults can lead to physical assault may be more willing to intervene earlier during hate events. Educating the public on how hate operates and escalates could inform their views of microaggressions and hate language as being more dire than just words. That information would enable them to see microaggressions and hate 159 language as meritorious of intervention when they before may have previously chosen not to act. By halting a hate event at its onset, bystanders may prevent escalation into violence. Study 2 Using the results of study 1, study 2 reported whether respondents would progress differently through the situational model based upon the victim’s identity which had been randomly assigned. It was hypothesized, based upon the public opinion literature as well as social categorization theory, that respondents would be most likely to help Asian victims followed by gay and lesbian victims then trans victims. The results of study 2 for this dissertation indicated partial support for the hypothesis that respondents would be most likely to assist Asian victims, followed by gay men and lesbians, and be least likely to assist transgender individuals. Bystanders to the Asian American woman microaggression were significantly more likely to see a problem compared to bystanders to the gay man, lesbian, and trans woman victims. Bystanders to the Asian man and trans man microaggressions also were significantly more likely to see a problem than those assigned to the lesbian microaggression. For the hate incident phase, subjects given the Asian woman scenario were only significantly more likely to see a problem than those witnessing the trans woman victim scenario. In terms of intervention, however, some significant differences persisted across all phases of the hate event. For the microaggression, hate incident, and hate crime phases, bystanders to an Asian woman victim were significantly more likely to intervene than bystanders to a gay man or trans woman. Likewise, bystanders to an Asian woman victim were significantly more likely to intervene than bystanders to a lesbian victim during the microaggression and hate incident phases. Both bystanders to an Asian man and bystanders to a trans man during the 160 microaggression phase were significantly more likely to intervene than bystanders to a lesbian victim. Examination of the latent classes into which bystanders could best be categorized highlighted the differences in how bystanders react based upon the victim’s identity. The majority of bystanders to the Asian woman victim could be classified as always interveners. There were significant differences in these classifications between bystanders to the Asian woman victim and bystanders to the gay man, lesbian, trans man, and trans woman victims. However, no other significant differences were observed. Thus, the hypothesis was partially supported as only bystanders to the Asian woman victim more frequently fell into the class of always interveners compared to sexual and gender minority victims with no observed favor towards sexual minority victims over gender minority victims. Overall, the results of study 2 reveal that bystanders indeed appraise hate events differently based upon the victim’s identity. The only manipulation was the victim’s identity in the vignettes; therefore, coupled with random assignment of respondents to one of the six victims, the victim’s identity alone appears to be responsible for these disparities in situational model progress. That individuals responded differently to hate events based on the identity of the victim signifies the importance of perceived victim worth and how perceptions of various minority groups affects bystander decision-making. These findings indicate the need for bystander meta-cognition in which bystanders question if they would react the same way to a hate event if the victim belonged to a different minority group. By mentally substituting the actual victim with an imagined victim, perhaps from a minority group the bystander holds in higher esteem, the bystander may then overcome whatever unconscious biases they have and progress further through the situational model to 161 ultimately intervene. Barnett and colleagues (1981) studied the possibility of this phenomenon. They found evidence for empathic transfer in which bystanders’ concern for one group can be transferred to concern for another group. Study 3 Study 3 examined whether bystander level characteristics impacted their situational model progress. These bystander traits most prominently include empathy (Nickerson et al., 2015; Secord Fredrick et al., 2020), bystander efficacy (Banyard, 2008; Banyard & Moynihan, 2011; Banyard et al., 2004), bystander decision-making (Jensen & Raver, 2020; Morgan, 1978; Shea et al., 2021), and demographic traits (Mainwaring et al., 2023; Ratcliff et al., 2023). In particular, it was hypothesized that majority group members (non-Hispanic white, heterosexual, cisgender males) would be the least likely to intervene compared to other groups. The main results of study 3 for this dissertation showed no support for the hypothesis that respondents who are cisgender, heterosexual, non-Hispanic white men would be less likely to see problems, intervene, and fall within latent classes that intervened. These individuals did not significantly differ from bystanders who belong to minority groups (e.g., women, transgender individuals, sexual minority individuals, racial minority individuals) for any of the analyses conducted. These included whether the respondent saw the phase as a problem, whether the respondent intervened for the phase, and the latent class membership of the respondent. When examining main effects in the sensitivity analysis, however, race, sexual orientation, and gender significantly predicted a few of the outcomes. Respondents who identified as male were significantly less likely than those who did not identify as male to see a problem with the hate incident phase. Heterosexuals were significantly less likely to intervene for the microaggression and hate crime phases compared to sexual minority individuals. 162 Heterosexual individuals were also significantly less likely to belong to the class of always interveners compared to their likelihoods of belonging to classes 2 and 3. Whites as opposed to non-whites were significantly more likely to intervene during the hate incident and hate crime phases. Thus, the results of the sensitivity analysis provided conflicting evidence regarding hypothesis 3 as gender and sexual orientation were in the predicted directions whereas race was not. These differences indicate that there is further nuance to examining intersectional identities and bystander intervention that needs to be elucidated. Regarding the other bystander variables, the results typically fell within the expected directions. The higher a respondent’s empathy, the more likely they were to see problems during the microaggression, hate incident, and hate crime phases. Moreover, these individuals were more likely to intervene for hate incidents and hate crimes. They were also more likely to belong to the classes that intervened as opposed to class 3 which did not intervene for any phase. Respondents who reported having previously intervened per the bystander behavior scale were significantly more likely to see problems during the microaggression and hate incident phases, intervene for the hate incident and hate crime phases, and be best classified as always interveners. Bystander efficacy did not predict whether a bystander found any of the phases to be a problem. It did, however, predict whether a bystander intervened during all of the hate event phases and whether they belonged to the classes of always interveners or those who intervened as the situation escalated compared to those who never intervened. Other-regarding as a decision-making process only differentiated whether respondents found microaggressions to be problems or intervened for hate crimes. Most importantly, whether bystanders felt favorably or unfavorably toward the victim’s group significantly predicted whether they would find the hate event phases to be problems, intervened for them, and belonged to the classes of always 163 interveners or those who intervened as the situation escalated compared to those who never intervened. This was the only variable that had a large effect size whereas all the other variables only had weak to small effect sizes. These findings frequently replicated in the stratification analyses conducted by splitting the sample based upon assigned victim, residing in a Trump 2020 state or Biden 2020 state, and being white or non-white. The strongest and most consistent indicator of situational model progress among these stratifications was the bystander’s perceived favorability of the victim’s group. The other most frequent indicators were bystander efficacy, bystander behavior, empathy, and other-regarding. In sum, the findings of study 3 show the complex interplay of bystander traits that ultimately impact situational model progress. Notably, three of these variables are related constructs: favorability toward the victim group, empathy, and other-regarding. Empathy in study 3 primarily consisted of affective empathy that occurs spontaneously. Other-regarding, however, was defined as a deliberate decision-making process operationalized as how fair or generous the respondent was in sets of dictator and ultimatum games. Studies have shown that empathy and perspective-taking, which is a more deliberate empathic process, can foster positive feelings towards outgroups and reduce racism as well as anti-immigrant attitudes (Miklikowska, 2017; Shih et al., 2009; Stephan & Finaly, 1999). Therefore, favorability towards the victim group may interact with empathy to affect situational model progress, though future studies would need to test this hypothesis. Thus, the findings of study 3 demonstrate promise for how empathy can be used to inspire bystanders to act during hate events. Collective Findings The findings for all three studies demonstrate that incident severity, victim identity, and 164 bystander characteristics all influence progress through the situational model for bystander intervention within the context of hate-based victimization. As an event becomes more dire, bystanders become significantly closer to, and often end up, intervening. Moreover, bystander progression through the situational model appears to depend on the identity of the victim; hence, why study 3 was stratified by the victim identity in addition to the inclusion of a variable for favorability toward the victim. Racial minority victims, in particular the Asian woman victim, in these scenarios received the most support from bystanders compared to sexual and gender minority victims. The effect of victim identity on situational model progress may have revealed how victim blameworthiness and implicit bias impact bystander intention to intervene. Indeed, the bystander’s favorability towards the victim’s group consistently significantly predicted seeing a problem, intervening, and being best classified into the latent class of always interveners. Lastly, traits intrinsic to the bystander such as empathy, bystander efficacy, and prior bystander intervention behaviors consistently predicted situational model progress. Incident severity, victim identity, and bystander characteristics thus all coalesce into informing whether a bystander proceeds through the decision-making process outlined by the situational model. Within the context of the latent class analysis model, these findings collectively reveal what factors differentiate between the individuals who always intervene (class 1), those who intervene as the situation escalates (class 2), and those who never intervene (class 3). Study 2 showed that bystanders were most likely to belong to class 1 when the victim was an Asian woman compared to the gay man, lesbian, trans man, and trans woman victims. Moreover, study 3 highlighted that empathy, bystander behavior, and feelings towards the victim group consistently distinguished between all classes. These findings together demonstrate the needs for (1) bystanders to contemplate their feelings towards various minority groups and (2) the 165 cultivation and subsequent leveraging of empathy towards minority groups. Culturally sensitive bystander intervention trainings have been found to increase intervention in cases of discrimination yet few trainings have incorporated cultural sensitivity (Stephens et al., 2023). Given its operationalization, the significance of previous bystander behavior to derogatory remarks signifies the importance of past intervention in establishing the habit of intervening against hate. Modifying bystanders’ decision-making processes to focus on the victim as a fellow human being rather than a member of a particular outgroup may enhance bystander progress in the situational model to compel action. Indeed, prior research has shown that intergroup bias is reduced when individuals are told to view members of outside groups as part of a superordinate group such as human beings (Gaertner et al., 1993; Paluck & Green, 2009). Throughout all three studies, significant differences often emerged most for the microaggression phase of the hate event. In study 1, nearly one-fourth of respondents failed to see a problem with the microaggression compared to nearly 9% for hate incident and nearly 5% for hate crime phases. In terms of intervention, approximately 37% intervened for the microaggression compared to 66% for the hate incident and 81% for the hate crime phases. In study 2, five pairwise significant differences based on the victim’s identity were observed for seeing a problem during the microaggression phase compared to only one pairwise significant difference in the hate incident phase and no pairwise significant differences in the hate crime phase. Intervening during the microaggression phase likewise showed five pairwise significant differences whereas there were only four for the hate incident phase and two for the hate crime phase. In study 3, four bystander variables significantly predicted seeing the microaggression as a hate crime compared to three variables for the hate incident and two for the hate crime. 166 These findings are paramount for preventing hate as microaggressions can be the precursors to violence in specific instances while more generally reinforcing prejudicial beliefs in society (Anti-Defamation League, 2003). The wealth of significant results in the microaggression phase reveals numerous ways in which bystanders can be compelled to intervene. In particular, past bystander behavior and feelings towards the victim group significantly predicted whether a bystander saw a problem and whether a bystander intervened when witnessing a microaggression. As will be discussed further in this chapter, bystander trainings must therefore incorporate elements of empathy training to better prompt bystanders to move towards intervening when witnessing hate. The programs that include empathy training have been shown to increase bystander willingness to intervene (Bell et al., 2019). Such trainings must also discuss microaggressions, how to detect them, and why they are harmful as most bystander intervention trainings centered around hate typically focus on hate incidents or hate crimes (Sue et al., 2019). Theoretical Implications Situational Model The findings from all three studies have implications for Latané and Darley’s (1970) situational model of bystander intervention. Of all the studies, study 1 most prominently tested the assertions made by them in their foundational work. The findings from study 1 provided mixed support for Latané and Darley’s (1970) situational model. In support of the model, bystanders tended to progress further through the process as the event escalated. However, the results contrasted with Latané and Darley’s (1970) expectation that subjects can commit themselves to not intervening through failing to intervene at the onset of an emergency. Latané and Darley originally reasoned that bystanders may continue to talk themselves into believing no 167 emergency is occurring to reinforce their initial decision not to perceive it as such. Thus, they posited the longer an event transpires, the less likely a bystander is to modify their decision-making. Respondents in the current study were subjected to an escalating situation which may have acted as an additional stimulus for them to react. Indeed, that a distinct latent class of individuals who progressed further through the situational model as the hate event escalated was found casts doubt on their belief that there is inertia in bystander intervention. Given that there were latent classes of bystanders, Latané and Darley’s (1970) situational model could be expanded to consider that different types of bystanders might exist. Notably, one latent class of people who were most likely to intervene for all phases was found along with a class of individuals who progressed further based on the phase’s severity. As further developed in study 3, the possibility of there being qualitatively different classes of bystanders represents that there are greater possibilities of individual level differences in bystanders than Latané and Darley (1970) originally believed. Their foundational work tended to focus on the severity of the incident as well as the presence of other bystanders as opposed to bystander-level traits. Findings from study 2 and study 3 reveal the need for the theoretical underpinnings of the situational model to be expanded to include more recent concepts such as implicit bias and integrated with theories such as social categorization theory. While the situational model contains descriptions of how victim blame may be attributed differently among potential victims, the mechanisms of that phenomenon were not detailed. It is critical that these mechanisms be further elucidated to ascertain how this effect can be nullified to increase situational model progress for all victims of hate rather than some. The integration of social categorization theory provides one avenue for completing this task. Social categorization theory describes the processes in which people categorize themselves and others based on in-group or outgroup 168 membership. Moreover, it emphasizes that individuals have a more global identity of being human over being subgroup members. Revealing how bystanders select which identity matters can demonstrate how to better leverage bystander intentions to halt hate. That the likelihoods of whether bystanders saw problems or intervened in study 2 tended to vary based upon the victim’s identity represents that the perceived severity of the incident may depend on who the victim is. In particular, bystanders to the Asian woman victim were significantly more likely to be best classified as the always interveners compared to bystanders of the gay man, lesbian, trans woman, and trans man victims. The importance of the victim’s type of minority status was further compounded by the results of study 3 which showed bystanders’ favorability towards the victim group was a strong predictor for seeing a problem, intervening, and latent class membership into latent classes that always intervened or progressed further through the model as the situation became more dire. Findings from study 3 further show the need for the situational model to be expanded to account for bystander level traits given they distinguished between interveners and non-interveners as well as each of the latent classes. While Latané and Darley (1970) briefly discuss empathy and altruism, the onus of their theoretical exploration into bystander intervention is on the number of bystanders nearby. Within the vignettes given for this dissertation, no mention was made as to the number of other bystanders in the area yet several traits such as empathy, bystander efficacy, and bystander behavior consistently predicted intervention. Further development of the situational model should explore why these bystander-level variables are of theoretical significance in progressing further through the model based on how empathic one is, how likely one is to believe intervention can work to stop violence, and one’s own past experiences voicing concerns over derogatory comments. Heightened empathy significantly 169 increased the likelihood of seeing a problem with each hate event phase. Empathy allows a bystander to perceive the discomfort of the victim in what is known as empathic distress (Hoffman, 1990).. Further, the degree of this discomfort may inform whether the bystander considers the situation to be a problem as well as their moral obligation to assist (Wang & Kim, 2021). Similarly, bystander efficacy supports the decision to intervene as the bystander knows how to respond and believes that intervention will be effective. Previous experience intervening, even to derogatory remarks made concerning minority groups, further amplifies noticing a problem, feeling obligated to assist, and knowing how to assist. Expansion of the model would provide greater insight into the best ways these traits can be cultivated and leveraged to increase the likelihood of intervening in hate-based scenarios. Social Categorization Theory Findings from studies 2 and 3 also have implications for social categorization theory. The aspects of studies 2 and 3 that focused on the victim’s identity found support for bystanders progressing differentially through the situational model based upon the victim’s group. For study 2, bystanders more frequently were classified as the always interveners for the Asian woman victim compared to sexual and gender minority victims. For study 3, feeling favorably as opposed to unfavorably toward the assigned victim group consistently predicted seeing a problem, intervening, and belonging to classes that intervened across all possible dependent variables. Respondents who thought favorably of the victim’s group likely had positive views towards them in other ways and therefore saw them as being more worthy victims. Thus, when centering the results on traits of the victim and the feelings elicited by them, these studies supported social categorization theory as victim group membership significantly influenced how bystanders responded to the hate event. 170 Furthermore, that the differences in study 2 were most pronounced between the Asian woman victim and the sexual and gender minority victims demonstrates how stereotypes of these groups may impact situational model progress. For example, hate crimes against Asian Americans were down 33.7% in 2022 compared to 2021 as shown in Table A1, suggesting that public sentiment towards Asian Americans may be increasing in parallel with the waning severity of COVID-19. Sexual and gender minority individuals, however, have not elicited the same positive public perception that Asian Americans sometimes have had. Homophobic and transphobic rhetoric has been growing with Gallup surveys showing consequent increases in homophobic and transphobic attitudes by the U.S. public (Jones, 2023a). Thus, as stereotypes towards Asian Americans tend to be positive, the results support the theoretical expectations from social categorization theory as Asian Americans, and the Asian woman in the vignettes in particular, received significantly more bystander progress through the situational model than sexual and gender minority individuals. However, contrary to the hypothesis for this study, trans men were more likely to have subjects view microaggressions against them as problems and ultimately intervene compared to lesbians. In contrast with social categorization theory, however, majority group membership in study 3 frequently did not predict situational model progress. When examining main effects from the sensitivity analysis, being heterosexual as opposed to another sexual orientation was associated with the most significant differences compared to race and gender identity. Heterosexual individuals were significantly less likely to intervene during the microaggression and hate crime phases compared to sexual minority individuals. They were also significantly less likely to be classified as the always interveners. Conversely, white respondents were significantly more likely to intervene during the hate incident and hate crime phases. 171 These findings suggest there is greater nuance for social categorization theory in that it may not always be as simple as in-group versus out-group membership for bystander intervention. The type of majority group (e.g., whites for race, heterosexuals for sexual orientation, and cisgender individuals for gender identity) may have different effects from one another. To that end, study 3 provided support for the existence of white allies who act against hate despite belonging to the majority group. White allies have been noted for their potential as upstanders because they can use the privilege of their majority group status to intercede on behalf of minority individuals (Sue et al., 2019). Social categorization theory thus ought to be expanded to demonstrate how majority group status may be used as a form of noblesse oblige to aid imperiled minority group members. Policy Implications The findings of this dissertation demonstrate elements that could be incorporated into bystander intervention training to enhance the likelihood of bystanders stopping hate. Current bystander intervention trainings such as Right to Be (n.d.) focus on providing attendants with toolkits of possible actions to take when witnessing hate. Per Bell et al. (2019), these programs significantly improve bystander efficacy and bystander behavior. As study 3 revealed, bystander efficacy as well as having prior histories of intervening consistently significantly predicted whether respondents would see problems or intervene across numerous stratifications and sensitivity analyses. However, some existing bystander intervention trainings fall short of leveraging empathy and other characteristics to improve willingness to intervene. One notable exception is the HateLess program implemented for adolescents to intervene against hate speech. In addition to reviewing methods of intervening, HateLess uses intergroup contact interventions through providing stories of victims to hate speech, as well as skills training to improve 172 participants’ empathy towards victims of hate (Beelmann & Lutterbach, 2020). This training significantly improved participants’ empathy and bystander efficacy that ultimately led to significantly higher likelihoods of using counter-speech (i.e., speech that discourages prejudicial statements or provides support for victims) when witnessing hate speech (Wachs et al., 2023). Thus, empirical evidence from the literature provides further support for the incorporation of empathy training into bystander intervention training. Bystander intervention trainings could additionally focus more on microaggressions as a form of hate. That so many individuals in study 1 did not see the microaggression as a problem or did not feel morally obligated to assist highlights the need to educate people on the harms of microaggression as well as how to identify them. Furthermore, the results of study 2 compound that implication as demonstrated with the bystanders often failing to see homophobic and transphobic microaggressions as problems compared to racist microaggressions. Bystander trainings should therefore explain to participants how microaggressions appear differently when enacted against different groups. In addition to bystander intervention trainings, the findings from studies 2 and 3 provide preliminary evidence for the need for legislative reform pertaining to sexual and gender minority individuals. Bystanders frequently were least likely to assist them compared to the racial minority victims. Moreover, favorability toward the victim’s minority group was the strongest predictor of intervention. Homophobic and transphobic rhetoric has been growing with Gallup surveys showing consequent increases in homophobic and transphobic attitudes by the U.S. public (Jones, 2023a). The “Don’t Say Gay” and “Don’t Say Trans” laws recently instituted in some states have been associated with large decreases in public sentiment towards gay and trans individuals, especially among Republicans (Jones, 2023b). These laws seemingly reinforce 173 negative stereotypes of sexual minority individuals and gender minority individuals. Their elimination should likewise signal acceptance towards them and hopefully reverse this trend in public sentiment, though future studies must be conducted using heterosexual and cisgender victims to verify if the disparity was due to sexual and gender minority status. In a similar vein, the findings from study 2 further represent the importance of expanding hate crime legislation and protected classes. Currently, only 35 states have laws classifying sexual minority individuals as a protected class for hate crime, 16 of which also protect gender minority individuals (Department of Justice, 2023). Conversely, 48 states have laws protecting individuals on the basis of race or national origin. The failure to include sexual and gender minority groups as protected classes in many states is likely reflective of public opinion towards them. The inclusion of these groups as protected classes might signal to the public that they are meritorious of protection. Indeed, legal scholars have noted that hate crime law represents a form of expressive law symbolizing the immorality of prejudice and the virtue of protecting all members of society (Mason, 2014). Thus, legislators ought to move toward including protections based on sexual orientation and gender identity in hate crime statutes to show that members of these groups deserve protection. The disparities unveiled by study 2 also highlight the additional needs that sexual and gender minority victims of hate crime may have for victims’ services. As previously noted, all forms of hate are associated with deleterious effects on victims’ physical and mental health (Nadal et al., 2014). Effective bystander intervention that includes aftercare for the victim can mitigate these consequences (Sue et al., 2018). Given that bystanders were significantly less likely to intervene for them than for the Asian woman victim, for example, shows that they might then require greater assistance after the victimization as they did not receive the benefit of 174 intervention. Victims’ services for hate crime should therefore have culturally competent care in anticipation of the lack of bystander intervention to attend to these heightened consequences (Palmer & Kutateladze, 2022). On a broader note, these findings from studies 2 and 3 signify the importance of diverse representation and enhancing intergroup contact. One method of doing so has been to bring community members of diverse backgrounds together for hate crime reduction programs (Freilich & Chermak, 2013). These programs foster intergroup bond formation to reduce the likelihood of hate crime perpetration. Such programs are premised on intergroup contact theory, which posits that knowing and having positive interactions with members of outgroups helps to improve attitudes towards those outgroups (Blau, 1977; Pettigrew, 1998). These programs would likely extend from not only preventing hate crime perpetration but also into increasing the likelihood of intervention as intergroup contact increases favorability towards other groups. Additionally, the fostering of the human identity as being the primary identity of an individual could help to reduce the disparity in intervention based upon victim identity. Limitations and Directions for Future Research The studies in this dissertation shared several limitations. First, the limitations from study 1 had downstream effects on studies 2 and 3 given that study 1 informed the dependent variables used in those later studies. These limitations were that respondents were recruited via quota sampling from Qualtrics rather than through a probability-based sampling method. This means that the respondents do not truly reflect the U.S. population as not every member had an equal chance of being selected, limiting the validity of the results as they may not extend to the true U.S. population. Future studies should attempt to draw probability-based samples to have a 175 representative sample. Additionally, future studies should examine whether there are any cross-cultural differences by using samples from outside the U.S. Second, there may be question order effects in which respondents went further through the situational model for the hate incident and hate crime phases given their responses for the prior phase(s). To fully explore how respondents would react to a hate event unfolding before them, it was necessary to include all phases and the situational model progress questions at the end of each phase. Each phase of the hate event ideally would have been randomly assigned to eliminate the possibility of order effects, but that would have made the resultant cell sizes too small for meaningful analysis when combined with the random assignment of victim characteristics. Future studies could mitigate this concern by obtaining samples large enough for this random assignment to be tested. Third, the latent class solution ultimately selected did not have the best goodness of fit in comparison to other class solutions, namely the five-class solution. While the three-class solution was selected based on its parsimony, theoretical relevance, and larger class sizes, it is possible that further nuance in studies 2 and 3 was obfuscated by using the three-class solution. Sensitivity analyses were conducted using the five-class solution, but its small class sizes still limit its usefulness. That could be ameliorated in the future through an even larger sample that would allow for greater comparisons between latent classes. Fourth, no items were included in the survey that asked respondents whether they felt safe or perceived any costs to their personal safety by intervening. This limitation is present in much of the bystander literature (Banyard, 2015). Its absence is notable as bystander safety may be a significant barrier informing whether one intervenes or not. McMahon and Dick (2011) discovered bystanders do consider the possible consequences intervention will have for 176 themselves along with the anxiety that comes from worrying about one’s own safety. That anxiety then inhibits bystanders from intervening. Moreover, bystanders have just cause for worrying about their safety as they can themselves become targets of violence when intervening (Hamby et al., 2015). Future studies should ask respondents how safe or unsafe they felt in the vignette to determine whether that impacts bystander decision-making. Fifth, the possible victims were limited to an Asian man, an Asian woman, a gay man, a lesbian, a trans man, or a trans woman. While these possible victims allowed for diversity in terms of race, sexual orientation, and gender identity, there are far more possible victims of hate crime based on other traits such as religious affiliation or disability status. Future studies should explore whether these differences continue to be observed if, for example, the victim were Black or Middle Eastern as opposed to Asian. Other sexual orientations and gender identities could similarly be examined to see what other differences might exist. Furthermore, studies should include religious minority individuals such as Jews given the increase in anti-Semitic hate crime (see Table A1 of the appendix). Sixth, respondents were provided with no information as to whether there were any other bystanders to the event. Given that the diffusion of responsibility greatly affects whether one intervenes (see Latané & Darley, 1970), future projects should specify whether any bystanders were present, how they reacted to the situation, and what other signals they were providing the subject. The public nature of hate offenses increases the likelihood of other bystanders being present, which may have impacted the results of these studies as that key detail was not specified in the vignettes. Seventh, subjects’ responses to a hypothetical vignette, particularly one administered online, may not be reflective of their behavior in reality. The literature contains mixed findings 177 as to whether respondents follow through with their intentions in a vignette scenario with their actions in real life. Work by Exum and Layana (2018) as well as Exum and colleagues (2012, 2014) demonstrated that subjects who claimed they would engage in classroom cheating or illegally downloading music rarely did so when given the opportunity in real-life. Intentions to act prosocially by abstaining from these activities, however, matched the respondents’ later actions. While bystander intervention is a prosocial act, it also requires the respondent to do something as opposed to choosing not to perform an unethical act. Thus, Exum’s work highlights a potential flaw in the external validity of this dissertation’s findings as it is unknown if respondents’ cognitive processes in real-life would match their survey responses. Future studies should attempt to replicate this dissertation’s findings through a variety of other methods such as real-life situations with confederates. Lastly, no measure for implicit bias was included in the survey instrument. Its absence precludes the possibility of testing if implicit bias was responsible for the differences observed in study 2. Although it is assumed that was the cause for sexual and gender minority victims receiving less help than racial minority victims, there is no direct test to empirically confirm that. Moreover, while the instrument did include feeling thermometers to ascertain respondents’ views towards the victim groups, feeling thermometers represent explicit bias rather than implicit bias. Ratings on feeling thermometers require respondents to be aware of and report their own potential biases whereas implicit bias occurs subconsciously without the individual’s own awareness (Greenwald & Banaji, 1995). Thus, the finding that favorability towards the victim group was the strongest predictor of progress through the situational model could have further nuance to be explored. Future studies should include measures of implicit bias to examine its relationship with explicit bias and effects on situational model progress. 178 Conclusion This dissertation has demonstrated that bystander decision-making is a nuanced process dependent on a confluence of contextual, victim, and bystander traits. Even if results did not always conform to theoretical expectations, significant findings were observed in every study. Bystander decision-making to hate-based scenarios requires further elucidation to better determine what factors should be cultivated to enhance the likelihood of intervention. While this dissertation found that incident severity, victim group membership, and several bystander traits impacted progress in the situational model, there are many other factors yet to be explored such as bystander perceptions of safety. Until such studies are completed and synthesized with the extant literature, there will remain an incomplete picture of what truly motivates people to assist in times of crisis. 179 References Aguinis, H., & Bradley, K. J. (2014). Best practice recommendations for designing and implementing experimental vignette methodology studies. Organization Research Methods, 17(4), 351-371. American Psychological Association (2009). Report of the Task Force on Gender Identity and Gender Variance. http://www.apa.org/pi/lgbt/resources/policy/gender-identity-report.pdf American Psychological Association (2012). Guidelines for psychological practice with lesbian, gay, and bisexual clients. American Psychologist, 67, 10-42. Anker, A. E., & Feeley, T. H. (2011). Are nonparticipants in prosocial behavior merely innocent bystanders? Health Communication, 26(1), 13-24. Anti-Defamation League and Survivors of the Shoah Visual History Foundation. (2003). Pyramid of Hate – Lesson for High School. https://sfi.usc.edu/sites/default/files/lessons/units/POH_Final_0.pdf Anti-Defamation League. (2021). No Place for Hate: The Program. https://www.noplaceforhate.org/the-program April, K., Cole, L. M., & Goldstein, N. E. S. (2023). Let’s “Talk” about the police: The role of race and police legitimacy attitudes in the legal socialization of youth. Current Psychology, 42, 15422-15437. Armstrong, R. A. (2014). When to use the Bonferroni correction. Ophthalmic & Physiological Optics, 34, 502-508. 180 Ash, E., Durante, R., Grebenschikova, M., & Schwarz, C. (2022). Visual representation and stereotypes in news media. CEPR Discussion Paper No. DP16624. https://ssrn.com/abstract=3960205 Ashburn-Nardo, L., Morris, K. A., & Goodwin, S. A. (2008). The confronting prejudiced responses (CPR) model: Applying CPR in organizations. Academy of Management Learning & Education, 7(3), 332-342. Bailey, A., & Jones, L. (2023). ‘Fairness versus inclusion’: Representations of transgender athletes in British newspaper reports. In K. File, & S. Schnurr (Eds.), The Language of Inclusion and Exclusion in Sport. De Gruyter. Banks, B. M., Adams, D. F., AuBuchon, S., Hynes, K., & Torres Gonzales, N. (2023). Factors influencing satisfaction with a microaggression bystander intervention. Journal of Human Behavior in the Social Environment. Advance online publication. https://doi.org/10.1080/10911359.2022.2027316 Banyard, V. (2008). Measurement and correlates of prosocial bystander behavior: The case of interpersonal violence. Violence and Victims, 23(1), 83-97. Banyard, V. L. (2015). Toward the next generation of bystander prevention of sexual and relationship violence: Action coils to engage communities. Springer Science + Business Media. https://doi.org/10.1007/978-3-319-23171-6 Banyard, V. L., Plante, E., & Moynihan, M. M. (2004). Bystander education: Bringing a broader community perspective to sexual violence prevention. Journal of Community Psychology, 32(1), 61-79. 181 Banyard, V., & Moynihan, M. (2011). Variation in bystander behavior related to sexual and intimate partner violence prevention: Correlates in a sample of college students. Psychology of Violence, 1(4), 287-301. Banyard, V., Moschella, E., Jouriles, E., & Grych, J. (2021). Exploring actions coils for bystander intervention: Modeling bystander consequences. Journal of American College Health, 69, 283-289. Barmann, J. (2021, June 2). Anti-LGBT 'Karen' video emerges from hotel pool in Sacramento; Oakland woman was allegedly upset over women kissing. SFist. https://sfist.com/2021/06/02/anti-lgbt-karen-video-emerges-from-hotel-pool-in-sacramento-oakland-woman-was-upset-over-women-kissing/ Barnett, M. A., Howard, J. A., King, L. M., & Dino, G. A. (1981). Helping behavior and the transfer of empathy. The Journal of Social Psychology, 115, 125-132. Batson, C. (1991). The altruism question: Toward a social-psychological answer. Erlbaum. Batson, C. (1997). Self-other merging and the empathy-altruism hypothesis: Reply to Neuberg et al. (1997). Journal of Personality and Social Psychology, 73, 517-522. Bayley, D. H. (1996). Police for the future. Oxford University Press. Beelmann, A., & Lutterbach, S. (2020). Preventing prejudice and the promotion of intergroup relations. In L. Benuto, M. Duckworth, A. Masuda, & W. O’Donohue (Eds.). Prejudice, stigma, privilege, and oppression: A behavioral health handbook (pp. 309-326). Springer Nature. Beemyn, G. (2013). A presence in the past: A transgender historiography. Journal of Women’s History, 25(4), 113-121. 182 Bell, S. C., Coker, A. L., & Clear, E. R. (2019). Bystander program effectiveness: A review of the evidence in educational settings (2007-2018). In W. O’Donohue & P. A. Schewe (Eds.), Handbook of sexual assault and sexual assault prevention (pp. 433-450). Springer Nature. Bennett, S., & Banyard, V. L. (2016). Do friends really help? The effect of relational factors and perceived severity on bystander perception of sexual violence. Psychology of Violence, 6(1), 64-72. Berntsen, D., Willert, M., & Rubin, D. C. (2003). Splintered memories or vivid landmarks? Qualities and organization of traumatic memories with and without PTSD. Applied Cognitive Psychology, 17, 675-693. Bettcher, T. M. (2007). Evil deceivers and make-believers: On transphobic violence and the politics of illusion. Hypatia, 22(3), 43-65. Bills, M. A., & Vaughn, M. S. (2023). A contemporary review of hate crime legislation in the United States. Criminal Justice Policy Review, 34(2), 115-139. Blackwell, H. (2022). The impact of empathy-building activities: Implementing the Anti-Defamation League’s No Place for Hate program. [Doctoral dissertation, George Fox University]. https://digitalcommons.georgefox.edu/edd/181/ Blau, P. M. (1977). Inequality and heterogeneity. Free Press. Blume, A. W., Lovato, L. V., Thyken, B. N., & Denny, N. (2012). The relationship of microaggressions with alcohol use and anxiety among ethnic minority college students in a historically White institution. Cultural Diversity and Ethnic Minority Psychology, 18(1), 45-54. 183 Blyth, C., & McRae, P. (2018). “Death by a thousand paper cuts”: Transphobia, symbolic violence, and conservative Christian discourse. In C. Blyth, E. Colgan, & K. B. Edwards (Eds.), Rape Culture, Gender Violence, and Religion (pp. 111-133). Palgrave Macmillan. Bonilla-Silva, E. (2004). From bi-racial to tri-racial: Towards a new system of racial stratification in the USA. Ethnic and Racial Studies, 27(6), 931-950. Botcherby, S., Glen, F., Iganski, P., Jochelson, K., & Lagou, S. (2011). Equality groups’ perceptions and experience of crime. Analysis of the British Crime Survey 2007-08, 2008-09 and 2009-10 (Briefing Paper 4). Equality and Human Rights Commission. Bradbury-Jones, C., Taylor, J., & Herber, O. R. (2014). Vignette development and administration: A framework for protecting research participants. International Journal of Social Research Methodology, 17(4), 427-440. Bradley, D. R., & Cucomb, S. (1977). Monte Carlo simulations and the chi-square test of independence. Behavior Research Methods & Instrumentation, 9(2), 193-201. Brennan Center for Justice. (2021). Exploring alternative approaches to hate crimes. Brennan Center for Justice. http://law.stanford.edu/wp-content/uploads/2021/06/Alternative-to-Hate-Crimes-Report_v09-final.pdf Brennan, P. K., Spohn, C. (2009). The joint effects of offender race/ethnicity and sex on sentence length decisions in federal courts. Race and Social Problems, 1, 200-217. Brewer, M. B. (1999). The psychology of prejudice: Ingroup love and outgroup hate?. Journal of Social Issues, 55, 429-444. Brewer, P. R., &Wilcox, C. (2005). The polls—Trends: Same-sex marriage and civil unions. Public Opinion Quarterly, 69, 599-616. 184 Britt, C. L., & Weisburd, D. (2010). Logistic regression models for categorical outcome variables. In A. Piquero & D. Weisburd (Eds.), Handbook of quantitative criminology (pp. 649-682). Springer. Brooks, C., & Harmon, A. (2022). Affect toward minority and majority groups in the era of Donald Trump. Socius, 8, 1-15. Brown, A. L., Banyard, V. L., & Moynihan, M. M. (2014). College students as helpful bystanders against sexual violence: Gender, race, and year in college moderate the impact of perceived peer norms. Psychology of Women Quarterly, 38(3), 350-362. Brown, T. A. (2015). Confirmatory factor analysis for applied research (2nd ed.). The Guilford Press. Brysbaert, M. (2019). How many participants do we have to include in properly powered experiments? A tutorial of power analysis with reference tables. Journal of Cognition, 2(1), 1-38. Budge, J. (2023). Are more educated states more gay-friendly? How the increase in college attainment promotes acceptance of gay men and lesbians. The Sociological Quarterly. Advance online publication https://doi.org/10.1080/00380253.2023.2189465 Burns, K. (2016). Judges, ‘common sense’ and judicial cognition. Griffith Law Review, 25(3), 319-351. Burns, K. (2020, June 3). A brutal video shows a black trans woman getting beat up by dozens of men. Vox. https://www.vox.com/identities/2020/6/3/21279407/video-black-trans-woman-iyanna-dior-beaten Burns, V. L., Eaton, A. A., Long, H., & Zapp, D. (2018). Exploring the role of race and gender on perceived bystander ability and intent: Findings before and after exposure to an online 185 training program to prevent sexual assault on campus. Violence Against Women, 25(8), 999-1017. California Penal Code § 190.2 (1978). Carbado, D. W., & Roithmayr, D. (2014). Critical race theory meets social science. Annual Review of Law and Social Science, 10, 149-167. Carré, A., Stefaniak, N., D’Ambrosio, F., Bensalah, L., & Besche-Richard, C. (2013). The basic empathy scale in adults (BES-A): A factor structure of a revised form. Psychological Assessment, 25(3), 679-691. Carson, J. V., & Politte, H. (2021). Implicit bias within public reporting: A virtual reality experiment examining “suspicious” activity. Crime & Delinquency, 67(12), 2135-2162. Casey, E. A., Lindhorst, T., & Storer, H. L. (2017). The situational-cognitive model of adolescent bystander behavior: Modeling bystander decision-making in the context of bullying and teen dating violence. Psychology of Violence, 7(1), 33-44. Center for Countering Digital Hate and the Human Rights Campaign. (2022). Digital hate: Social media’s role in amplifying dangerous lies about LGBTQ+ people. Center for Countering Digital Hate, Inc. https://hrc-prod-requests.s3-us-west-2.amazonaws.com/CCDH-HRC-Digital-Hate-Report-2022-single-pages.pdf Center for the Study of Hate and Extremism. (2021). Report to the nation: Anti-Asian prejudice & hate crime. California State University San Bernardino. https://www.csusb.edu/sites/default/files/Report%20to%20the%20Nation%20-%20Anti-Asian%20Hate%202020%20Final%20Draft%20-%20As%20of%20Apr%2030%202021%206%20PM%20corrected.pdf 186 Chabot, H. F., Tracy, T. L., Manning, C. A., & Poisson, C. A. (2009). Sex, attribution, and severity influence intervention decisions of informal helpers in domestic violence. Journal of Interpersonal Violence, 24, 1696-1713. Chiricos, T., Welch, K., & Gertz, M. (2004). Racial typification of crime and support for punitive measures. Criminology, 42(2), 358-390. Civil Rights Act of 1968, 18 U.S.C. § 245 (1968). Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Routledge Academic. Cohen, J. E. (2018). From antisemitism to philosemitism? Trends in American attitudes toward Jews from 1964 to 2016. Religions, 9, 107-127. Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. Wiley. Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology, 78(1), 98-104. Costa, P. L., McDuffie, J. W., Brown, S. E. V., He, Y., Ikner, B. N., Sabat, I. E., & Miner, K. N. (2023). Microaggressions: Mega problems or micro issues? A meta-analysis. Journal of Community Psychology, 51, 137-153. Councilor, K. C. (2021). Public and political discourse about “bathroom bills”. In M. N. Goins (Ed.), The Routledge handbook of gender and communication (pp. 274-288). Routledge. COVID-19 Hate Crimes Act, Pub.L. 117-13 (2021). Cox, A., & Adam, A. (2018). The bystander effect in non-emergency situations: Influence of gender and group size. Modern Psychological Studies, 23(2), 3. 187 Czopp, A. M. (2010). Studying is lame when he got game: Racial stereotypes and the discouragement of Black student-athletes from schoolwork. Social Psychology of Education, 13(4), 485-498. Czopp, A. M. (2019). The consequences of confronting prejudice. In R. K. Mallett and M. J. Monteith (Eds.), Confronting prejudice and discrimination: The science of changing minds and behaviors (pp. 201-221). Academic Press. Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8(4), 377-383. Dasgupta, N., & Greenwald, A. G. (2001). On the malleability of automatic attitudes: Combating automatic prejudice with images of admired and disliked individuals. Journal of Personality and Social Psychology, 81, 800-814. Davidovic, A., Joinson, A., Talbot, C., & Hamilton-Giachritsis, C. (2023). To intervene or not to intervene: Young adults’ views on when and how to intervene in online harassment. Journal of Computer-Mediated Communication. Advance online publication. Davis, R. L., & O’Neill, P. (2016, May). “The Hate Crimes Reporting Gap: Low Numbers Keep Tensions High,”. The Police Chief , 83. https://www.policechiefmagazine.org/the-hate-crimes/?ref=0ca264135c970468c26b0e5118245b4d De Houwer, J. (2019). Implicit bias is behavior: A functional-cognitive perspective on implicit bias. Perspectives on Psychological Science, 14(5), 835-840. Deacon, J., Reichler, M., Saylor, K., & Maxwell, P. (2020). 104-Gender differences in bystander responses to male violence against a transwoman. GREAT Day. https://knightscholar.geneseo.edu/great-day-symposium/great-day-2020/posters-2020/104 188 Decety, J., & Michalska, K. J. (2010). Neurodevelopmental changes in the circuits underlying empathy and sympathy from childhood to adulthood. Developmental Science, 13(6), 886-899. Del Visco, S. (2019). Yellow peril, red scare: Race and communism in National Review. Ethnic and Racial Studies, 42(4), 626-644. Dennison, C. R., & Finkeldey, J. G. (2021). Self-reported experiences and consequences of unfair treatment by police. Criminology, 59, 254-290. Department of Justice. (n.d.) Learn about hate crimes. https://www.justice.gov/hatecrimes/learn-about-hate-crimes Department of Justice. (2023, December 31). Laws and policies. https://www.justice.gov/hatecrimes/laws-and-policies DeSante, C. D., & Smith, C. W. (2020). Racial stasis: The millennial generation and the stagnation of racial attitudes in American politics. The University of Chicago Press. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56(1), 5-18. Dovidio, J. F., & Gaertner, S. L. (2010). Intergroup bias. In S.T. Fiske, D.T. Gilbert, & G. Lindzey (Eds.), Handbook of social psychology (5th ed., pp. 1084-1121). Wiley. Eisenberg, A. (2014). Expressive enforcement. UCLA Law Review, 61, 858-926. Erentzen, C., Schuller, R. A., & Gardner, R. C. (2021). Model victims of hate: Victim blaming in the context of Islamophobic hate crime. Journal of Interpersonal Violence, 36(11-12), 5422-5445. 189 Erfanian, F., Roudsar, R. L., Haidari, A., & Bahmani, M. N. D. (2020). A narrative on using vignettes: Its advantages and disadvantages. Journal of Midwifery & Reproductive Health, 8(2), 2134-2145. Exum, M. L., Bailey, D., & Wright, E. L. (2014). False positive and false negative rates in self-reported intentions to offend: A replication and extension. Journal of Criminal Justice, 42, 1-9. Exum, M. L., & Layana, M. C. (2018). A test of the predictive validity of hypothetical intentions to offend. Journal of Crime and Justice, 41(2), 136-154. Exum, M. L., Turner, M. G., & Hartman, J. L. (2012). Self-reported intentions to offend: All talk and no action? American Journal of Criminal Justice, 37, 523-543. Famouri, M. L., Hernandez, S., Omlor, R. L., Lane-Brown, M., Evans, S. M., McIntosh, D., & Denizard-Thompson, N. (2023). Active bystander training: Using standardized patient methodology to teach residents to navigate microaggressions in patient encounters. MedEdPORTAL, 19, 11298. Farrell, A., & Lockwood, S. (2023). Addressing hate crime in the 21st century: Trends, threats, and opportunities for intervention. Annual Review of Criminology, 6, 107-130. Fischer, P., Krueger, J. I., Greitemeyer, T., Vogrincic, C., Kastenmüller, A., Frey, D., Heene, M., Wicher, M., & Kainbacher, M. (2011). The bystander-effect: A meta-analytic review on bystander intervention in dangerous and non-dangerous emergencies. Psychological Bulletin, 137(4), 517-537. Franklin, T. W., & Fearn, N. E. (2015). Sentencing Asian offenders in state courts: The influence of a prevalent stereotype. Crime & Delinquency, 61(1), 96-120. 190 Freilich, J. D., & Chermak, S. M. (2013). Hate crimes. Problem-oriented guides for police, problem-specific guides (Series, No. 72). Office of Community Oriented Policing Service. Washington, DC: U.S. Department of Justice. Retrieved from https://popcenter.asu.edu/sites/default/files/sites/default/files/problems/pdfs/hate_crimes.pdf Gaertner, S. L., Dovidio, J. F., & Johnson, G. (1982). Race of victim, nonresponsive bystanders, and helping behavior. The Journal of Social Psychology, 117, 69-77. Gaertner, S. L., Dovidio, J. F., Anastasio, P. A., Bachman, B. A., & Rust, M. C. (1993). The common ingroup identity model: Recategorization and the reduction of intergroup bias. European Review of Social Psychology, 4(1), 1-26. Gamberini, L., Chittaro, L., Spagnolli, A., & Carlesso, C. (2015). Psychological response to an emergency in virtual reality: Effects of victim ethnicity and emergency type on helping behavior and navigation. Computers in Human Behavior, 48, 104-113. Gans, H. J. (2012). “Whitening” and the changing American racial hierarchy. Cambridge University Press. Gettys, C. F., Pliske, R. M., Manning, C., & Casey, J. T. (1987). An evaluation of human act generation performance. Organizational behavior and Human Decision Processes, 39, 23-51. Gini, G., Albiero, P., Benelli, B., & Altoè, G. (2007). Does empathy predict adolescents’ bullying and defending behavior? Aggressive Behavior, 33, 467-476. Godzisz, P., & Mazurczak, J. (2023). Not ideal, but still acknowledged: A 10-country survey on empathy for victims of anti-LGBT violence. Journal of Interpersonal Violence, 38(9-10), 6751-6772. 191 Gordon, S. L. (2020). Understanding xenophobic hate crime in South Africa. Journal of Public Affairs, 20(3), e2076. Gover, A. R., Harper, S. B., & Langon, L. (2020). Anti-Asian hate crime during the COVID-19 pandemic: Exploring the reproduction of inequality. American Journal of Criminal Justice, 45, 647-667. Graham, A., Haner, M., Sloan, M. M., Cullen, F. T., Kulig, T. C., & Jonson, C. L. (2020). Race and worrying about police brutality: The hidden injuries of minority status in America. Victims & Offenders, 15(5), 549-573. Graham, K., LaChance, A., & Wormwood, J. B. (2023). Bystander intervention in intimate partner violence: An audio vignette study of heterosexual, gay, and lesbian dating partners. Journal of Interpersonal Violence, 38(3-4), NP3979-NP4006. Grasso, J. C., Jenness, V., & Vogler, S. (2023). Understanding the context of police avoidance: The impact of sexual identity, police legitimacy and legal cynicism on willingness to report hate crime. Current Issues in Criminal Justice, 35(2), 269-289. Grattet, R., Jenness, V., & Curry, T. R. (1998). The homogenization and differentiation of hate crime law in the United States, 1978 to 1995: Innovation and diffusion in the criminalization of bigotry. American Sociological Review, 63(2), 286-307. Greenwald, A. G., & Krieger, L. H. (2006). Implicit bias: Scientific foundations. California Law Review, 94(4), 945-967. Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102(1), 4-27. Greitemeyer, T., Fischer, P., Kastenmüller, A., & Frey, D. (2006). Civil courage and helping behavior: Differences and similarities. European Psychologist, 11, 90-98. 192 Griffin, R., Quasem, M., Sides, J., & Tesler, M. (2021). Racing apart: Partisan shifts on racial attitudes over the last decade. Democracy Fund Voter Study Group. https://search.issuelab.org/resources/39124/39124.pdf Grossman, A., D’Augelli, A., Salter, N., & Hubbard, S. (2005). Comparing gender expression, gender nonconformity, and parents’ responses of female-to-male and male-to-female transgender youth: Implications for counseling. Journal of LGBT Issues in Counseling, 1, 41-59. Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2001). Inside the judicial mind. Cornell Law Review, 86(4), 777-830. Halvorson, E. (2021). ‘Sticks and stones’: Experiencing microaggressions from the perspectives of the victim, bystander, and perpetrator. [Honors thesis, University of Tennessee at Chattanooga]. UTC Scholar. https://scholar.utc.edu/honors-theses/295/ Hamby, S., Weber, M. C., Grych, J., & Banyard, V. (2015). What difference do bystanders make? The association of bystander involvement with victim outcomes in a community sample. Psychology of Violence, 6(1), 91-102. Han, S., Riddell, J. R., & Piquero, A. R. (2023). Anti-Asian American hate crimes spike during the early stages of the COVID-19 pandemic. Journal of Interpersonal Violence, 38(3-4), 3513-3533. Hartman, C. (2023). Don’t Say Gay: How laws are tools for hate, discrimination, and violence. Electronic Theses, Projects, and Dissertations. 1572. Hate Crimes Sentencing Enhancement Act, 28 U.S.C. § 994 (1994). Hate Crimes Statistics Act, 28 U.S.C. § 534 (1990). 193 Haynes-Baratz, M. C., Metinyurt, T., Li, Y. L., Gonzales, J., & Bond, M. A. (2021). Bystander training for faculty: A promising approach to tackling microaggressions in the academy. New Ideas in Psychology, 63, 100882. Heen, M. S. J., Leiberman, J. D., & Miethe, T. D. (2014). A comparison of different online sampling approaches for generating national samples. Center for Crime and Justice Policy. https://www.unlv.edu/sites/default/files/page_files/27/ComparisonDifferentOnlineSampling.pdf Herek, G. M., Cogan, J. C., & Gillis, J. R. (2002). Victim experiences in hate crimes based on sexual orientation. Journal of Social Issues, 58(2), 319-339. Herek, G. M., Gillis, J. R., Cogan, J. C., & Glunt, E. K. (1997). Hate crime victimization among lesbian, gay, and bisexual adults: Prevalence, psychological correlates, and methodological issues. Journal of Interpersonal Violence, 12(2), 195-215. Hersh, E., & Royden, L. (2023). Antisemitic attitudes across the ideological spectrum. Political Research Quarterly, 76(2), 697-711. Hoffman, M. L. (1990). Empathy and justice motivation. Motivation and Emotion, 14(2), 151-172. Hosmer, D. W., & Lemeshow, S. (1989). Applied logistic regression. Wiley. Howerton, A. (2006). Police response to crime: Differences in the application of law by race. Journal of Ethnicity in Criminal Justice, 4(3), 51-63. Hoxmeier, J. C., Mennicke, A., & McMahon, S. (2022). Bystander intervention opportunities and prosocial behaviors among gender and sexual minority college students. Journal of Interpersonal Violence, 37(9-10), NP6439-NP6465. 194 Huang, L., Li, W., Xu, Z., Sun, H., Ai, D., Hu, Y., Wang, S., Li, Y., & Zhou, Y. (2023). The severity of cyberbullying affects bystander intervention among college students: The roles of feelings of responsibility and empathy. Psychology Research and Behavior Management, 16, 893-903. Hubbell, J. T. (2024). Elucidating intimate partner violence rate disparities between same- and opposite-sex couples: A demographic approach. Journal of Interpersonal Violence, 39(3-4), 651-675. Hughes, R., & Huby, M. (2004). The construction and interpretation of vignettes in social research. Social Work & Social Sciences Review, 11(1), 36-51. Huston, T. L., Ruggiero, M., Conner, R., & Geis, G. (1981). Bystander intervention into crime: A study based on naturally-occurring episodes. Social Psychology Quarterly, 44(1), 14-23. Huynh, V. W. (2012). Ethnic microaggressions and the depressive and somatic symptoms of Latino and Asian American adolescents. Journal of Youth and Adolescence, 41, 831-846. Iganski, P. (2002). Hate crimes hurt more. American Behavioral Scientist, 45, 626-638. Imhoff, R., & Banse, R. (2009). Ongoing victim suffering increase prejudice: The case of secondary anti-Semitism. Psychological Science, 20(12), 1443-1447. International Association of Chiefs of Police. (2019). Responding to hate crimes: A police officer’s guide to investigation and prevention [Brochure]. https://www.theiacp.org/sites/default/files/2019-01/IACP-Hate_Crimes_Brochure.pdf Ipsos. (2020). New center for public integrity/Ipsos poll finds most Americans say the Coronavirus pandemic is a natural disaster. https://www.ipsos.com/en-us/newspolls/center-for-public-integrity-poll-2020 195 Jackson, J., McKay, T., Cheliotis, L., Bradford, B., Fine, A., & Trinkner, R. (2023). Centering race in procedural justice theory: Structural racism and the under- and overpolicing of black communities. Law and Human Behavior, 47(1), 68-82. Jenkins, L. N., Fredrick, S. S., & King, J. (2023). Testing the bystander intervention model in cyberbullying across students of color and white students. International Journal of Bullying Prevention. Advance online publication. https://doi.org/10.1007/s42380-023-00162-0 Jenkins, L. N., Synder Kaminski, S., & Miller, M. (2021). Bystander intervention in bullying: Differences across latent profiles. International Journal of Bullying Prevention, 3, 130-137. Jensen, J., & Raver, J. (2020). A policy capturing investigation of bystander decisions to intervene against workplace incivility. Journal of Business and Psychology, 36(5), 883-901. Jeong, D., Aggarawal, S., Ronbinson, J., Kumar, N., Spearot, A., & Sungho Park, D. (2023). Exhaustive or exhausting? Evidence on respondent fatigue in long surveys. Journal of Development Economics, 161, 102992. Jia, Y., Wu, Y., Jin, T., & Zhang, L. (2022). How are bystanders involved in cyberbullying? A latent class analysis of the cyberbystander and their characteristics in different intervention stages. International Journal of Environmental Research and Public Health, 19(23), 16083. Johnson, B. D., & Betsinger, S. (2009). Punishing the “model minority”: Asian-American criminal sentencing outcomes in federal district courts. Criminology, 47(4), 1045-1090. 196 Jones, J. M. (2023a, June 12). More say birth gender should dictate sports participation. Gallup. https://news.gallup.com/poll/507023/say-birth-gender-dictate-sports-participation.aspx Jones, J. M. (2023b, June 16). Fewer in U.S. say same-sex relations morally acceptable. Gallup. https://news.gallup.com/poll/507230/fewer-say-sex-relations-morally-acceptable.aspx Jones, J. M. (2023c, February 22). U.S. LGBT identification steady at 7.2%. Gallup. https://news.gallup.com/poll/470708/lgbt-identification-steady.aspx Jones, L. (2015). Masculinity in lesbian discourse: The case of butch and femme. In Tommasso M. Milani (Ed.) Language and Masculinities: Performances, intersections, dislocations (pp. 174-196). Routledge. Jones, P. E., Brewer, P. R., Young, D. G., Lambe, J. L., & Hoffman, L. H. (2018). Explaining public opinion toward transgender people, rights, and candidates. Public Opinion Quarterly, 82(2), 252-278. Kahneman, D., & Tversky, A. (1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3, 430-454. Katz, J., Federici, D., & Brown, D. (2022). Effects of humor and bystander gender on responses to antigay harassment. Journal of Homosexuality, 69(6), 947-966. Katz, J., Federici, D., & Ramos-Dries, T. (2019a). Amnestic heterosexism and bystander responses to anti-gay bullying. Journal of Homosexuality, 66(1), 1-16. Katz, J., Grant, C., & Merrilees, C. (2019b). Just joking? White college students’ responses to different types of racist comments. Journal of Diversity in Higher Education, 12(4), 341-350. 197 Katz, J., Merriliees, C., Hoxmeier, J. C., & Motisi, M. (2017). White female bystanders’ responses to a black woman at risk for incapacitated sexual assault. Psychology of Women Quarterly, 41(2), 273-285. Katz, J., Pazienza, R., Olin, R., & Rich, H. (2015). That’s what friends are for: Bystander responses to friends or strangers at risk for party rape victimization. Journal of Interpersonal Violence, 30(16), 2775-2792. Keleher, A., & Smith, E. R. A. N. (2012). Growing support for gay and lesbian equality since 1990. Journal of Homosexuality, 59(9), 1307-1326. Kena, G., & Thompson, A. (2021). Hate crime victimization, 2005–2019. Rep. NCJ 30054, US Department of Justice, Washington, DC. https://bjs.ojp.gov/sites/g/files/xyckuh236/files/media/document/hcv0519_1.pdf Kettrey, H. H., & Marx, R. A. (2021). Effects of bystander sexual assault prevention programs on promoting intervention skills and combatting the bystander effect: A systematic review and meta-analysis. Journal of Experimental Criminology, 17(3), 343-367. Kim, C., Lee, C. S., & Lim, H. (2023). Hate-motivated crime/incidents against Asians in the United States of America: A systematic review. Race and Justice, 13(1), 9-31. Kim, J. Y. (2006). Hate crime law and the limits of inculpation. Nebraska Law Review, 84, 846-894. Kim, J. Y., & Kesari, A. (2021). Misinformation and hate speech: The case of anti-Asian hate speech during the COVID-19 pandemic. Journal of Online Trust and Safety, 1(1), https://doi.org/10.54501/jots.v1i1.13 198 Kim, N. (2007). Asian Americans’ experiences of “race” and racism. In H. Vera, & J. R. Feagin (Eds.), Handbook of the Sociology of Racial and Ethnic Relations (pp. 131-144). Springer. Kim, N. Y. (2007). Critical thoughts on Asian American assimilation in the whitening literature. In M. Zhou & A. C. Ocampo (Eds.), Contemporary Asian America (pp. 554-575). New York University Press. Kirk-Provencher, K. T., Spillane, N. S., Schick, M. R., Chalmers, S. J., Hawes, C., & Orchowski, L. M. (2023). Sexual and gender minority inclusivity in bystander intervention programs to prevent violence on college campuses: A critical review. Trauma, Violence, & Abuse, 24(1), 110-124. Kosciw, J. G., Clark, C. M., & Menard, L. (2022). The 2021 National School Climate Survey: The experiences of LGBTQ+ youth in our nation’s schools. GLSEN. Kreuter, F., Presser, S., & Tourangeau, R. (2008). Social desirability bias in CATI, IVR, and web surveys: The effects of mode and question sensitivity. Public Opinion Quarterly, 72(5), 847-865. Kunstman, J. W., & Plant, E. A. (2008). Racing to help: Racial bias in high emergency helping situations. Journal of Personality and Social Psychology, 95, 1499-1510. Labhardt, D., Holdsworth, E., Brown, S., & Howat, D. (2017). You see but you do not observe: A review of bystander intervention and sexual assault on university campuses. Aggression and Violent Behavior, 35, 13-25. Lantz, B., & Kim, J. (2019). Hate crimes hurt more, but so do co-offenders: Separating the influence of co-offending and bias on hate-motivated physical injury. Criminal Justice and Behavior, 46(3), 437-456. 199 Lantz, B., & Wenger, M. R. (2020). Bias and hate crime victimization during the COVID-19 pandemic: Project report. Florida State University Hate Crime Research and Policy Institute. Lantz, B., & Wenger, M. R. (2022). Are Asian victims less likely to report hate crime victimization to the police? Implications for research and policy in the wake of the COVID-19 pandemic. Crime & Delinquency, 68(8), 1292-1319. Lantz, B., Gladfelter, A. S., & Ruback, R. B. (2019). Stereotypical hate crimes and criminal justice processing: A multi-dataset comparison of bias crime arrest patterns by offender and victim race. Justice Quarterly, 36(2), 193-224. Latané, B. & Darley, J. M. (1968). Group inhibition of bystander intervention in emergencies. Journal of Personality and Social Psychology, 10(3), 215-221. Latané, B. & Darley, J. M. (1970). The unresponsive bystander: Why doesn’t he help? New York: Appleton-Century-Crofts. Latané, B., & Nida, S. (1981). Ten years of research on group size and helping. Psychological Bulletin, 89(2), 308-324. Lee, E. (2007). The “yellow peril” and Asian exclusion in the Americas. Pacific Historical Review, 76(4), 537-562. Leets, L. (2002). Experiencing hate speech: Perceptions and responses to anti-Semitism and antigay speech. Journal of Social Issues, 58(2), 341-361. Levine, M., Cassidy, C., Brazier, G., & Reicher, S. (2002). Self-categorization and bystander non-intervention: Two experimental studies. Journal of Applied Social Psychology, 32(7), 1452-1463. 200 Li, Y., & Nicholson Jr., H. L. (2021). When “model minorities” become “yellow peril”—Othering and the racialization of Asian Americans in the COVID-19 pandemic. Sociology Compass, 15(2), e12849. Libby, C. (2022). Sympathy, fear, hate: Trans-exclusionary radical feminism and evangelical Christianity. Transgender Studies Quarterly, 9(3), 425-442. Liebst, L. S., Philpot, R., Bernasco, W., Dausel, K. L., Ejbye-Ernst, P., Nicolaisen, M. H., & Lindegaard, M. R. (2019). Social relations and presence of others predict bystander intervention: Evidence from violent incidents captured on CCTV. Aggressive Behavior, 45, 598-609. Lilienfeld, S. O. (2017). Microaggressions: Strong claims, inadequate evidence. Perspectives on Psychological Science, 12, 138-169. Locantore, D. C., & Wasarhaley, N. E. (2020). Mentally ill, HIV-positive, or sexual predator? Determining myths perceived as representative of transgender people. Group Processes & Intergroup Relations, 23(3), 418-440. Mainwaring, C., Gabbert, F., & Scott, A. J. (2023). A systematic review exploring variables related to bystander intervention in sexual violence contexts. Trauma, Violence, & Abuse, 24(3), 1727-1742. Mallory, C., Hasenbush, A., & Sears, B. (2015). Discrimination and harassment by law enforcement officers in the LGBT communities. http://williamsinstitute.law.ucla.edu/wp-content/uploads/LGBT-Discrimination-andHarassment-in-Law-Enforcement-March-2015.pdf Marsden, P. V., Smith, T. W., & Hout, M. (2020). Tracking US social change over a half-century: The General Social Survey at fifty. Annual Review of Sociology, 46, 109-134. 201 Mason, G. (2014). The symbolic purpose of hate crime law: Ideal victims and emotion. Theoretical Criminology, 18(1), 75-92. Masucci, M., & Langton, L. (2017). Hate crime victimization, 2004-2015. U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics. Matal v. Tam, 582 U.S. (2017) Matthew Shepard and James Byrd, Jr. Hate Crimes Prevention Act, 18 U.S.C. § 249 (2009). Mazar, L. (2019). History and theoretical understanding of bystander intervention. In W. T. O’Donohue and P. A. Schewe (Eds.), Handbook of sexual assault and sexual assault prevention (pp. 423-432). Springer. Mazurczak, J., & Godzisz, P. (2019). Hates and allies: Attitudes towards LGBT people as victims of crimes in 10 EU states. In P. Godzisz & G. Viggiani (Eds.), Awareness of Anti-LGBT Hate Crimes in the European Union (pp. 26-52). Lambda Warsaw. McClure, E. (2020). Escalating linguistic violence: From microaggressions to hate speech. In L. Freeman & J. Weekes Schroer (Eds.) Microaggressions and philosophy (pp. 121-145). Routledge. McCutcheon, A. L. (1987). Latent class analysis. Sage Publications. McDevitt, J., Cuevas, C. A., Wills, C., Lawshe, N., Ikimalo, O., Iwama, J., deDufour, A., Shukur, P., Kraus, M., & Feinman, A. (2021). School hate crime resource guide. Massachusetts Executive Office of Public Safety and Security. Retrieved from https://www.mass.gov/doc/school-hate-crime-resource-guide/download McDevitt, J., Levin, J., Nolan, J., & Bennett, S. (2010). Hate crime offenders. In N. Chakraborti (Ed.), Hate crime: Concepts, policy, future directions (pp. 125–145). Willan Publishing. McDonald, J. H. (2014). Handbook of biological statistics (3rd ed.). Sparky House Publishing. 202 McMahon, S., & Dick, A. (2011). “Being in a room with like-minded men”: An exploratory study of men’s participation in a bystander intervention program to prevent intimate partner violence. The Journal of Men’s Studies, 19, 3-18. McMahon, S., Peterson, N. A., Winter, S. C., Palmer, J. E., Postmus, J. L., & Koenick, R. A. (2015). Predicting bystander behavior to predict sexual assault on college campuses: The role of self-efficacy and intent. American Journal of Community Psychology, 56, 46-56. McMahon, S., Treitler, P., Peterson, N. A., & O’Connor, J. (2019). Bystander intentions to intervene and previous sexual violence education: A latent class analysis. Psychology of Violence, 9(1), 117-126. Meyer, D. (2012). An intersectional analysis of lesbian, gay, bisexual, and transgender (LGBT) people’s evaluations of anti-queer violence. Gender & Society, 26(6), 849-873. Miklikowska, M. (2017). Empathy trumps prejudice: The longitudinal relation between empathy and anti-immigrant attitudes in adolescence. Developmental Psychology, 54(4), 703-717. Mize, T. D. (2019). Best practices for estimating, interpreting, and presenting nonlinear interaction effects. Sociological Science, 6, 81-117. Morgan, C. J. (1978). Bystander intervention: Experimental test of a formal model. Journal of Personality and Social Psychology, 36, 43-55. Moroz, S., & Campbell, L. (2018). Where there’s smoke, there’s fire: Error management theory as a framework for the perception of heterosexist microaggressions. PsyArXiv. https://doi.org/10.31234/osf.io/t5uan Moschella, E. A., & Banyard, V. L. (2021). Action and reaction: The impact of consequences of intervening in situations of interpersonal violence. Journal of Interpersonal Violence, 36(7-8), NP3820-NP3843. 203 Motivans, M. (2021). Federal hate crime prosecutions, 2005-19. Bureau of Justice Statistics. Department of Justice. Moynihan, M., Banyard, V. J., Arnold, J. S., Eckstein, R. P., & Stapleton, J. G. (2011). Sisterhood may be powerful for reducing sexual and intimate partner violence: An evaluation of the brining in the bystander in-person program with sorority members. Violence Against Women, 17(6), 703-719. Mullen, B., & Smyth, J. M. (2004). Immigrant suicide rates as a function of ethnophaulilsms: Hate speech predicts death. Psychosomatic Medicine, 66(3), 343-348. Mullen, B., Brown, R., & Smith, C. (1992). Ingroup bias as a function of salience, relevance, and status: An integration. European Journal of Social Psychology, 22, 103-122. Mulvey, K. L., Palmer, S. B., & Abrams, D. (2016). Race-based humor and peer group dynamics in adolescence: Bystander intervention and social exclusion. Child Development, 87(5), 1379-1391. Mutz, L. (2006). A fairy tale: The myth of the homosexual lifestyle in anti-gay-and-lesbian rhetoric. Women’s Rights Law Reporter, 27(2), 69-86. Nadal, K. L. (2011). The Racial and Ethnic Microaggressions Scale (REMS): Construction, reliability, and validity. Journal of Counseling Psychology, 58, 470-480. Nadal, K. L., Griffin, K. E., Wong, Y., Davidoff, K. C., & Davis, L. S. (2017). The injurious relationship between racial microaggressions and physical health: Implications for social work. Journal of Ethnic & Cultural Diversity in Social Work, 26(1-2), 6-17. Nadal, K. L., Griffin, K. E., Wong, Y., Hamit, S., & Rasmus, M. (2014). The impact of racial microaggressions on mental health: Counseling implications for clients of color. Journal of Counseling and Development, 92(1), 57-66. 204 Nadal, K. L., Rivera, D. P., & Corpus, M. J. H. (2010). Sexual orientation and transgender microaggressions in everyday life: Experiences of lesbians, gays, bisexuals, and transgender individuals. In D. W. Sue (Ed.), Microaggressions and marginality: Manifestation, dynamics, and impact (pp. 217-240). John Wiley & Sons. Nadal, K. L., Skolnik, A., & Wong, Y. (2012). Interpersonal and systemic microaggressions toward transgender people: Implications for counseling. Journal of LGBT Issues in Counseling, 6(1), 55-82. Nagin, D. S. (2005). Group-based modeling of development. Harvard University Press. Nam, H. H., Sawyer, K., & Style, H. K. (2022). Understanding anti-Asian sentiment and political behavior in the wake of COVID-19. Politics, Groups, and Identities. Advance online publication. https://doi.org/10.1080/21565503.2022.2137051 Newell, A., Shaw, J. C., & Simon, H. A. (1957). Empirical explorations of the logic theory machine: A case study in heuristic. Proceedings of the 1957 Western Joint Computer Conference, February 26-28, 230-240. Nickerson, A. B., Aloe, A. M., & Werth, J. M. (2015). The relation of empathy and defending in bullying: A meta-analytic investigation. School Psychology Review, 44(4), 372-390. Nickerson, A. B., Aloe, A. M., Livingston, J. A., & Feeley, T. H. (2014). Measurement of the bystander intervention model for bullying and sexual harassment. Journal of Adolescence, 37, 391-400. Nisbett, R. E., Zukier, H., & Lemley, R. E. (1981). The dilution effect: Nondiagnostic information weakens the implications of diagnostic information. Cognitive Psychology, 13, 248-277. 205 Nolan, J. J., McDevitt, J., Cronin, S., & Farrell, A. (2004). Learning to see hate crimes: A framework for understanding and clarifying ambiguities in bias crime classification. Criminal Justice Studies, 17(1), 91-105. Nolan, J. J., Nicewarner, R. L., & Mom, R. E. (2020). Rural crime prevention. Routledge. Norton, A. T., & Herek, G. M. (2013). Heterosexuals’ attitudes toward transgender people: Findings from a national probability sample of U.S. adults. Sex Roles, 68, 738-753. Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14(4), 535-569. Oakes, P. J. (1987). The salience of social categories. In J. C. Turner, M. A. Hogg, P. J. Oakes, S. D. Reicher, S., & Wetherell, M. S. (Eds.). Rediscovering the social group: A Self-categorization theory (pp. 117-141). Blackwell. Oakes, P. J., Turner, J. C., & Haslam, S. A. (1991). Perceiving people as group members: The role of fit in the salience of social categorizations. British Journal of Social Psychology, 30, 125-144. Oakley, C. M. (2022). Curriculum censorship of LGBTQ+ identity: Modern adaptation of vintage “Save Our Children” rhetoric is still just discrimination. Loyola University Chicago Law Journal, 54, 641-681. Okihiro, G. Y. (2014). Margins and mainstreams: Asians in American history and culture. University of Washington Press. Olivier, J., & Bell, M. L. (2013). Effect sizes for 2x2 contingency tables. PLOS One, 8(3), e58777. 206 Olson, L. R., Cadge, W., & Harrison, J. T. (2006). Religion and public opinion about same-sex marriage. Social Science Quarterly, 87(2), 340-360. Ong, A. D., Burrow, A. L., Fuller-Rowell, T. E., Ja, N. M., & Sue, D. W. (2013). Racial microaggressions and daily well-being among Asian Americans. Journal of Counseling Psychology, 60(2), 188-199. Osborn, M. (2022). U.S. news coverage of transgender victims of fatal violence: An exploratory content analysis. Violence Against Women, 28(9), 2033-2056. Osswald, S., Greitemeyer, T., Fischer, P., & Frey, D. (2010). What is moral courage? Definition, explication, and classification of a complex construct. In C. L. Pury & S. J. Lopez (Eds.), The psychology of courage: Modern research on an ancient virtue (pp. 149-164). American Psychological Association. Otten, S., Mummendey, A., & Buhl, T. (1998). Information processing and social discrimination: Accuracy in information processing and positive-negative symmetry in social discrimination. Revue Internationale de Psychologie Sociale, 11, 69-96. Owuamalam, C. K., & Matos, A. S. (2020). When might heterosexual men be passive or compassionate toward gay victims of hate crime? Integrating the bystander and social loafing explanations. Archives of Sexual Behavior, 49, 1693-1709. Owuamalam, C. K., & Matos, A. S. (2022). Heterosexual men in Trump’s America downplay compassion more for masculine (than for feminine) gay victims of hate crime: Why?. European Journal of Social Psychology, 52, 280-304. Palmer, N. A., & Kutateladze, B. L. (2022). What prosecutors and the police should do about underreporting of anti-LGBTQ hate crime. Sexuality Research and Social Policy, 19, 1190-1204. 207 Paluck, E. L., & Green, D. P. (2009). Prejudice reduction: What works? A review and assessment of research and practice. Annual Review of Psychology, 60, 339-367. Papieska, J., Spilt, J. L., Roorda, D. L., & Laevers, F. (2019). Promoting socioemotional competence in primary school classrooms: Intervention effects of the EMOscope. European Journal of Developmental Psychology, 16(1), 97-112. Parekh, B. (2012). Is there a case for banning hate speech? In M. Herz, & P. Molnar (Eds.), The Content and Context of Hate Speech: Rethinking Regulation and Responses (pp. 37-56). Cambridge University Press. Park, S., & Kim., S. (2023). A systematic review and meta-analysis of bystander intervention programs for intimate partner violence and sexual assault. Psychology of Violence, 13(2), 93-106. Parker, L., Webb, S., & Chonody, J. M. (2023). Bullying toward LGBTQI+ students in Australian schools: Understanding teachers’ intentions to intervene. Journal of LGBT Youth, 20(3), 561-584. Paternoster, R., Mamayek Jaynes, C., & Wilson, T. (2017). Rational choice theory and interest in the “fortune of others”. Journal of Research in Crime and Delinquency, 54(6), 847-868. Perry, B., & Alvi, S. (2012). ‘We are all vulnerable’: The in terrorem effects of hate crimes. International Review of Victimology, 18(1), 57-71. Peterson, W. (1966, January 6). Success story: Japanese American style. New York Times Magazine, p. 11. Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65-85. Pezzella, F. S., & Fetzer, M. D. (2017). The likelihood of injury among bias crimes: An analysis of general and specific bias types. Journal of Interpersonal Violence, 32(5), 703-729. 208 Pezzella, F. S., & Fetzer, M. D. (2021). The measurement of hate crimes in America. Springer. Pezzella, F., S., Fetzer, M. D., & Keller, T. (2019). The dark figure of hate crime underreporting. American Behavioral Scientist. Advance online publication. https://doi.org/10.1177/0002764218823844 Piliavin, J. A., Dovidio, J. F., Gaertner, S. L., & Clark, R. D., III. (1981). Emergency intervention. Academic Press. Platt, L. F., & Lenzen, A. L. (2013). Sexual orientation microaggressions and the experiences of sexual minorities. Journal of Homosexuality, 60, 1011-1034. Quinn, A. (2021, October 22). Bed-Stuy man indicted in anti-gay attack at Brooklyn bodega: DA. Patch. https://patch.com/new-york/bed-stuy/bed-stuy-man-indicted-anti-gay-attack-brooklyn-bodega-da Ratcliff, J. J., Andrus, T., Miller, A. K., Olowu, F., & Capellupo, J. (2023). When potential allies and targets do (and do not) confront anti-Asian prejudice: Reactions to blatant and subtle prejudice during the COVID-19 pandemic. Journal of Interpersonal Violence. Advance online publication. https://doi.org/10.1177/08862605231188057 Rennison, C. M., Gover, A. R., Bosick, S. J., & Dodge, M. (2011). Reporting violent victimization to the police: A focus on Black, White, Asian, and Hispanic adolescent victims. The Open Family Studies Journal, 4, 54-67. Reny, T. T., & Barreto, M. A. (2022). Xenophobiai in the time of the pandemic: Othering, anti-Asian attitudes, and COVID-19. Politics, Groups, and Identities, 10(2), 209-232. Right to Be. (n.d.). The 5Ds of Bystander Intervention. https://righttobe.org/guides/bystander-intervention-training/ 209 Ruback, R. B., Gladfelter, A. S., & Lantz, B. (2015). An examination of hate and bias incidents in Pennsylvania, 1999-2012. The Center for Rural Pennsylvania. Rubin, D. I., & Wilson, F. A. (2021). A time of covidiocy: Media, politics, and social upheaval. Brill. Rutland, A., Killen, M., & Abrams, D. (2010). A new social-cognitive developmental perspective on prejudice: The interplay between morality and group identity. Perspectives on Psychological Science, 5(3), 279-291. Sagir, M., & Mockabee, S. T. (2023). Public attitudes toward immigration: Was there a Trump effect?. American Politics Research, 51(3), 381-396. Schafer, J. R., & Navarro, J. (2003). The seven-stage hate model: The psychopathology of hate groups. FBI Law Enforcement Bulletin, 72, 1-7. Schipper, K. (2022). Conservative political tactics to obstruct gender-affirming healthcare for transgender youth in the USA [Master’s thesis, Columbia University]. Columbia Academic Commons. https://academiccommons.columbia.edu/doi/10.7916/keds-5594 Schultze-Krumbholz, A., Hess, M., Pfetsch, J., & Scheithauer, H. (2018). Who is involved in cyberbullying? Latent class analysis of cyberbullying roles and their associations with aggression, self-esteem, and empathy. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 12(4). Schweppe, J., & Perry, B. (2022). A continuum of hate: Delimiting the field of hate studies. Crime, Law, and Social Change, 77, 503-528. Seaton, E. K., Gee, G. C., Neblett, E., & Spanierman, L. (2018). New directions for racial discrimination research as inspired by the integrative model. American Psychologist, 73(6), 768-780. 210 Secord Fredrick, S., Jenkins, L. N., & Ray, K. (2020). Dimensions of empathy and bystander intervention in bullying in elementary school. Journal of School Psychology, 79, 31-42. Shea, C., Malone, M., Griffith, J. A., Staneva, V., Graham, K., & Banyard, V. (2021). Please feel free to intervene: A longitudinal analysis of the consequences of bystander behavioral expectations. Journal of Diversity in Higher Education. Advance online publication. https://doi.org/10.1037/dhe0000348 Shih, M., Wang, E., Trahan Bucher, A., & Stotzer, R. (2009). Perspective taking: Reducing prejudice towards general outgroups and specific individuals. Group Processes & Intergroup Relations, 12(5), 565-577. Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review, 63(2), 129-138. Sinclair, S. (2021). Bystander reactions to workplace incivility: The role of gender and discrimination claims. Europe’s Journal of Psychology, 17(1), 134-144. Sisak, M. R. (2021, April 6). Doormen fired for failing to intervene in anti-Asian attack. The Associated Press. https://apnews.com/article/doormen-fired-failing-to-intervene-anti-asian-attack-fe5dc2a00f7177b9f0963f5b27034783 Slaby, R. G., Wilson-Brewer, R., & Dash, K. (1994). Aggressors, victims, and bystanders: Thinking and acting to prevent violence. Education Development Center. Slocum, L. A. (2018). The effect of prior police contact on victimization reporting: Results from the police-public contact and national crime victimization surveys. Journal of Quantitative Criminology, 34, 535-589. Smith, T. W., Davern, M., Freese, J., & Morgan, S. L. (2019). General Social Surveys, 1972-2018: Cumulative Codebook. Principal Investigator, Tom W. Smith; Co-Principal 211 Investigators, Michael Davern, Jeremy Freese and Stephen L. Morgan. -- Chicago: NORC, 2019. 3,758 pp., 28cm. -- (National Data Program for the Social Sciences Series, no. 25). Smith, W. A., Hung, M., & Franklin, J. D. (2011). Racial battle fatigue and the miseducation of Black men: Racial microaggressions, societal problems, and environmental stress. The Journal of Negro Education, 63-82. Song, M. (2021). Is there evidence of ‘whitening’ for Asian/white multiracial people in Britain? Journal of Ethnic and Migration Studies, 47(4), 934-950. Steffensmeier, D., Painter-Davis, N., & Ulmer, J. (2017). Intersectionality of race, ethnicity, gender, and age on criminal punishment. Sociological Perspectives, 60(4), 810-833. Stephen, W. G., & Finaly, K. (1999). The role of empathy in improving intergroup relations. Journal of Social Issues, 55(4), 729-743. Stephens, K. C., Redman, T., Williams, R., Bandstra, B., & Shah, R. (2023). Considering culture and conflict: A novel approach to active bystander intervention. MedEdPORTAL, 19, 11338. Stroebe, W., & Frey, B. S. (1982). Self-interest and collective action: The economics and psychology of public goods. British Journal of Social Psychology, 21, 121-137. Stryker, S. (1998). The transgender issue: An introduction. GLQ: A Journal of Lesbian and Gay Issues, 4(2), 145-158. Sue, D. W. (2010). Microaggressions in everyday life: Race, gender, and sexual orientation. Wiley. 212 Sue, D. W., Alsaidi, S., Awad, M. N., Glaeser, E., Caelle, C. Z., & Mendez, N. (2019). Disarming racial microaggressions: Microintervention strategies for targets, white allies, and bystanders. American Psychologist, 74(1), 128-142. Sue, D. W., Capodilupo, C. M., Torino, G. C., Bucceri, J. M., Holder, A. M. B., Nadal, K. L., & Esquilin, M. (2007). Racial microaggressions in everyday life: Implications for clinical practice. American Psychologist, 62(4), 271-286. Sue, D., Capodilupo, C., & Holder, A. (2008). Racial microaggressions in the life experience of black Americans. Professional Psychology: Research and Practice, 39(3), 329-336. Tadlock, B. L., Flores, A. R., Haider-Markel, D. P., Lewis, D. C., Miller, P. R., & Taylor, J. K. (2017). Testing contact theory and attitudes on transgender rights. Public Opinion Quarterly, 81(4), 956-972. Thompson, C. G., Kim, R. S., Aloe, A. M., & Becker, B. J. (2017). Extracting the variance inflation factor and other multicollinearity diagnostics from typical regression results. Basic and Applied Social Psychology, 39(2), 81-90. Travis, S. (2023a, February 11). State questions Broward schools’ use of ADL anti-hate program. Sun Sentinel. https://www.sun-sentinel.com/2023/02/10/state-questions-broward-schools-use-of-adl-anti-hate-program/ Travis, S. (2023b, March 7). State won’t investigate Broward schools’ anti-hate program. Sun Sentinel. https://www.sun-sentinel.com/2023/03/06/state-wont-investigate-broward-schools-anti-hate-program/ Triplet, R. G. (1992). Discriminatory biases in the perception of illness: The application of availability and representativeness heuristics to the AIDS crisis. Basic and Applied Social Psychology, 13(3), 303-322. 213 Trout, M. (2015). Federalizing hate: Constitutional and practical limitations to the Matthew Shepard and James Byrd, Jr. Hate Crimes Prevention Act of 2009. American Criminal Law Review, 52(1), 131-154. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Blackwell. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5, 207-232. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124-1131. Uniform Crime Reporting. (2024). “Hate Crime Statistics, 2021”. Accessed 11 February 2024. van Earp, K. J. P. M., Gevers, J. M. P., Rispens, S., & Demerouti, E. (2018). Empowering public service workers to face bystander conflict: Enhancing resources through a training intervention. Journal of Occupational and Organizational Psychology, 91(1), 84-109. Vera, E. M., Galvin, S., Millon, F., Drapcho, R., & Payne, H. (2023). Culturally-motivated bullying and bystander intervention: The role of witness situational awareness, intervention history, and victim race. Journal of Prevention and Health Promotion, 4(1), 3-30. Vergani, M., Perry, B., Freilich, J., Chermak, S., Scrivens, R., & Link, R. (2022). PROTOCOL: Mapping the scientific knowledge and approaches to defining and measuring hate crime, hate speech, and hate incidents. Campbell Systematic Reviews, 18:e1228. Violent Crime Control and Law Enforcement Act, 42 U.S.C. § 136 (1994). 214 Wachs, S., Krause, N., Wright, M. F., & Gámez-Guadix, M. (2023). Effects of the prevention program “HateLess. Together against hatred” on adolescents’ empathy, self-efficacy, and countering hate speech. Journal of Youth and Adolescence, 52, 1115-1128. Waldron, J. (2012). The harm in hate speech. Harvard University Press. Wang, J., Leu, J., & Shoda, Y. (2011). When the seemingly innocuous “stings”: Racial microaggressions and their emotional consequences. Personality and Social Psychology Bulletin, 37(12), 1666-1678. Wang, S., & Kim, K. J. (2021). Effects of victimization experience, gender, and empathic distress on bystanders’ intervening behavior in cyberbullying. The Social Science Journal, 1-10. Waters, E. A., Pachur, T., Pogge, G., Hunleth, J., Webster, G. D., & Shepperd, J. A. (2023). Linking cognitive and affective heuristic cues to interpersonal risk perceptions and behavior. Risk Analysis. Advance online publication. https://doi-org/10.1111/risa.14101 Weller, B. E., Bowen, N. K., & Faubert, S. J. (2020). Latent class analysis: A guide to best practice. Journal of Black Psychology, 46(4), 287-311. Wenger, M. R., Lantz, B., & Gallardo, G. (2022). The role of hate crime victimization, fear of victimization, and vicarious victimization in COVID-19-related depression. Criminal Justice and Behavior,ili 49(12), 1746-1762. Whitehead, K. A., Bowman, B., & Raymond, G. (2018).”Risk factors” in action: The situated constitution of “risk” in violent interactions. Psychology of Violence, 8(3), 329-338.l White House. (n.d.). The Clinton Administration: Drawing a line against hate crimes. https://clintonwhitehouse4.archives.gov/Initiatives/OneAmerica/whc-accomp.html 215 Williams, M. T. (2021). Microaggressions are a form of aggression. Behavior Therapy, 52(3), 709-719. Wolff, K. B., & Cokely, C. L. (2007). “To protect and serve?”: An exploration of police conduct in relation to the gay, lesbian, bisexual, and transgender community. Sexuality and Culture, 11(2), 1-23. Wong, A. (2023, March 27). Education model wanted students ‘to be nice.’ That’s controversial in Florida now under DeSantis. USA Today. https://www.usatoday.com/story/news/education/2023/03/27/desantis-administrations-war-woke-targets-sel-florida/11414547002/ Wong, F., & Halgin, R. (2006). The “model minority”: Bane or blessing for Asian Americans? Journal of Multicultural Counseling and Development, 34(1), 38-49. Woods, J. B. (2008). Taking the “hate” out of hate crimes: Applying unfair advantage theory to justify the enhanced punishment of opportunistic bias crimes. UCLA Law Review, 56, 489-541. Worthen, M. G. F. (2013). An argument for separate analyses of attitudes toward lesbian, gay, bisexual men, bisexual women, MtF and FtM transgender individuals. Sex Roles, 68, 703-723. Worthen, M. G. F. (2016). Hetero-cis-normativity and the gendering of transphobia. International Journal of Transgenderism, 17(1), 31-57. Wu, D. J., Syropoulos, S., Rivera-Rodriguez, A., & Dasgupta, N. (2023). From model minority to yellow peril: How threat perceptions and disgust predict anti-Asian prejudice during COVID-19. Social and Personality Psychology Compass. Advance online publication https://doi.org/10.1111/spc3.12833 216 Wypych, M., & Bilewicz, M. (2022). Psychological toll of hate speech: The role of acculturation stress in the effects of exposure to ethnic slurs on mental health among Ukrainian immigrants in Poland. Cultural Diversity and Ethnic Minority Psychology. Advance online publication. https://doi.org/10.1037/cdp0000522 Xie, M., & Baumer, E. P. (2019). Crime victims’ decisions to call the police: Past research and new directions. Annual Review of Criminology, 2, 217-240. Xie, T., & Galliher, R. V. (2023). Responding to microaggressions: Social cost of bystander intervention strategies. The Counseling Psychologist, 51(2), 242-269. Xu, J., & Lee, J. C. (2013). The marginalized “model” minority: An empirical examination of the racial triangulation of Asian Americans. Social Forces, 91(4), 1363-1397. Yang, A. S. (1997). Trends: Attitudes toward homosexuality. The Public Opinion Quarterly, 61(3), 477-507. York, R. (2018). Control variables and causal inference: A question of balance. International Journal of Social Research Methodology, 21(6), 675-684. York, M., Langford, K., Davidson, M., Hemingway, C., Russell, R., Neeley, M., & Fleming, A. (2021). Becoming active bystanders and advocates: Teaching medical students to respond to bias in the clinical setting. MedEdPORTAL, 17, 11175. Zempi, I., Hamilton, P., Krulisova, K., & Trickett, L. (2021). Evaluation of “Citizens at the Heart: A citizen centered approach to tackling hate crime” project report. Nottingham Trent University. https://irep.ntu.ac.uk/id/eprint/42669/ 217 Appendix Table A1. Hate Crime Changes Over Time by Bias Type 2018-2022 2018 2019 2020 2021 2022 5-Year Average Change n n (Δ%) n (Δ%) n (Δ%) n (Δ%) % Bias Type Anti-American Indian or Alaska Native 183 101 (-44.8%) 108 (6.9%) 144 (33.3%) 194 (34.7%) 7.5% Anti-Arab 80 101 (26.3%) 80 (-20.8%) 103 (28.8%) 92 (-10.7%) 5.88% Anti-Asian 148 188 (27.0%) 334 (77.7%) 753 (125.4%) 499 (-33.7%) 49.1% Anti-Atheism/Agnosticism 4 6 (50%) 12 (100%) 21 (75%) 14 (-33.3%) 47.9% Anti-Bisexual 20 31 (55%) 19 (-38.7%) 37 (94.7%) 35 (-5.4%) 26.4% Anti-Black or African American 1949 2,182 (12.0%) 3,499 (60.4%) 3,297 (-5.8%) 3,424 (3.9%) 17.6% Anti-Buddhist 9 7 (-22.2%) 21 (200%) 30 (42.9%) 20 (-33.3%) 46.8% Anti-Catholic 51 68 (33.3%) 76 (11.8%) 99 (30.3%) 107 (8.1%) 20.9% Anti-Church of Jesus Christ 8 11 (37.5%) 8 (-27.3%) 16 (100%) 30 (87.5%) 49.4% Anti-Eastern Orthodox (Russian Greek Other) 31 54 (74.2%) 53 (-1.9%) 55 (3.8%) 78 (41.8%) 29.5% 218 Anti-Female 40 56 (40%) 56 (0%) 79 (41.1%) 77 (-2.5%) 19.6% Anti-Gay (Male) 744 765 (2.8%) 763 (-0.3%) 950 (24.5%) 1,077 (13.4%) 10.1% Anti-Gender Non-Conforming 26 42 (61.5%) 61 (45.2%) 103 (68.9%) 131 (27.2%) 50.7% Anti-Heterosexual 21 18 (-14.3%) 12 (-33.3%) 13 (8.3%) 22 (69.2%) 7.5% Anti-Hindu 13 6 (-53.9%) 10 (66.7%) 12 (20%) 25 (108.3%) 35.3% Anti-Hispanic or Latino 494 562 (13.8%) 610 (8.5%) 694 (13.8%) 738 (6.3%) 10.6% Anti-Islamic (Muslim) 192 185 (-3.6%) 129 (-30.3%) 153 (18.6%) 158 (3.3%) -3.0% Anti-Jehovah’s Witness 10 6 (-40%) 7 (16.7%) 6 (-14.3%) 14 (133.3%) 23.9% Anti-Jewish 851 1,053 (23.7%) 854 (-18.9%) 824 (-3.5%) 1,124 (36.4%) 9.4% Anti-Lesbian (Female) 131 115 (-12.2%) 128 (11.3%) 189 (47.7%) 191 (1.1%) 12.0% Anti-Lesbian Gay Bisexual or Transgender (Mixed Group) 306 304 (-0.7%) 341 (12.2%) 522 (53.1%) 622 (19.2%) 20.9% Anti-Male 22 16 (-27.3%) 21 (31.3%) 18 (-14.3%) 18 (0%) -2.6% Anti-Mental Disability 103 100 (-2.9%) 93 (-7%) 100 (7.5%) 97 (-3%) -1.3% Anti-Multiple Races Group 134 132 (-1.5%) 234 (77.3%) 196 (-16.2%) 232 (18.4%) 19.5% Anti-Multiple Religions Group 46 40 (-13.0%) 52 (30%) 37 (-28.9%) 42 (13.5%) 0.4% 219 Anti-Native Hawaiian or Other Pacific Islander 14 21 (50%) 20 (-4.8%) 50 (150%) 26 (-48%) 36.8% Anti-Other Christian 50 51 (2%) 58 (13.7%) 54 (-6.9%) 97 (79.6%) 22.1% Anti-Other Race/Ethnicity/Ancestry 268 317 (18.3%) 351 (10.7%) 338 (-3.7%) 399 (18.0%) 10.8% Anti-Other Religion 95 88 (-7.4%) 81 (-8.0%) 75 (-7.5%) 91 (21.3%) -0.3% Anti-Physical Disability 61 46 (-24.6%) 58 (26.1%) 54 (-6.9%) 74 (37.0%) 7.9% Anti-Protestant 34 24 (-29.4%) 38 (58.3%) 40 (5.3%) 63 (57.5%) 22.9% Anti-Sikh 41 53 (29.3%) 92 (73.6%) 191 (107.6%) 181 (-5.2%) 51.3% Anti-Transgender 143 159 (11.2%) 222 (39.6%) 250 (12.6%) 338 (35.2%) 24.7% Anti-White 781 679 (-13.1%) 1,083 (59.5%) 1,076 (-0.6%) 966 (-10.2%) 8.9% Multiple Bias 77 288 (274.0%) 368 (27.8%) 312 (-15.2%) 347 (11.2%) 74.5% Total 7,180 7,875 (9.7 %) 9,952 (26.4%) 10,891 (9.4%) 11,643 (6.9%) 13.1% Note: The year-on-year change, noted in the table as Δ, for each year after 2018 was calculated as the difference in number of hate crimes between the years divided by the number of hate crimes in the previous year. Hate crime rates could not be calculated given that many of these populations, primarily LGBTQ+ populations, are not enumerated by the U.S. decennial census nor by the American Community Survey (see Hubbell, 2024). 220 Table A2. Hate Crime Changes Over Time by Bias Category Type 2018-2022 2018 2019 2020 2021 2022 5-Year Average Change n n (Δ%) n (Δ%) n (Δ%) n (Δ%) % Bias Category Type Anti-Disability 164 146 (-11.0%) 151 (3.4%) 154 (2.0%) 171 (11.0%) 1.4% Anti-Gender 62 72 (16.1%) 77 (6.9%) 97 (26.0%) 95 (-2.1%) 11.7% Anti-Gender Identity 169 201 (18.9%) 283 (40.8%) 353 (24.7%) 469 (32.9%) 29.3% Anti-Race/Ethnicity/Ancestry 4,051 4,283 (5.7%) 6,319 (47.5%) 6,651 (5.3%) 6,570 (-1.2%) 14.3% Anti-Religion 1,427 1,641 (15.0%) 1,483 (-9.6%) 1,597 (7.7%) 2,014 (26.1%) 9.8% Anti-Sexual Orientation 1,222 1,233 (0.9%) 1,263 (2.4%) 1,711 (35.5%) 1,947 (13.8%) 13.1% Note: The year-on-year change, noted in the table as Δ, for each year after 2018 was calculated as the difference in number of hate crimes between the years divided by the number of hate crimes in the previous year. Hate crime rates could not be calculated given that many of these populations, primarily LGBTQ+ populations, are not enumerated by the U.S. decennial census nor by the American Community Survey (see Hubbell, 2024). 221 Table A3. Two-Class Solution Latent Class Marginal Means (1) (2) Latent Class Name Always interveners Only intervene for hate crime Class Membership Size (n=1,001) 691 (69.03%) 310 (30.97%) Average Posterior Probability 97.91% 97.39% Microaggression Did not see problem 0.1222 0.4911 Saw a problem 0.3071 0.4200 Felt morally obligated to assist 0.0273 0.0326 Considered options to intervene 0.0100 0.0289 Intervened 0.5334 0.0276 Hate Incident Did not see problem 0.0030 0.2723 Saw a problem 0.0081 0.5704 Felt morally obligated to assist 0.0436 0.0355 Considered options to intervene 0.0000 0.0885 Intervened 0.9453 0.0333 Hate Crime Did not see problem 0.0000 0.1548 Saw a problem 0.0097 0.2606 Felt morally obligated to assist 0.0247 0.0732 Considered options to intervene 0.0037 0.0457 Intervened 0.9619 0.4656 222 Table A4. Omnibus Chi-Square of Victim Identity and Latent Class for the Five-Class Solution Class 1 Class 2 Class 3 Class 4 Class 5 Class Description Always interveners Intervene as the situation escalates beyond micro-aggression Only intervene once physical violence occurs Never see a problem Progress through the model as expected but rarely intervene n (%) n (%) n (%) n (%) n (%) Victim Asian Man 77 (45.56%) 46 (27.22%) 24 (14.20%) 18 (10.65%) 4 (2.37%) Asian Woman 92 (58.23%) 44 (27.85%) 13 (8.23%) 5 (3.16%) 4 (2.53%) Gay Man 85 (48.85%) 38 (21.84%) 33 (18.97%) 14 (8.05%) 4 (2.30%) Lesbian 62 (37.13%) 44 (26.35%) 39 (23.35%) 17 (10.18%) 5 (2.99%) Trans Man 63 (44.68%) 36 (25.53%) 25 (17.73%) 14 (9.93%) 3 (2.13%) Trans Woman 75 (39.06%) 39 (20.31%) 46 (23.96%) 29 (15.10%) 3 (1.56%) χ2(20) 44.399 p-value .001 Cramér’s V .1053 223 Table A5. OLS Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) Microaggression Hate Incident Hate Crime Coefficient (SD) β Coefficient (SD) β Coefficient (SD) β Empathy .0072** (.0025) .0918 .0066*** (.0016) .1264 .0032* (.0013) .0801 Bystander Behavior .0276** (.0084) .1057 .0127* (.0056) .0732 .0070 (.0043) .0529 Bystander Efficacy .0004 (.0020) .0056 .0005 (.0013) .0122 .0003 (.0010) .0109 Dictator Game Average -.0129* (.0055) -.0746 -.0019 (.0037) -.0163 -.0031 (.0028) -.0349 Other-Regarding .0620* (.0287) .0693 .0113 (.0192) .0189 .0055*** (.01478) .0121 Feeling Toward Victim Group .2332*** (.0363) .2010 .1729*** (.0242) .2231 .1144*** (.0187) .1947 Majority Group Member -.0276 (.02926) -.0293 .0197 (.0195) .0314 .0001 (.0151) .0002 Constant .3321 (.0797) .5655*** (.0533) .7593*** (.0411) Model Fit F(7, 993) 14.06 13.85 8.78 p-value <.0001 <.0001 <.0001 R2 .0902 .0825 .0517 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 224 Table A6. OLS Regressions of Seeing a Problem for Each Phase on Bystander Characteristics (n=1,001) Microaggression Hate Incident Hate Crime Coefficient (SD) β Coefficient (SD) β Coefficient (SD) β Empathy -.0031 (.0027) -.0354 .0077** (.0026) .0889 .0100*** (.0022) .1377 Bystander Behavior .0515*** (.0094) .1742 .0566*** (.0087) .1949 .0245** (.0075) .1010 Bystander Efficacy .0121*** (.0022) .1692 .0125*** (.0021) .1774 .0081*** (.0018) .1384 Dictator Game Average -.0037 (.0061) -.0189 .0018 (.0057) .0993 -.0040 (.0049) -.0250 Other-Regarding .0362 (.0319) .0356 .0294 (.0297) .0295 .0570* (.0257) .0686 Feeling Toward Victim Group .2385*** (.0404) .1811 .3424*** (.0376) .2648 .2587*** (.0325) .2400 Majority Group Member -.0184 (.0326) -.0173 .0355 (.0303) .0339 .0403 (.0262) .0463 Constant -.2559** (.0887) -.3463*** (.0826) .0317 (.0713) Model Fit F(7, 993) 20.33 38.24 26.48 p-value <.0001 <.0001 <.0001 R2 .1253 .2123 .1573 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 225 Table A7. Logistic Regressions for Seeing a Problem for Each Hate Event Phase Using Main Effects Microaggression Hate Incident Hate Crime Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.0371* (0.016) 0.067** (0.025) 0.06 (0.032) Bystander Behavior 0.1544** (0.0499) 0.1655* (0.0766) 0.1593 (0.1008) Bystander Efficacy 0.0022 (0.012) 0.0102 (0.0168) 0.0085 (0.021) Dictator Game Average -0.0735 (0.0377) -0.0265 (0.0393) -0.0528 (0.0411) Other-Regarding 0.3712* (0.1826) 0.1761 (0.2859) 0.1014 (0.3716) Feeling Toward Victim Group 1.0953*** (0.1898) 1.4344*** (0.2502) 1.5916*** (0.3228) Hispanic 0.0299 (0.2168) -0.4654 (0.2954) -0.9821** (0.3622) White 0.1307 (0.1861) 0.4386 (0.2587) -0.2278 (0.3614) Cisgender -0.8372 (1.0773) 0.2472 (1.0986) --a Heterosexual -0.5468 (0.3129) -0.271 (0.4352) -0.759 (0.6614) Male -0.269 (0.1673) -0.5207* (0.2577) -0.4513 (0.3407) Constant 0.4 (1.1684) -0.7278 (1.2796) 1.3972 (1.099) Model Fit n 1,001 1,001 988a χ2(7) 92.87 84.54 54.88 p-value <.0001 <.0001 <.0001 Pseudo R2 .0844 .1408 .1408 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 aThis variable was collinear with the outcome and therefore excluded from the model along with 13 observations as all transgender and gender non-conforming respondents considered the hate crime phase to be a problem. The reference groups for the main effect variables were: non-Hispanic individuals for the Hispanic variable, non-whites for the White variable, transgender and gender non-conforming individuals for the cisgender variable, sexual minority individuals for the heterosexual variable, and individuals who were not male presenting (i.e., not cisgender or transgender men) for the male variable. 226 Table A8. Logistic Regressions for Intervening for Each Hate Event Phase Using Main Effects Microaggression Hate Incident Hate Crime Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy -0.0116 (0.0143) 0.0377* (0.0156) 0.0684*** (0.0187) Bystander Behavior 0.2534*** (0.0478) 0.2866*** (0.047) 0.173** (0.0555) Bystander Efficacy 0.0604*** (0.0122) 0.071*** (0.0124) 0.0568*** (0.0134) Dictator Game Average -0.0257 (0.0283) 0.0182 (0.0493) -0.0342 (0.0314) Other-Regarding 0.1679 (0.154) 0.1174 (0.1761) 0.4445* (0.2122) Feeling Toward Victim Group 1.5396*** (0.2616) 1.6549*** (0.2059) 1.364*** (0.2027) Hispanic 0.1925 (0.1907) 0.0407 (0.2114) -0.0716 (0.2398) White -0.0774 (0.1733) 0.3654* (0.182) 0.5295** (0.2006) Cisgender -0.8308 (0.7068) -1.0223 (0.8851) -0.8107 (1.1485) Heterosexual -0.7765** (0.2466) -0.3725 (0.2882) -0.8633* (0.3833) Male 0.1056 (0.1523) 0.0136 (0.1639) 0.0249 (0.1906) Constant -2.6878** (0.8544) -3.348** (1.0054) -2.0325 (1.2621) Model Fit n 1,001 1,001 1,001 χ2(7) 163.03 229.61 161.05 p-value <.0001 <.0001 <.0001 Pseudo R2 .1232 .1782 .1631 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 The reference groups for the main effect variables were: non-Hispanic individuals for the Hispanic variable, non-whites for the White variable, transgender and gender non-conforming individuals for the cisgender variable, sexual minority individuals for the heterosexual variable, and individuals who were not male presenting (i.e., not cisgender or transgender men) for the male variable. 227 Table A9. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics Using Main Effects Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0223 (0.0153) 1.0226 -0.0455* (0.0197) 0.9555 0.0678*** (0.0189) 1.0702 Bystander Behavior -0.1941*** (0.0511) 0.8236 -0.3817*** (0.0608) 0.6827 0.1876** (0.0564) 1.2064 Bystander Efficacy -0.0577*** (0.0131) 0.9439 -0.0824*** (0.0155) 0.9209 0.0247 (0.0138) 1.025 Dictator Game Average 0.0278 (0.0317) 1.0282 0.0434 (0.0369) 1.0444 -0.0156 (0.0353) 0.9845 Other-Regarding -0.0291 (0.1638) 0.9713 -0.3764 (0.2168) 0.6863 0.3473 (0.2091) 1.4152 Feeling Toward Victim Group -0.9105** (0.286) 0.4023 -2.2692*** (0.2818) 0.1034 1.3587*** (0.2195) 3.8911 Hispanic -0.2071 (0.2078) 0.8129 -0.2336 (0.2606) 0.7917 0.0265 (0.2592) 1.0269 White 0.3624 (0.1925) 1.4368 -0.2353 (0.2215) 0.7903 0.5977** (0.2184) 1.8179 Cisgender 0.3934 (0.7393) 1.482 0.2266 (0.9734) 1.2543 0.1669 (0.9969) 1.1816 Heterosexual 0.5801* (0.267) 1.7862 1.064** (0.3781) 2.8979 -0.4839 (0.3863) 0.6164 Male -0.1625 (0.1635) 0.85 0.0594 (0.2036) 1.0612 -0.2218 (0.1952) 0.8011 Constant 1.5352 (0.9137) 4.6423 4.8415 (1.1574) 126.6592 -3.3062** (1.1372) 0.0367 Model Fit n 1,001 χ2(22) 294.24 p-value <.0001 Pseudo R2 .1364 Note: Two tailed significance *p<.05, **p<.01, ***p<.001 The coefficients indicating the difference in logarithmic-odds were converted into relative risk ratios (RRR) in the likelihoods of class membership. From study 1 of this dissertation, class 1 contains the always interveners, class 2 contains those who intervene as the event escalates, and class 3 contains the individuals who never intervene. The reference groups for the main effect variables were: non-Hispanic individuals for the Hispanic variable, non-whites for the White variable, transgender and gender non-conforming individuals for the cisgender variable, sexual minority individuals for the heterosexual variable, and individuals who were not male presenting (i.e., not cisgender or transgender men) for the male variable. 228 Table A10. Logistic Regressions for Seeing a Problem for Microaggression by Victim Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.044* (2.4) 0.066 (0.0563) 0.0137 (0.0305) 0.0498 (0.0346) 0.0026 (0.0475) 0.11** (0.0409) Bystander Behavior 0.0867 (0.1348) 0.0761 (0.1899) 0.0086 (0.1137) 0.2064 (0.1081) 0.0517 (0.1465) 0.3002* (0.1283) Bystander Efficacy -0.0046 (0.0317) -0.0034 (0.0488) 0.0597* (0.0302) -0.0188 (0.0262) -0.0601 (0.0411) -0.0054 (0.026) Dictator Game Average -0.0455 (0.3098) -0.2709 (0.4028) -0.0787 (0.0716) -0.0139 (0.115) -0.0222 (0.1802) -0.0406 (0.0817) Other-Regarding -0.0531 (0.5814) 0.7186 (0.7901) 0.4634 (0.4133) 0.4373 (0.4159) 0.0995 (0.6059) 0.6215 (0.478) Feeling Toward Victim Group 2.0041** (0.6502) 1.2216 (0.988) 0.7329 (0.4896) 0.3994 (0.49) 1.1223* (0.4885) 1.7099*** (0.3974) Majority Group Member 0.4921 (0.4951) -0.1324 (0.6257) 0.287 (0.4122) -0.604 (0.3713) -0.9471 (0.5058) -0.0982 (0.4134) Constant -2.7647 (1.4617) -0.2052 (2.1356) -2.0398 (1.1299) -0.7316 (0.9233) 2.6928 (1.7217) -2.7451* (1.1383) Model Fit n 169 158 174 167 141 192 χ2(7) 20.14 5.42 12.10 18.30 11.89 49.19 p-value .0053 .6086 .0974 .0107 .1044 <.0001 Pseudo R2 .1207 .0547 .0561 .0830 .0863 .2278 229 Table A11. Logistic Regressions for Seeing a Problem for Hate Incident by Victim Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.1471* (0.0745) 0.1179 (0.1187) 0.0769 (0.0528) 0.0859 (0.0601) 0.1236 (0.0717) 0.0785 (0.0441) Bystander Behavior -0.0431 (0.21) 0.4663 (0.3493) -0.0627 (0.203) -0.079 (0.1961) 0.0665 (0.2193) 0.36* (0.1471) Bystander Efficacy 0.0201 (0.0454) 0.1057 (0.0848) 0.0218 (0.0537) 0.006 (0.0405) -0.0108 (0.055) -0.0289 (0.0294) Dictator Game Average 0.7439 (0.5221) -0.0271 (0.7932) -0.005 (0.0598) 0.8242 (0.6792) -0.2236 (0.1573) -0.0484 (0.0878) Other-Regarding -0.3783 (1.0124) -0.7268 (1.485) 0.2561 (0.76) -0.9551 (1.2206) 1.153 (1.0294) 0.1555 (0.5099) Feeling Toward Victim Group 3.304*** (0.8153) --a 1.6732* (0.6639) 0.9596 (0.6967) 1.2952* (0.6466) 1.015* (0.4536) Majority Group Member 1.2346 (0.8353) 1.0844 (1.3303) 0.6285 (0.7712) -0.3767 (0.6389) -0.5403 (0.7028) 0.2368 (0.4768) Constant -4.992* (2.2422) -2.905 (3.9094) -1.1135 (1.9858) -0.6339 (1.3569) -0.7611 (2.177) -0.3867 (1.2245) Model Fit n 169 151 174 167 141 192 χ2(7) 30.36 9.71 11.52 10.82 15.26 20.45 p-value .0001 .1374 .1174 .1466 .0328 .0047 Pseudo R2 .2998 .2211 .1183 .1125 .1760 .1282 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 230 Table A12. Logistic Regressions for Seeing a Problem with Hate Crime by Victim Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.1174 (0.0848) 0.0214 (0.2361) 0.1187 (0.0727) 0.025 (0.0829) 0.0387 (0.0918) 0.141* (0.0682) Bystander Behavior -0.1856 (0.2552) -1.0426 (1.4827) 0.1315 (0.2505) -0.0876 (0.2595) 0.0267 (0.3232) 0.4892* (0.221) Bystander Efficacy 0.0134 (0.0549) 1.8187 (2.1763) 0.0539 (0.068) 0.0313 (0.0514) -0.0422 (0.086) -0.0069 (0.0378) Dictator Game Average 1.0099 (0.6545) 3.3726 (3.6252) 0.2457 (0.501) -0.3015* (0.1443) -0.358* (0.1804) -0.1059 (0.0927) Other-Regarding -1.2644 (1.2291) --a -0.1989 (1.1602) 0.3491 (1.0065) 0.535 (1.2282) 0.2739 (0.7137) Feeling Toward Victim Group 3.5738*** (0.86) --a 1.2504 (0.8501) 1.5614 (0.9153) 2.8083** (0.9852) 0.5022 (0.6219) Majority Group Member 0.6814 (0.883) --a 0.5816 (0.9449) 0.6221 (0.9216) -2.1044* (0.9265) 0.4669 (0.6823) Constant -3.4785 (2.3854) -36.996 (45.1812) -2.5628 (2.6918) 0.7711 (1.8069) 3.0915 (2.7805) -1.2101 (1.682) Model Fit n 169 158 174 167 141 192 χ2(7) 24.54 6.39 7.63 7.15 19.99 15.25 p-value .0009 .1717 .3664 .4138 .0056 .0329 Pseudo R2 .3016 .5274 .1176 .1113 .3254 .1603 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 aThese variables were collinear with the outcome and had to be removed from the model for the Asian woman victim. 231 Table A13. Logistic Regressions for Intervening for Microaggression by Victim Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy -0.0086 (0.0322) -0.0023 (0.0324) -0.0675 (0.035) -0.0193 (0.0413) 0.0398 (0.0413) 0.0265 (0.0359) Bystander Behavior 0.0796 (0.1068) 0.0699 (0.1154) 0.2546* (0.1254) 0.2599 (0.1383) 0.2909* (0.1333) 0.5496*** (0.1324) Bystander Efficacy 0.0564* (0.0267) 0.0696* (0.03) 0.1424*** (0.037) 0.0939* (0.0381) 0.0667* (0.0336) 0.0351 (0.0277) Dictator Game Average 0.1193 (0.2499) 0.3399 (0.2905) -0.1427 (0.1616) 0.1411 (0.1278) 0.1424 (0.1579) 0.0201 (0.1134) Other-Regarding -0.2618 (0.4493) -0.0341 (0.5057) 0.272 (0.4307) 0.2395 (0.4428) 0.1935 (0.49) -0.5408 (0.419) Feeling Toward Victim Group 0.9405 (0.7139) -0.7137 (0.906) 1.8908** (0.7265) 1.6435 (1.0793) 2.2359** (0.6633) 2.171*** (0.5452) Majority Group Member 0.2939 (0.3681) -0.1223 (0.3854) 0.1221 (0.43) -0.3721 (0.4719) -0.8564 (0.5189) 0.5289 (0.4123) Constant -3.2203* (1.2829) -1.8361 (1.4282) -6.0003*** (1.4217) -6.0984*** (1.6357) -6.2212*** (1.7466) -5.34*** (1.2663) Model Fit n 169 158 174 167 141 192 χ2(7) 9.21 10.54 41.39 26.03 37.07 62.63 p-value .2382 .1600 <.0001 <.001 <.0001 <.0001 Pseudo R2 .0406 .0484 .1857 .1453 .1975 .2486 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 232 Table A14. Logistic Regressions for Intervening for Hate Incident by Victim Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.0641 (0.0378) 0.0198 (0.0434) 0.0463 (0.0363) 0.0378 (0.0353) 0.105* (0.0449) 0.0443 (0.0341) Bystander Behavior 0.2038 (0.1146) 0.0798 (0.1432) 0.3044* (0.1203) 0.1994 (0.1083) 0.4864*** (0.1376) 0.3898*** (0.1115) Bystander Efficacy 0.0772** (0.0297) 0.0682 (0.0376) 0.1175** (0.0367) 0.0586* (0.0291) 0.0343 (0.035) 0.0881** (0.0274) Dictator Game Average 0.199 (0.2906) -0.367 (0.3352) 0.0269 (0.0874) 0.2193 (0.1755) 0.0365 (0.1427) -0.0308 (0.0935) Other-Regarding 0.1354 (0.5272) 0.8377 (0.6135) -0.6292 (0.4506) 0.1492 (0.4435) 0.1435 (0.5477) 0.3112 (0.4068) Feeling Toward Victim Group 1.3986 (0.7326) 0.3118 (0.9467) 2.5106*** (0.5983) 0.792 (0.5181) 2.0267*** (0.5023) 1.4941*** (0.3957) Majority Group Member 1.1772* (0.4614) -0.1282 (0.4869) 0.5998 (0.4714) -0.2654 (0.3965) 0.0953 (0.4955) 0.1976 (0.3914) Constant -5.3896*** (1.4877) -1.3273 (1.6934) -6.6084*** (1.5056) -3.4816** (1.0958) -5.6323** (1.6976) -5.4594*** (1.1992) Model Fit n 169 158 174 167 141 192 χ2(7) 32.68 7.05 53.57 30.41 43.54 66.67 p-value <.0001 .4238 <.0001 <.001 <.0001 <.0001 Pseudo R2 .1555 .0468 .2434 .1362 .2359 .2515 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 233 Table A15. Logistic Regressions for Intervening for Hate Crime by Victim Asian Man Asian Woman Gay Man Lesbian Trans Man Trans Woman Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.1228* (0.0513) 0.0796 (0.0601) 0.0582 (0.0367) 0.0415 (0.0437) 0.1477** (0.0552) 0.1003* (0.0403) Bystander Behavior 0.0705 (0.1485) -0.2932 (0.2375) 0.2776* (0.1284) 0.0668 (0.1368) 0.2399 (0.156) 0.382** (0.1277) Bystander Efficacy 0.0963** (0.0354) -0.0077 (0.0526) 0.0467 (0.0348) 0.0409 (0.0309) 0.0031 (0.0406) 0.1065*** (0.0291) Dictator Game Average -0.301 (0.3363) -0.8308* (0.4193) 0.2079 (0.2285) -0.0675 (0.1339) -0.1773 (0.1541) -0.1223 (0.0827) Other-Regarding 0.7421 (0.645) 1.5063 (0.8102) -0.1657 (0.5557) 0.166 (0.5381) 0.8864 (0.7039) 1.2235* (0.5206) Feeling Toward Victim Group 2.7667*** (0.7591) 0.557 (1.0568) 1.4789** (0.5251) 1.3508* (0.5271) 1.2695* (0.507) 0.8353* (0.4055) Majority Group Member 0.4412 (0.5379) -0.0062 (0.6829) 0.4366 (0.4841) 0.1909 (0.4849) -0.3687 (0.5388) 0.983* (0.467) Constant -6.6645*** (1.7917) 1.8859 (2.3614) -3.5544* (1.4087) -1.8568 (1.1448) -3.1993 (1.784) -5.8375*** (1.2812) Model Fit n 169 158 174 167 141 192 χ2(7) 38.12 11.36 26.56 16.72 28.19 62.89 p-value <.0001 .1236 <.001 .0193 <.001 <.0001 Pseudo R2 .2367 .1201 .1434 .1043 .2006 .2804 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 234 Table A16. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Asian Man Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0289 (0.0353) 1.0293 -0.1025* (0.0519) 0.9026 0.1314* (0.0522) 1.1404 Bystander Behavior -0.0785 (0.1166) 0.9245 -0.1206 (0.1526) 0.8864 0.0421 (0.1516) 1.043 Bystander Efficacy -0.0643* (0.03) 0.9377 -0.0984** (0.0373) 0.9063 0.0341 (0.0355) 1.0347 Dictator Game Average -0.1026 (0.282) 0.9025 -0.1615 (0.3544) 0.8509 0.0589 (0.3716) 1.0607 Other-Regarding 0.4424 (0.4955) 1.5564 0.0529 (0.6677) 1.0543 0.3895 (0.6884) 1.4762 Feeling Toward Victim Group 0.7755 (1.1916) 2.1717 -2.3469** (0.8246) 0.0957 3.1224** (1.1287) 22.7008 Majority Group Member -0.183 (0.4054) 0.8328 -0.5026 (0.5417) 0.605 0.3197 (0.5434) 1.3767 Constant 0.8311 (1.733) 2.2958 7.215*** (1.9082) 1359.674 -6.3839** (1.9472) 0.0017 Model Fit n 169 χ2(14) 41.43 p-value <.001 Pseudo R2 .1166 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 235 Table A17. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Asian Woman Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.016 (0.0348) 1.0161 -0.0507 (0.0622) 0.9506 0.0666 (0.0646) 1.0689 Bystander Behavior -0.0149 (0.1253) 0.9852 -0.2506 (0.1952) 0.7783 0.2358 (0.2028) 1.2659 Bystander Efficacy -0.0791* (0.0327) 0.9239 -0.0795 (0.0515) 0.9236 0.0005 (0.0536) 1.0005 Dictator Game Average -0.4363 (0.3402) 0.6464 0.4483 (0.4172) 1.5656 -0.8846 (0.4685) 0.4129 Other-Regarding 0.1029 (0.5788) 1.1084 -0.8471 (0.7833) 0.4287 0.95 (0.8625) 2.5857 Feeling Toward Victim Group 0.9395 (1.1771) 2.5587 -0.9352 (1.0353) 0.3925 1.8747 (1.359) 6.5189 Majority Group Member 0.408 (0.4078) 1.5038 -0.4868 (0.7497) 0.6146 0.8948 (0.7693) 2.4468 Constant 1.2444 (1.6692) 3.4709 3.0934 (2.2527) 22.0519 -1.849 (2.428) 0.1574 Model Fit n 158 χ2(14) 22.50 p-value .0690 Pseudo R2 .0770 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 236 Table A18. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Gay Man Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0573 (0.0353) 1.059 0.0332 (0.0439) 1.0338 0.0241 (0.0403) 1.0244 Bystander Behavior -0.1559 (0.1288) 0.8556 -0.3221* (0.1506) 0.7246 0.1661 (0.1365) 1.1807 Bystander Efficacy -0.116** (0.0372) 0.8905 -0.1496** (0.0441) 0.8611 0.0336 (0.0379) 1.0342 Dictator Game Average 0.1443 (0.1615) 1.1552 0.1071 (0.1721) 1.113 0.0372 (0.0743) 1.0379 Other-Regarding -0.5167 (0.4489) 0.5965 -0.0765 (0.5583) 0.9264 -0.4402 (0.5124) 0.6439 Feeling Toward Victim Group -0.831 (0.8193) 0.4356 -3.1066*** (0.8001) 0.0448 2.2756*** (0.6348) 9.7338 Majority Group Member 0.2214 (0.4402) 1.2478 -0.8871 (0.5933) 0.4118 1.1085* (0.5499) 3.0298 Constant 3.611* (1.4893) 37.003 7.0619*** (1.7573) 1166.66 -3.4509 *(1.4745) 0.0317 Model Fit n 174 χ2(14) 60.36 p-value <.0001 Pseudo R2 .1606 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 237 Table A19. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Lesbian Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0049 (0.0422) 1.0049 -0.0652 (0.0486) 0.9369 0.07 (0.0399) 1.0725 Bystander Behavior -0.1832 (0.1372) 0.8326 -0.3085* (0.1542) 0.7345 0.1253 (0.1232) 1.1335 Bystander Efficacy -0.09* (0.0379) 0.9139 -0.0955* (0.0419) 0.9089 0.0056 (0.0297) 1.0056 Dictator Game Average -0.1031 (0.1305) 0.902 -0.1922 (0.1789) 0.8251 0.0891 (0.1706) 1.0932 Other-Regarding 0.0554 (0.45) 1.057 -0.4434 (0.5743) 0.6419 0.4988 (0.508) 1.6467 Feeling Toward Victim Group -1.2294 (1.1172) 0.2925 -2.2549* (1.0991) 0.1049 1.0255 (0.5332) 2.7885 Majority Group Member 0.2731 (0.4772) 1.314 0.4518 (0.5454) 1.5711 -0.1787 (0.4365) 0.8364 Constant 5.0034** (1.6885) 148.9186 7.585* (1.8175) 1968.447 -2.5815* (1.1295) 0.0757 Model Fit n 167 χ2(14) 46.96 p-value <.0001 Pseudo R2 .1314 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 238 Table A20. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Trans Man Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0351 (0.0448) 1.0357 -0.1313* (0.0569) 0.877 0.1664** (0.0561) 1.181 Bystander Behavior -0.1953 (0.1433) 0.8226 -0.5352** (0.1702) 0.5856 0.3399* (0.1602) 1.4048 Bystander Efficacy -0.081* (0.0362) 0.9222 -0.0511 (0.0442) 0.9502 -0.0299 (0.0427) 0.9705 Dictator Game Average -0.5564 (0.3533) 0.5733 -0.0513 (0.1512) 0.95 -0.5051 (0.3574) 0.6034 Other-Regarding 0.4572 (0.6561) 1.5796 -0.2135 (0.671) 0.8078 0.6708 (0.8028) 1.9558 Feeling Toward Victim Group -1.2239* (0.6105) 0.2941 -2.4049*** (0.6448) 0.0903 1.181* (0.5499) 3.2576 Majority Group Member 0.8591 (0.5384) 2.361 0.6679 (0.6139) 1.9501 0.1913 (0.5611) 1.2108 Constant 3.4966 (1.795) 33.0031 7.069** (2.1091) 1174.973 -3.5724 (1.9025) 0.0281 Model Fit n 141 χ2(14) 56.00 p-value <.0001 Pseudo R2 .1840 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 239 Table A21. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for Trans Woman Victim Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0239 (0.0397) 1.0242 -0.0866 (0.0451) 0.917 0.1105** (0.0405) 1.1168 Bystander Behavior -0.5195*** (0.1479) 0.5948 -0.7255*** (0.1564) 0.4841 0.206 (0.1307) 1.2288 Bystander Efficacy -0.0067 (0.0309) 0.9933 -0.0649 (0.0338) 0.9372 0.0582* (0.0291) 1.0599 Dictator Game Average -0.4118 (0.3306) 0.6625 0.049 (0.101) 1.0502 -0.4609 (0.3301) 0.6307 Other-Regarding 1.5638* (0.6256) 4.7769 0.099 (0.5285) 1.1041 1.4648* (0.6593) 4.3267 Feeling Toward Victim Group -2.0283** (0.5913) 0.1316 -2.5025*** (0.5922) 0.0819 0.4742 (0.4226) 1.6067 Majority Group Member -0.2741 (0.4549) 0.7603 -0.1113 (0.4891) 0.8947 -0.1627 (0.441) 0.8498 Constant 2.5163 (1.4208) 12.3827 7.1504*** (1.5309) 1274.616 -4.634*** (1.2546) 0.0097 Model Fit n 192 χ2(14) 100.50 p-value <.0001 Pseudo R2 .2392 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 240 Table A22. Logistic Regressions for Seeing a Problem for Each Hate Event Phase by 2020 Election Results Microaggression Hate Incident Hate Crime Trump State Biden State Trump State Biden State Trump State Biden State Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.0467 (0.0244) 0.0412* (0.0192) 0.1076* (0.0429) 0.0839** (0.028) 0.1113* (0.0558) 0.0612 (0.0369) Bystander Behavior 0.1352 (0.0744) 0.1804** (0.0663) 0.2695* (0.1226) 0.1319* (0.0962) 0.1695 (0.1577) 0.1833 (0.1279) Bystander Efficacy 0.013 (0.0188) -0.0083 (0.0154) 0.0195 (0.0287) -0.0054 (0.0209) -0.0032 (0.0368) 0.0003 (0.0259) Dictator Game Average -0.0302 (0.0816) -0.0792 (0.0508) -0.1836 (0.0971) 0.0491 (0.1115) -0.0194 (0.1489) -0.0368 (0.037) Other-Regarding -0.046 (0.286) 0.6884** (0.2482) 0.3648 (0.5057) 0.1169 (0.3745) 0.6303 (0.7152) -0.2034 (0.4444) Feeling Toward Victim Group 1.2863*** (0.2787) 0.9911*** (0.2638) 1.7593*** (0.4138) 1.3484*** (0.3263) 1.3759** (0.5274) 1.7618*** (0.4077) Majority Group Member -0.1225 (0.2671) -0.1621 (0.225) 0.3279 (0.445) 0.1583 (0.3285) 0.4025 (0.5862) -0.2298 (0.406) Constant -1.444 (0.7474) -0.7611 (0.6043) -2.0212 (1.1918) -0.8163 (0.8186) -0.6129 (1.5104) 0.1137 (1.0066) Model Fit n 424 577 424 577 424 577 χ2(7) 39.46 51.13 47.44 36.22 19.58 30.43 p-value <.0001 <.0001 <.0001 <.0001 .0065 <.0001 Pseudo R2 .0848 .0805 .2004 .0997 .1314 .1260 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 241 Table A23. Logistic Regressions for Intervening for Each Hate Event Phase by 2020 Election Results Microaggression Hate Incident Hate Crime Trump State Biden State Trump State Biden State Trump State Biden State Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy -0.0078 (0.0216) -0.0226 (0.0173) 0.0374 (0.0246) 0.0522** (0.0184) 0.0926** (0.0306) 0.0778*** (0.0217) Bystander Behavior 0.2102** (0.0704) 0.2858*** (0.0635) 0.2779*** (0.073) 0.295*** (0.0612) 0.1946* (0.0889) 0.171* (0.0707) Bystander Efficacy 0.0745*** (0.0191) 0.0606*** (0.0155) 0.0773*** (0.0196) 0.0648*** (0.0156) 0.0348 (0.0216) 0.0648*** (0.017) Dictator Game Average 0.0347 (0.0784) -0.0278 (0.0333) 0.0823 (0.0958) 0.0049 (0.0437) 0.0606 (0.1116) -0.0324 (0.0306) Other-Regarding 0.2293 (0.2435) 0.0833 (0.2019) 0.0214 (0.2873) 0.1527 (0.2208) 0.5741 (0.3868) 0.2926 (0.261) Feeling Toward Victim Group 1.6008*** (0.3701) 1.4264*** (0.3659) 1.9905*** (0.304) 1.3774*** (0.2841) 1.5043*** (0.3083) 1.3081*** (0.2759) Majority Group Member -0.036 (0.2424) -0.1587 (0.2158) 0.2272 (0.2792) 0.155 (0.2194) 0.6777 (0.3495) 0.1214 (0.2487) Constant -4.669*** (0.7869) -3.901*** (0.6857) -4.8623 (0.8425) -4.4384*** (0.6692) -3.3546*** (0.9411) -3.7985*** (0.7053) Model Fit n 424 577 424 577 424 577 χ2(7) 70.17 77.29 114.98 112.75 69.39 88.04 p-value <.0001 <.0001 <.0001 <.0001 <.0001 <.0001 Pseudo R2 .1238 .1023 .2137 .1504 .1811 .1466 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 242 Table A24. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for 2020 Trump States Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.034 (0.0227) 1.0346 -0.0778* (0.0327) 0.9251 0.1118*** (0.0311) 1.1183 Bystander Behavior -0.1789* (0.0742) 0.8362 -0.3653*** (0.096) 0.694 0.1864* (0.0905) 1.2049 Bystander Efficacy -0.0726*** (0.0198) 0.93 -0.075** (0.0251) 0.9277 0.0024 (0.0226) 1.0024 Dictator Game Average -0.0322 (0.0893) 0.9683 0.0064 (0.1043) 1.0064 -0.0385 (0.1064) 0.9622 Other-Regarding -0.0754 (0.2578) 0.9274 -0.4007 (0.3753) 0.6699 0.3253 (0.3694) 1.3844 Feeling Toward Victim Group -0.8951* (0.3948) 0.4086 -2.5746*** (0.4047) 0.0762 1.6795*** (0.3274) 5.3629 Majority Group Member 0.1973 (0.2549) 1.2181 -0.2116 (0.3535) 0.8093 0.4088 (0.3376) 1.505 Constant 2.7708** (0.8265) 15.9714 6.3013*** (1.0902) 545.2803 -3.5305*** (0.9628) 0.0293 Model Fit n 424 χ2(14) 131.80 p-value <.0001 Pseudo R2 .1469 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 243 Table A25. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for 2020 Biden States Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0374* (0.019) 1.0381 -0.0394 (0.0227) 0.9614 0.0768** (0.0222) 1.0798 Bystander Behavior -0.2048** (0.0689) 0.8148 -0.4064*** (0.0782) 0.666 0.2016** (0.0725) 1.2234 Bystander Efficacy -0.0609*** (0.017) 0.9409 -0.09*** (0.0195) 0.9139 0.0291 (0.0175) 1.0295 Dictator Game Average 0.0307 (0.0365) 1.0312 0.0353 (0.0433) 1.0359 -0.0046 (0.0369) 0.9954 Other-Regarding 0.0636 (0.2166) 1.0657 -0.274 (0.2738) 0.7603 0.3377 (0.2616) 1.4017 Feeling Toward Victim Group -0.8223* (0.4114) 0.4394 -2.0653*** (0.3944) 0.1268 1.243*** (0.3062) 3.466 Majority Group Member 0.2769 (0.2349) 1.319 0.1753 (0.2712) 1.1916 0.1016 (0.2519) 1.1069 Constant 2.2289** (0.7721) 9.2896 5.9678*** (0.8573) 390.6453 -3.739*** (0.726) 0.0238 Model Fit n 577 χ2(14) 147.60 p-value <.0001 Pseudo R2 .1176 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 244 Table A26. Logistic Regressions for Seeing a Problem for Each Hate Event Phase by White or non-White Respondents Microaggression Hate Incident Hate Crime White non-White White non-White White non-White Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy 0.0247 (0.0177) 0.1019** (0.0382) 0.0699* (0.028) 0.14* (0.0551) 0.0691* (0.0335) 0.1591 (0.0894) Bystander Behavior 0.175** (0.0568) 0.1378 (0.1032) 0.1762 (0.0921) 0.2238 (0.1307) 0.1875 (0.112) 0.1355 (0.2039) Bystander Efficacy 0.009 (0.0138) -0.0173 (0.0237) -0.0049 (0.0207) 0.0212 (0.0281) -0.0107 (0.0248) 0.0354 (0.0405) Dictator Game Average -0.0696 (0.0531) -0.0676 (0.0549) -0.0766 (0.0654) 0.0198 (0.1147) -0.0907 (0.0707) -0.0284 (0.0601) Other-Regarding 0.2964 (0.2065) 0.6989 (0.4089) 0.0747 (0.3388) 0.4114 (0.5606) 0.1164 (0.4131) 0.3011 (0.8459) Feeling Toward Victim Group 0.9865*** (0.2249) 1.3368*** (0.3608) 1.7119*** (0.3105) 0.9936*** (0.426) 1.7036*** (0.3739) 1.338* (0.6194) Majority Group Member -0.2758 (0.1937) (omitted) 0.0236 (0.3102) (omitted) 0.0966 (0.3766) (omitted) Constant -0.7003 (0.5618) -1.8695 (0.9578) -0.3831 (0.8375) -2.7512* (1.255) 0.2745 (0.9961) -2.3118 (1.8428) Model Fit n 746 255 746 255 746 255 χ2(7) 53.04 38.73 50.66 26.65 34.69 15.18 p-value <.0001 <.0001 <.0001 <.001 <.0001 .0189 Pseudo R2 .0658 .1319 .1274 .1356 .1202 .1478 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 245 Table A27. Logistic Regressions for Intervening for Each Hate Event Phase by White or non-White Respondents Microaggression Hate Incident Hate Crime White non-White White non-White White non-White Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Coefficient (SD) Empathy -0.0144 (0.0162) -0.0002 (0.0287) 0.0251 (0.0179) 0.0973** (0.0335) 0.0476* (0.0209) 0.1604*** (0.043) Bystander Behavior 0.2172*** (0.0544) 0.3902*** (0.0988) 0.2591*** (0.0546) 0.4247*** (0.0951) 0.1756** (0.0657) 0.2483* (0.1051) Bystander Efficacy 0.0857*** (0.0144) 0.0114 (0.0221) 0.0878*** (0.0148) 0.028 (0.0225) 0.0463** (0.0158) 0.0768** (0.025) Dictator Game Average 0.0415 (0.0588) -0.0516 (0.0527) 0.0005 (0.0641) 0.0746 (0.159) -0.0765 (0.0577) -0.0116 (0.0498) Other-Regarding -0.0181 (0.1823) 0.7554* (0.3281) 0.1192 (0.2042) 0.1049 (0.3998) 0.4105 (0.2483) 0.7068 (0.439) Feeling Toward Victim Group 1.5623*** (0.3203) 1.4145** (0.4563) 1.8538*** (0.2481) 1.0428** (0.3787) 1.5601*** (0.239) 0.774* (0.3809) Majority Group Member -0.0426 (0.1791) (omitted) 0.003 (0.195) (omitted) 0.0681 (0.2297) (omitted) Constant -4.897*** (0.6414) -3.0939** (0.9008) -4.6391*** (0.6437) -4.4023*** (0.9579) -2.4913*** (0.6665) -5.7209*** (1.1332) Model Fit n 746 255 746 255 746 255 χ2(7) 111.31 45.37 171.71 62.01 96.77 59.63 p-value <.0001 <.0001 <.0001 <.0001 <.0001 <.0001 Pseudo R2 .1144 .1308 .1821 .1815 .1415 .2030 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 246 Table A28. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for White Respondents Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0259 (0.0172) 1.0262 -0.0213 (0.0223) 0.979 0.0471* (0.0211) 1.0483 Bystander Behavior -0.1599** (0.0574) 0.8522 -0.3524*** (0.0713) 0.703 0.1925** (0.065) 1.2123 Bystander Efficacy -0.0811*** (0.0151) 0.9221 -0.1033*** (0.0186) 0.9018 0.0222 (0.0161) 1.0224 Dictator Game Average -0.0898 (0.0789) 0.9141 0.046 (0.0644) 1.047 -0.1358 (0.0822) 0.873 Other-Regarding 0.1469 (0.2005) 1.1582 -0.3146 (0.2541) 0.7301 0.4615 (0.2522) 1.5865 Feeling Toward Victim Group -0.9488** (0.3496) 0.3872 -2.4887*** (0.3484) 0.083 1.5399*** (0.2531) 4.6639 Majority Group Member 0.0844 (0.1898) 1.088 0.2115 (0.2432) 1.2356 -0.1272 (0.2281) 0.8806 Constant 3.3886*** (0.6825) 29.6238 5.95*** (0.8165) 383.7568 -2.5614*** (0.6762) 0.0772 Model Fit n 746 χ2(14) 198.14 p-value <.0001 Pseudo R2 .1245 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 247 Table A29. Multinomial Logistic Regression of Latent Class Membership on Bystander Characteristics for non-White Respondents Class 2 vs. Class 1 Class 3 vs. Class 1 Class 2 vs. Class 3 Coefficient (SD) RRR Coefficient (SD) RRR Coefficient (SD) RRR Empathy 0.0365 (0.0322) 1.0371 -0.1524** (0.0452) 0.8586 0.1889*** (0.0465) 1.2079 Bystander Behavior -0.301 (0.1128) 0.7401 -0.5404*** (0.1182) 0.5825 0.2394* (0.1194) 1.2705 Bystander Efficacy -0.0086 (0.0261) 0.9914 -0.0421 (0.0271) 0.9588 0.0335 (0.0277) 1.034 Dictator Game Average 0.0624 (0.0605) 1.0643 0.0104 (0.12) 1.0104 0.052 (0.1105) 1.0533 Other-Regarding -0.3839 (0.3697) 0.6812 -0.5925 (0.4606) 0.553 0.2085 (0.4794) 1.2319 Feeling Toward Victim Group -0.8741 (0.5112) 0.4172 -1.5745** (0.481) 0.2071 0.7003 (0.4379) 2.0144 Constant 0.6753 (1.0771) 1.9646 6.6184*** (1.2375) 748.7342 -5.9431*** (1.2325) 0.0026 Model Fit n 255 χ2(14) 86.86 p-value <.0001 Pseudo R2 .1586 Note: Two-tailed significance tests: *p<.05, **p<.01, ***p<.001 248 Table A30. Stratifications by Randomly Assigned Victim Identity, 2020 Election State Residency, and White versus non-White Respondents Microaggression Hate Incident Hate Crime Latent Class Membership See a Problem Intervene See a Problem Intervene See a Problem Intervene Class 2 vs Class 1 Class 3 vs Class 1 Class 2 vs Class 3 Empathy AM, TW, BSR, NW AM, TSR, BSR, W, NW TM, BSR, NW TW, TSR, W AM, TM, TW, TSR, BSR, W, NW BSR AM, TM, TSR, NW AM, TM, TW, TSR, BSR, W, NW Bystander Behavior TW, BSR, W GM, TM, TW, TSR, BSR, W, NW TW, TSR, BSR GM, TM, TW, TSR, BSR, W, NW TW GM, TW, TSR, BSR, W, NW TW, TSR, BSR, W GM, L, TM, TW, TSR, BSR, W, NW TM, TSR, BSR, W, NW Bystander Efficacy GM AM, AW, GM, L, TM, TSR, BSR, W AM, GM, L, TW, TSR, BSR, W AM, TW, BSR, W, NW AM, AW, GM, L, TM, TSR, BSR, W AM, GM, L, TSR, BSR, W TW Dictation Game Average L, TM Other-Regarding BSR GM, TM, TW, NW TW TW TW Feeling Toward Victim Group AM, TM, TW, TSR, BSR, W, NW GM, TM, TW, TSR, BSR, W, NW AM, GM, TM, TW, TSR, BSR, W, NW GM, TM, TW, TSR, BSR, W, NW AM, TM, TSR, BSR AM, GM, L, TM, TW, TSR, BSR, W, NW TM, TW, TSR, BSR, W AM, GM, L, TM, TW, TSR, BSR, W, NW AM, GM, TM, TSR, BSR, W Majority Group Member AM TM TW GM 249 Note: Two-tailed significance tests were used for the above. Initials are given for any stratification with a p-value below 0.05. AM=Asian Man victim, AW=Asian Woman victim, GM=Gay Man victim, L=Lesbian victim, TM=Trans Man victim, TW=Trans Woman victim, TSR=Trump State Resident 2020 election, BSR=Biden State Resident 2020 election, W=White respondent, NW=non-White respondent 250 Figure A2. Year-on-Year Percent Changes in Hate Crime by Bias Category Type 2018-2022 251 Vignettes Microaggression Scenario Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young Asian man trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and asks "Aren't you people supposed to be good at math?" The Asian man looks uncomfortable but says nothing. Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young Asian woman trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind her and asks "Aren't you people supposed to be good at math?" The Asian woman looks uncomfortable but says nothing. Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young man who appears to be gay wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and says, "Oh, someone is very out loud and proud!" The gay man looks uncomfortable but says nothing. Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young woman who appears to be lesbian wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks 252 behind her and says, "Oh, someone is very out loud and proud!" The lesbian woman looks uncomfortable but says nothing. Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young person who appears to be a transgender man wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and says, "You know you're still a woman, right?" The transgender man looks uncomfortable but says nothing. Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young person who appears to be a transgender woman wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind her and says, "You know you're still a man, right?" The transgender woman looks uncomfortable but says nothing. Hate Incident Scenario The young white man continues as the other customer appears to ignore him. He now says to the Asian man, "Get out of here, you stupid chink! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." The young white man continues as the other customer appears to ignore him. He now says to the Asian woman, "Get out of here, you stupid chink! No one wants you here!" 253 The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." The young white man continues as the other customer appears to ignore him. He now says to the gay man, "Get out of here, you stupid faggot! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." The young white man continues as the other customer appears to ignore him. He now says to the lesbian woman, "Get out of here, you stupid dyke! No one wants you here!" The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." The young white man continues as the other customer appears to ignore him. He now says to the transgender man, "Get out of here, you stupid tranny! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." The young white man continues as the other customer appears to ignore him. He now says to the transgender woman, "Get out of here, you stupid tranny! No one wants you here!" The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." 254 Hate Crime Scenario The young white man appears to grow angrier that the Asian man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" The young white man appears to grow angrier that the Asian woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" The young white man appears to grow angrier that the gay man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" The young white man appears to grow angrier that the lesbian woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" The young white man appears to grow angrier that the transgender man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" The young white man appears to grow angrier that the transgender woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" 255 Survey Instrument The survey instrument used in Qualtrics is shown below. Items retained after confirmatory factor analysis for the bystander efficacy, bystander behavior, and empathy scales are highlighted. Start of Block: Informed Consent Information Q1.1 INFORMED CONSENT INFORMATION FOR RESEARCH PARTICIPATION Study Title: Bystander Intervention in Incidents Principal Investigator: Theodore Wilson, Assistant Professor Co-Principal Investigator: James Hubbell, PhD Candidate IRB Study Number: 23X079 I am a faculty member at the University at Albany, in the School of Criminal Justice. I am planning to conduct a research study, which I invite you to take part in. This form has important information about the reasons for doing this study, what we will ask you to do, and the way we would like to use any information about you that we collect. Why are you doing this study? You are being asked to participate in a research study about how you would respond to possible bias situations that might occur in public. The purpose of the study is to see how characteristics of the people involved in these incidents influence your perceptions of the events and any actions you may or may not take. Why am I eligible to participate in this study? You are eligible to participate because you are over 18 years of age and a resident of the U.S. What will I do if I choose to be in this study? You will read scenarios of possible bias incidents that you might observe in public. You will be asked questions about how you would interpret these possible bias incidents and what you may or may not do if you witnessed them. Lastly, you will be asked some demographic questions as well as some questions regarding your decision-making rationale. You have the right not to answer any question, and to stop participation at any time. For how long will I participate? Study participation will take approximately 10-15 minutes. 256 Where will I need to go to participate? All study procedures will take place online through the Qualtrics survey platform. Are there any costs I should be aware of? Participation is not expected to incur cost to individual participants. What are the possible risks or discomforts? As with all research, there is a chance that confidentiality of the information we collect from you could be breached – we will take steps to minimize this risk, as discussed in more detail below in this form. You may feel uncomfortable being asked to imagine witnessing some of these scenarios. However, the risks are likely to be minimal because the entire survey documents hypothetical cases in plain text (as opposed to showing you traumatic or graphic pictures or videos). What are the possible benefits for me or others? You are not likely to have any direct benefit from being in this research study. This study is designed to learn more about bystander behavior during possible bias incidents. The study results may be used to help other people in the future. How will you protect the information you collect about me, and how will that information be shared? Results of this study may be used in publications and presentations. Your study data will be handled as confidentially as possible. If results of this study are published or presented, individual names and other personally identifiable information will not be used. To minimize the risks to confidentiality, your responses will be anonymous. We will not collect any information that can be used to identify you, and your responses will only be accessed from a secure server by members of the research team. De-identified data collected as a part of the current will be shared with other investigators for future research purposes. The results of this study may be used in reports, presentations, or publications but your name will not be used. All results will only be shared in the aggregate form. Will my data be used in future research? Your responses to the survey questions may be used in future research. However, all your data will be anonymous. What are my rights as a research participant? Participation in this study is voluntary. You do not have to answer any question you do not want to answer. If at any time and for any reason, you would prefer not to participate in this study, please feel free not to. If at any time you would like to stop participating, please tell me. We can take a break, stop and continue at a later date, or stop altogether. You may withdraw from this study at any time, and you will not be penalized in any way for deciding to stop participation. If you decide to withdraw from this study, the researchers will ask you if the information already 257 collected from you can be used. What if I am a University at Albany student or employee? You may choose not to participate or to stop participating in this research at any time. This will not affect your class standing, grades, employment, or any other aspects of your relationship with the University at Albany. Who can I contact if I have questions or concerns about this research study? If you have questions, you are free to ask them now. If you have questions later, you may contact the primary investigator, Dr. Theodore Wilson, at thwilson@albany.edu or 518-442-5176 or the co-investigator, James Hubbell, at jhubbell@albany.edu or 518-442-5210. If you have any questions about your rights as a participant in this research, you can contact the following office at the University at Albany: Institutional Review Board University at Albany Office of Regulatory and Research Compliance 1400 Washington Ave, ES 244 Albany, NY 12222 Phone: 1-866-857-5459 Email: rco@albany.edu Consent I have read this form and the research study has been explained to me. I have been given the opportunity to ask questions and my questions have been answered. If I have additional questions, I have been told whom to contact. I agree to participate in the research study described above. I understand that I will receive a copy of this information. Q1.2 The information in the consent document and any other written information was accurately explained to and understood by me, and I give consent freely. By clicking 'next', I agree to participate in the study. o I do not consent to take the survey (1) o Next (2) Skip To: End of Survey If Q1.2 = I do not consent to take the survey End of Block: Informed Consent Information 258 Start of Block: Asian Man Scenario Q2.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young Asian man trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and asks "Aren't you people supposed to be good at math?" The Asian man looks uncomfortable but says nothing. Q2.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q2.9 If Q2.2 = No Q2.3 Do you feel obligated to help the Asian man in this situation? o Yes (1) o No (2) Skip To: Q2.9 If Q2.3 = No Q2.4 Is there anything you could do to help the Asian man in this situation? o Yes (1) o No (2) Skip To: Q2.9 If Q2.4 = No 259 Q2.5 Please list what you think you could do in this event to help the Asian man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q2.6 Would you do anything in this situation to help the Asian man? o Yes (1) o No (2) Skip To: Q2.9 If Q2.6 = No Q2.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q2.5/ChoiceTextEntryValue/1} (1) o ${Q2.5/ChoiceTextEntryValue/2} (2) o ${Q2.5/ChoiceTextEntryValue/3} (3) o ${Q2.5/ChoiceTextEntryValue/4} (4) Page Break 260 Q2.8 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q2.9 The young white man continues as the other customer appears to ignore him. He now says to the Asian man, "Get out of here, you stupid chink! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q2.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q2.17 If Q2.10 = No Q2.11 Do you feel obligated to help the Asian man in this situation? o Yes (1) o No (2) Skip To: Q2.17 If Q2.11 = No Q2.12 Is there anything you could do to help the Asian man in this situation? o Yes (1) o No (2) 261 Skip To: Q2.17 If Q2.12 = No Q2.13 Please list what you think you could do in this event to help the Asian man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q2.14 Would you do anything in this situation to help the Asian man? o Yes (1) o No (2) Skip To: Q2.17 If Q2.14 = No Q2.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q2.13/ChoiceTextEntryValue/1} (1) o ${Q2.13/ChoiceTextEntryValue/2} (2) o ${Q2.13/ChoiceTextEntryValue/3} (3) o ${Q2.13/ChoiceTextEntryValue/4} (4) Page Break 262 Q2.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q2.17 The young white man appears to grow angrier that the Asian man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q2.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q2.18 = No Q2.19 Do you feel obligated to help the Asian man in this situation? o Yes (1) o No (2) Skip To: End of Block If Q2.19 = No Q2.20 Is there anything you could do to help the Asian man in this situation? o Yes (1) o No (2) 263 Skip To: End of Block If Q2.20 = No Q2.21 Please list what you think you could do in this event to help the Asian man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q2.22 Would you do anything in this situation to help the Asian man? o Yes (1) o No (2) Skip To: End of Block If Q2.22 = No Q2.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q2.21/ChoiceTextEntryValue/1} (1) o ${Q2.21/ChoiceTextEntryValue/2} (2) o ${Q2.21/ChoiceTextEntryValue/3} (3) o ${Q2.21/ChoiceTextEntryValue/4} (4) Page Break 264 Q2.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q2.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q2.25 != No, I would not contact the police. Q2.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Asian Man Scenario Start of Block: Asian Woman Scenario Q3.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young Asian 265 woman trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind her and asks "Aren't you people supposed to be good at math?" The Asian woman looks uncomfortable but says nothing. Q3.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q3.9 If Q3.2 = No Q3.3 Do you feel obligated to help the Asian woman in this situation? o Yes (1) o No (2) Skip To: Q3.9 If Q3.3 = No Q3.4 Is there anything you could do to help the Asian woman in this situation? o Yes (1) o No (2) Skip To: Q3.9 If Q3.4 = No 266 Q3.5 Please list what you think you could do in this event to help the Asian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q3.6 Would you do anything in this situation to help the Asian woman? o Yes (1) o No (2) Skip To: Q3.9 If Q3.6 = No Q3.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q3.5/ChoiceTextEntryValue/1} (1) o ${Q3.5/ChoiceTextEntryValue/2} (2) o ${Q3.5/ChoiceTextEntryValue/3} (3) o ${Q3.5/ChoiceTextEntryValue/4} (4) Page Break 267 Q3.8 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q3.9 The young white man continues as the other customer appears to ignore him. He now says to the Asian woman, "Get out of here, you stupid chink! No one wants you here!" The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q3.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q3.17 If Q3.10 = No Q3.11 Do you feel obligated to help the Asian woman in this situation? o Yes (1) o No (2) Skip To: Q3.17 If Q3.11 = No Q3.12 Is there anything you could do to help the Asian woman in this situation? o Yes (1) o No (2) 268 Skip To: Q3.17 If Q3.12 = No Q3.13 Please list what you think you could do in this event to help the Asian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q3.14 Would you do anything in this situation to help the Asian woman? o Yes (1) o No (2) Skip To: Q3.17 If Q3.14 = No Q3.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q3.13/ChoiceTextEntryValue/1} (1) o ${Q3.13/ChoiceTextEntryValue/2} (2) o ${Q3.13/ChoiceTextEntryValue/3} (3) o ${Q3.13/ChoiceTextEntryValue/4} (4) Page Break 269 Q3.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q3.17 The young white man appears to grow angrier that the Asian woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q3.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q3.18 = No Q3.19 Do you feel obligated to help the Asian woman in this situation? o Yes (1) o No (2) Skip To: End of Block If Q3.19 = No Q3.20 Is there anything you could do to help the Asian woman in this situation? o Yes (1) o No (2) 270 Skip To: End of Block If Q3.20 = No Q3.21 Please list what you think you could do in this event to help the Asian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q3.22 Would you do anything in this situation to help the Asian woman? o Yes (1) o No (2) Skip To: End of Block If Q3.22 = No Q3.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q3.21/ChoiceTextEntryValue/1} (1) o ${Q3.21/ChoiceTextEntryValue/2} (2) o ${Q3.21/ChoiceTextEntryValue/3} (3) o ${Q3.21/ChoiceTextEntryValue/4} (4) Page Break 271 Q3.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q3.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q3.25 != No, I would not contact the police. Q3.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Asian Woman Scenario Start of Block: Gay Man Scenario 272 Q4.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young man who appears to be gay wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and says, "Oh, someone is very out loud and proud!" The gay man looks uncomfortable but says nothing. Q4.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q4.9 If Q4.2 = No Q4.3 Do you feel obligated to help the gay man in this situation? o Yes (1) o No (2) Skip To: Q4.9 If Q4.3 = No Q4.4 Is there anything you could do to help the gay man in this situation? o Yes (1) o No (2) Skip To: Q4.9 If Q4.4 = No 273 Q4.5 Please list what you think you could do in this event to help the gay man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q4.6 Would you do anything in this situation to help the gay man? o Yes (1) o No (2) Skip To: Q4.9 If Q4.6 = No Q4.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q4.5/ChoiceTextEntryValue/1} (1) o ${Q4.5/ChoiceTextEntryValue/2} (2) o ${Q4.5/ChoiceTextEntryValue/3} (3) o ${Q4.5/ChoiceTextEntryValue/4} (4) Page Break 274 Q4.8 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q4.9 The young white man continues as the other customer appears to ignore him. He now says to the gay man, "Get out of here, you stupid faggot! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q4.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q4.17 If Q4.10 = No Q4.11 Do you feel obligated to help the gay man in this situation? o Yes (1) o No (2) Skip To: Q4.17 If Q4.11 = No Q4.12 Is there anything you could do to help the gay man in this situation? o Yes (1) o No (2) 275 Skip To: Q4.17 If Q4.12 = No Q4.13 Please list what you think you could do in this event to help the gay man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q4.14 Would you do anything in this situation to help the gay man? o Yes (1) o No (2) Skip To: Q4.17 If Q4.14 = No Q4.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q4.13/ChoiceTextEntryValue/1} (1) o ${Q4.13/ChoiceTextEntryValue/2} (2) o ${Q4.13/ChoiceTextEntryValue/3} (3) o ${Q4.13/ChoiceTextEntryValue/4} (4) Page Break 276 Q4.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q4.17 The young white man appears to grow angrier that the gay man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q4.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q4.18 = No Q4.19 Do you feel obligated to help the gay man in this situation? o Yes (1) o No (2) Skip To: End of Block If Q4.19 = No Q4.20 Is there anything you could do to help the gay man in this situation? o Yes (1) o No (2) Skip To: End of Block If Q4.20 = No 277 Q4.21 Please list what you think you could do in this event to help the gay man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q4.22 Would you do anything in this situation to help the gay man? o Yes (1) o No (2) Skip To: End of Block If Q4.22 = No Q4.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q4.21/ChoiceTextEntryValue/1} (1) o ${Q4.21/ChoiceTextEntryValue/2} (2) o ${Q4.21/ChoiceTextEntryValue/3} (3) o ${Q4.21/ChoiceTextEntryValue/4} (4) Page Break 278 Q4.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q4.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q4.25 != No, I would not contact the police. Q4.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Gay Man Scenario Start of Block: Lesbian Woman Scenario 279 Q5.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young woman who appears to be lesbian wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind her and says, "Oh, someone is very out loud and proud!" The lesbian woman looks uncomfortable but says nothing. Q5.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q5.9 If Q5.2 = No Q5.3 Do you feel obligated to help the lesbian woman in this situation? o Yes (1) o No (2) Skip To: Q5.9 If Q5.3 = No Q5.4 Is there anything you could do to help the lesbian woman in this situation? o Yes (1) o No (2) Skip To: Q5.9 If Q5.4 = No 280 Q5.5 Please list what you think you could do in this event to help the lesbian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q5.6 Would you do anything in this situation to help the lesbian woman? o Yes (1) o No (2) Skip To: Q5.9 If Q5.6 = No Q5.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q5.5/ChoiceTextEntryValue/1} (1) o ${Q5.5/ChoiceTextEntryValue/2} (2) o ${Q5.5/ChoiceTextEntryValue/3} (3) o ${Q5.5/ChoiceTextEntryValue/4} (4) Page Break 281 Q5.8 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q5.9 The young white man continues as the other customer appears to ignore him. He now says to the lesbian woman, "Get out of here, you stupid dyke! No one wants you here!" The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q5.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q5.17 If Q5.10 = No Q5.11 Do you feel obligated to help the lesbian woman in this situation? o Yes (1) o No (2) Skip To: Q5.17 If Q5.11 = No Q5.12 Is there anything you could do to help the lesbian woman in this situation? o Yes (1) o No (2) 282 Skip To: Q5.17 If Q5.12 = No Q5.13 Please list what you think you could do in this event to help the lesbian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q5.14 Would you do anything in this situation to help the lesbian woman? o Yes (1) o No (2) Skip To: Q5.17 If Q5.14 = No Q5.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q5.13/ChoiceTextEntryValue/1} (1) o ${Q5.13/ChoiceTextEntryValue/2} (2) o ${Q5.13/ChoiceTextEntryValue/3} (3) o ${Q5.13/ChoiceTextEntryValue/4} (4) Page Break 283 Q5.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q5.17 The young white man appears to grow angrier that the lesbian woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q5.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q5.18 = No Q5.19 Do you feel obligated to help the lesbian woman in this situation? o Yes (1) o No (2) Skip To: End of Block If Q5.19 = No Q5.20 Is there anything you could do to help the lesbian woman in this situation? o Yes (1) o No (2) Skip To: End of Block If Q5.20 = No 284 Q5.21 Please list what you think you could do in this event to help the lesbian woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q5.22 Would you do anything in this situation to help the lesbian woman? o Yes (1) o No (2) Skip To: End of Block If Q5.22 = No Q5.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q5.21/ChoiceTextEntryValue/1} (1) o ${Q5.21/ChoiceTextEntryValue/2} (2) o ${Q5.21/ChoiceTextEntryValue/3} (3) o ${Q5.21/ChoiceTextEntryValue/4} (4) Page Break 285 Q5.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q5.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q5.25 != No, I would not contact the police. Q5.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Lesbian Woman Scenario Start of Block: Transgender Man Scenario 286 Q6.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young person who appears to be a transgender man wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind him and says, "You know you're still a woman, right?" The transgender man looks uncomfortable but says nothing. Q6.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q6.9 If Q6.2 = No Q6.3 Do you feel obligated to help the transgender man in this situation? o Yes (1) o No (2) Skip To: Q6.9 If Q6.3 = No Q6.4 Is there anything you could do to help the transgender man in this situation? o Yes (1) o No (2) Skip To: Q6.9 If Q6.4 = No 287 Q6.5 Please list what you think you could do in this event to help the transgender man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q6.6 Would you do anything in this situation to help the transgender man? o Yes (1) o No (2) Skip To: Q6.9 If Q6.6 = No Q6.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q6.5/ChoiceTextEntryValue/1} (1) o ${Q6.5/ChoiceTextEntryValue/2} (2) o ${Q6.5/ChoiceTextEntryValue/3} (3) o ${Q6.5/ChoiceTextEntryValue/4} (4) Page Break 288 Q6.8 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q6.9 The young white man continues as the other customer appears to ignore him. He now says to the transgender man, "Get out of here, you stupid tranny! No one wants you here!" The other customer now appears scared as he says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q6.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q6.17 If Q6.10 = No Q6.11 Do you feel obligated to help the transgender man in this situation? o Yes (1) o No (2) Skip To: Q6.17 If Q6.11 = No Q6.12 Is there anything you could do to help the transgender man in this situation? o Yes (1) o No (2) 289 Skip To: Q6.17 If Q6.12 = No Q6.13 Please list what you think you could do in this event to help the transgender man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q6.14 Would you do anything in this situation to help the transgender man? o Yes (1) o No (2) Skip To: Q6.17 If Q6.14 = No Q6.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q6.13/ChoiceTextEntryValue/1} (1) o ${Q6.13/ChoiceTextEntryValue/2} (2) o ${Q6.13/ChoiceTextEntryValue/3} (3) o ${Q6.13/ChoiceTextEntryValue/4} (4) Page Break 290 Q6.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q6.17 The young white man appears to grow angrier that the transgender man hasn't left the store yet. He quickly moves closer and shoves him violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q6.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q6.18 = No Q6.19 Do you feel obligated to help the transgender man in this situation? o Yes (1) o No (2) Skip To: End of Block If Q6.19 = No Q6.20 Is there anything you could do to help the transgender man in this situation? o Yes (1) o No (2) Skip To: End of Block If Q6.20 = No 291 Q6.21 Please list what you think you could do in this event to help the transgender man. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q6.22 Would you do anything in this situation to help the transgender man? o Yes (1) o No (2) Skip To: End of Block If Q6.22 = No Q6.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q6.21/ChoiceTextEntryValue/1} (1) o ${Q6.21/ChoiceTextEntryValue/2} (2) o ${Q6.21/ChoiceTextEntryValue/3} (3) o ${Q6.21/ChoiceTextEntryValue/4} (4) Page Break 292 Q6.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q6.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q6.25 != No, I would not contact the police. Q6.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Transgender Man Scenario Start of Block: Transgender Woman Scenario 293 Q7.1 Imagine that you're out purchasing a few items at your local pharmacy before going to meet with some friends. While you're trying to figure out the best deals, you see a young person who appears to be a transgender woman wearing a rainbow pride shirt trying to choose between two different types of chips on the same aisle as you. A young white man then walks behind her and says, "You know you're still a man, right?" The transgender woman looks uncomfortable but says nothing. Q7.2 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q7.9 If Q7.2 = No Q7.3 Do you feel obligated to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: Q7.9 If Q7.3 = No Q7.4 Is there anything you could do to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: Q7.9 If Q7.4 = No 294 Q7.5 Please list what you think you could do in this event to help the transgender woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q7.6 Would you do anything in this situation to help the transgender woman? o Yes (1) o No (2) Skip To: Q7.9 If Q7.6 = No Q7.7 Which of the actions you previously listed (reproduced below) would you take? o ${Q7.5/ChoiceTextEntryValue/1} (1) o ${Q7.5/ChoiceTextEntryValue/2} (2) o ${Q7.5/ChoiceTextEntryValue/3} (3) o ${Q7.5/ChoiceTextEntryValue/4} (4) Page Break Q7.8 Would you contact the police in this situation? o Yes (1) o No (2) 295 Page Break Q7.9 The young white man continues as the other customer appears to ignore him. He now says to the transgender woman, "Get out of here, you stupid tranny! No one wants you here!" The other customer now appears scared as she says, "Look, I'm just trying to buy a few things before I head home. I'll be out of here soon enough." Q7.10 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: Q7.17 If Q7.10 = No Q7.11 Do you feel obligated to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: Q7.17 If Q7.11 = No Q7.12 Is there anything you could do to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: Q7.17 If Q7.12 = No 296 Q7.13 Please list what you think you could do in this event to help the transgender woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q7.14 Would you do anything in this situation to help the transgender woman? o Yes (1) o No (2) Skip To: Q7.17 If Q7.14 = No Q7.15 Which of the actions you previously listed (reproduced below) would you take? o ${Q7.13/ChoiceTextEntryValue/1} (1) o ${Q7.13/ChoiceTextEntryValue/2} (2) o ${Q7.13/ChoiceTextEntryValue/3} (3) o ${Q7.13/ChoiceTextEntryValue/4} (4) Page Break 297 Q7.16 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q7.17 The young white man appears to grow angrier that the transgender woman hasn't left the store yet. He quickly moves closer and shoves her violently to the ground before yelling, "I'm gonna kick your ass if you don't leave!" Q7.18 Do you consider what you've observed to be a problem? o Yes (1) o No (2) Skip To: End of Block If Q7.18 = No Q7.19 Do you feel obligated to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: End of Block If Q7.19 = No Q7.20 Is there anything you could do to help the transgender woman in this situation? o Yes (1) o No (2) Skip To: End of Block If Q7.20 = No 298 Q7.21 Please list what you think you could do in this event to help the transgender woman. (Up to 4 actions) o Action #1 (1) __________________________________________________ o Action #2 (2) __________________________________________________ o Action #3 (3) __________________________________________________ o Action #4 (4) __________________________________________________ Q7.22 Would you do anything in this situation to help the transgender woman? o Yes (1) o No (2) Skip To: End of Block If Q7.22 = No Q7.23 Which of the actions you previously listed (reproduced below) would you take? o ${Q7.21/ChoiceTextEntryValue/1} (1) o ${Q7.21/ChoiceTextEntryValue/2} (2) o ${Q7.21/ChoiceTextEntryValue/3} (3) o ${Q7.21/ChoiceTextEntryValue/4} (4) Page Break 299 Q7.24 Would you contact the police in this situation? o Yes (1) o No (2) Page Break Q7.25 If you were to encounter a scenario like described above that escalated across the three stages (microaggression, derogatory slur, physical violence), would you contact the police? If so, at which point of escalation? o No, I would not contact the police. (1) o Yes, from the microaggression onwards. (5) o Yes, from the derogatory slur onwards. (6) o Yes, once physical violence was used. (7) Skip To: End of Block If Q7.25 != No, I would not contact the police. Q7.26 You indicated that you would not contact the police. Why wouldn't you contact the police in a scenario like the one described above? o I have an issue or issues with the police (e.g., bad experiences, lack of trust) (1) o I would use an alternate reporting mechanism (e.g., 211 in Los Angeles County) (2) o I didn't see a need to contact the police in that scenario (3) o Something else (please specify) (4) __________________________________________________ End of Block: Transgender Woman Scenario Start of Block: Bystander Efficacy 300 Q8.1 Please answer the following questions about what you think about "violence prevention." Violence is when people fight or hurt others on purpose. Violence prevention means keeping violence from happening or stopping violence before it starts. Choose only one answer that best describes your response: Q8.2 People's violent behavior can be prevented. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) Q8.3 There are certain things a person can do to help prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) 301 Q8.4 I myself can make a difference in helping to prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) Q8.5 People can be taught to help prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) 302 Q8.6 Doing or saying certain kinds of things can help prevent violence o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) Q8.7 I can learn to do or say the kinds of things that help prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) 303 Q8.8 People can learn to become someone who helps others to avoid violence o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) Q8.9 Even people who are not involved in a fight can do things that help prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) 304 Q8.10 Even when I'm not involved and it's not about me, I can make a difference in helping to prevent violence. o Agree completely (1) o Agree a lot (2) o Agree a little (3) o Disagree a little (4) o Disagree a lot (5) o Disagree completely (6) End of Block: Bystander Efficacy Start of Block: Bystander Behavior Q9.1 For these next four questions, please indicate a yes or no response based upon whether you have done what is indicated in the PAST TWO MONTHS. Q9.2 I indicate my displeasure when I hear sexist jokes. o Yes (1) o No (2) Q9.3 I indicate my displeasure when I hear racist jokes. o Yes (1) o No (2) 305 Q9.4 I indicate my displeasure when I hear homophobic jokes. o Yes (1) o No (2) Q9.5 I indicate my displeasure when I hear cat-calls. o Yes (1) o No (2) Page Break Q9.6 For these three questions, please indicate a yes or no response based upon whether you have done what is indicated in your lifetime. Q9.7 Have you ever intervened when you saw a violent crime unfolding? o Yes (1) o No (2) Q9.8 Have you ever intervened when you saw a non-violent crime unfolding? o Yes (1) o No (2) 306 Q9.9 Have you ever intervened when you saw a hate crime unfolding? (whether violent or non-violent) o Yes (1) o No (2) End of Block: Bystander Behavior Start of Block: Feeling Thermometers Q10.1 Report your feelings toward groups that are in the news these days. To do so, use a feeling thermometer. Ratings between 50 degrees and 100 degrees mean that you feel favorable and warm toward the group. Ratings between 0 degrees and 50 degrees mean that you don't feel favorable toward the group and that you don't care too much for that group. If you don't feel particularly warm or cold toward the group, you would rate that group at the 50-degree mark. 0 10 20 30 40 50 60 70 80 90 100 Straight men () Straight women () Gay men () Lesbian women () Transgender men () Transgender women () White men () White women () Asian men () Asian women () End of Block: Feeling Thermometers Start of Block: Basic Empathy Scale - Adult 307 Q11.1 Please indicate the extent to which you agree or disagree with the following statements. Strongly agree (1) Somewhat agree (2) Neither agree nor disagree (3) Somewhat disagree (4) Strongly disagree (5) My friends' emotions don't affect me much. (1) o o o o o After being with a friend who is sad about something, I usually feel sad. (2) o o o o o I can understand my friend's happiness when he/she does well at something. (3) o o o o o I get frightened when I watch characters in a good scary movie. (4) o o o o o I get caught up in other people's feelings easily. (5) o o o o o 308 I find it hard to know when my friends are frightened. (6) o o o o o I don't become sad when I see other people crying. (7) o o o o o When someone is feeling down, I can usually understand how they feel. (8) o o o o o I can usually tell when my friends are scared. (9) o o o o o I often become sad when watching sad things on TV or in films. (10) o o o o o I can often understand how people are feeling even before they tell me. (11) o o o o o 309 Seeing a person who has been angered has no effect on my feelings. (12) o o o o o I can usually tell when people are cheerful. (13) o o o o o I tend to feel scared when I am with friends who are afraid. (14) o o o o o I can usually realize quickly when a friend is angry. (15) o o o o o I often get swept up in my friends' feelings. (16) o o o o o My friends' unhappiness doesn't make me feel anything. (17) o o o o o I am not usually aware of my friends' feelings. (18) o o o o o I have trouble figuring out when my friends are happy. (19) o o o o o 310 End of Block: Basic Empathy Scale - Adult Start of Block: Dictator and Ultimatum Games Q12.1 Suppose you are given a sum of money which you must decide how to share with another person. You must decide how to split the total up, and you may do this any way you like. But, there is one rule: Should the other person not like your offer, they cannot punish you in any way and they have no choice, they have to accept your offer. Here are some sums of money you are given. Please write the amount that you would offer the other person under this rule in dollars. (NUMERIC ENTRY VALIDATION) You offer... (1) ...$10,000 (13) ...$5,000 (14) ...$500 (15) ...$100 (16) Page Break 311 Q12.2 Now we are going to change the rule. The new rule in this game is that once you decide how to split the money, the other person can either accept your offer in which case you both get what is proposed, or the other person can reject the offer, in which case you both would get $0. Here are some sums of money you are given. Please enter the amount that you would offer the other person under this new rule in dollars. (NUMERIC ENTRY VALIDATION) You offer... (1) ...$10,000 (13) ...$5,000 (14) ...$500 (15) ...$100 (16) End of Block: Dictator and Ultimatum Games Start of Block: Demographics 312 Q13.1 What is your age group? o 18-29 years old (1) o 30-39 years old (2) o 40-49 years old (3) o 50-59 years old (4) o 60-69 years old (5) o 70 years or older (6) Q13.2 Are you of Hispanic, Latino, or Spanish origin? o No, not of Hispanic, Latino, or Spanish origin (1) o Yes, Mexican, Mexican American, Chicano (2) o Yes, Puerto Rican (3) o Yes, Cuban (4) o Yes, another Hispanic, Latino, or Spanish origin (e.g., Salvadoran, Dominican, Colombian, Guatemalan, Spaniard, Ecuadorian, etc.) (5) 313 Q13.3 What is your race? (please select only one option that best reflects your race) o American Indian or Alaska Native (1) o Asian or Asian American (2) o Black or African American (3) o Native Hawaiian or other Pacific Islander (4) o White (5) o Multiracial (6) o Something else (please specify) (7) __________________________________________________ Q13.4 What is your gender? o Male (1) o Female (2) o Transgender male (3) o Transgender female (4) o Non-binary/gender non-conforming (5) o Something else (please specify) (6) __________________________________________________ 314 Q13.5 What is your sexual orientation? o Heterosexual or straight (1) o Gay or lesbian (2) o Bisexual (3) o Something else (please specify) (4) __________________________________________________ Q13.6 What is the highest degree or level of education you have completed? o Less than high school graduate (1) o High school graduate (including equivalency or GED) (2) o Some college credit or associate's degree (e.g., AA, AS) (3) o Bachelor's degree (e.g., BA, BS) (4) o Graduate or professional degree (e.g., MA, MS, MD, DDS, JD, PhD) (5) Q13.7 Please provide your employment status o Employed full time or part time (1) o Unemployed and seeking work (2) o Unemployed and NOT seeking work (3) o Retired (4) o Something else (please specify) (5) __________________________________________________ End of Block: Demographics 315 Start of Block: Other controls Q14.1 Regarding your political orientation, would you consider yourself to be: o Strongly liberal (1) o Somewhat liberal (2) o Slightly liberal (3) o Neither liberal nor conservative (4) o Slightly conservative (5) o Somewhat conservative (6) o Strongly conservative (7) Q14.2 Are you registered with a political party? If so, which party? o Yes, democratic party (1) o Yes, republican party (2) o Yes, a different party (please specify) (3) __________________________________________________ o No, I am not registered with a political party (4) 316 Q14.3 What is your religious affiliation? o Protestant (1) o Roman Catholic (2) o Mormon (3) o Jewish (4) o Muslim (5) o Hindu (6) o Buddhist (7) o Another religion (please specify) (8) __________________________________________________ o No religious affiliation (9) Skip To: Q14.5 If Q14.3 = No religious affiliation Q14.4 How important, if at all, would you say religion is in your own life? o Very important (1) o Somewhat important (2) o Not at all important (3) Q14.5 Have you ever been arrested for any offense(s) other than traffic violations? o Yes (1) o No (2) 317 Q14.6 Have you ever spent time incarcerated in a jail, prison, or juvenile detention center? o Yes (1) o No (2) Q14.7 Please give your best estimate of the TOTAL AMOUNT of income received by all members of your household BEFORE TAXES (combine income for yourself and anyone who is living or staying at your address as their primary residence) during the PAST 12 MONTHS. o Less than $10,000 (1) o $10,000 - $19,999 (2) o $20,000 - $29,999 (3) o $30,000 - $39,999 (4) o $40,000 - $49,999 (5) o $50,000 - $59,999 (6) o $60,000 - $69,999 (7) o $70,000 - $79,999 (8) o $80,000 - $89,999 (9) o $90,000 - $99,999 (10) o $100,000 - $149,999 (11) o More than $150,000 (12) End of Block: Other controls |
| Clean Full Text | (not set) |
| Language | (not set) |
| Doi | (not set) |
| Arxiv | (not set) |
| Mag | (not set) |
| Acl | (not set) |
| Pmid | (not set) |
| Pmcid | (not set) |
| Pub Date | 2024-01-01 08:00:00 |
| Pub Year | 2024 |
| Journal Name | (not set) |
| Journal Volume | (not set) |
| Journal Page | (not set) |
| Publication Types | (not set) |
| Tldr | (not set) |
| Tldr Version | (not set) |
| Generated Tldr | (not set) |
| Search Term Used | Jehovah's AND yearPublished>=2024 |
| Reference Count | (not set) |
| Citation Count | (not set) |
| Influential Citation Count | (not set) |
| Last Update | 2024-11-10 00:00:00 |
| Status | 0 |
| Aws Job | (not set) |
| Last Checked | (not set) |
| Modified | 2025-01-13 22:07:06 |
| Created | 2025-01-13 22:07:06 |