GENDER IDENTITY THREAT AND HARASSMENT: ASSESSING THE IMPACT OF IDENTITY THREAT ON RESPONSES TOWARDS GENDER NONCONFORMING OTHERS By Catherine Viola Dodson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology – Master of Arts 2014 ABSTRACT GENDER IDENTITY THREAT AND HARASSMENT: ASSESSING THE IMPACT OF IDENTITY THREAT ON RESPONSES TOWARDS GENDER NONCONFORMING OTHERS By Catherine Viola Dodson Although studies concerning male-female sexual harassment have proliferated, same-sex genderbased harassment remains an understudied area in social psychology. I sought to expand upon the gender harassment and identity threat literature by examining the impact of gender identity threat and partner gender nonconformity on expressions of harassment and incivility by men in same-sex virtual encounters. I also attempted to replicate the findings of Funk and Werhun (2011) regarding the impact of gender identity threat on men’s cognitive functioning and identity salience. Results indicated that men who experienced gender identity threat (being told by a female confederate that they “had a grip like a girl”) were more likely to send homophobic jokes to their interaction partner, compared to men who did not experience gender identity threat. Additionally, men who were exposed to a gender nonconforming male interaction partner were more likely to send sexist jokes to their partner compared to men with gender conforming interaction partner. However, neither gender identity threat nor partner gender nonconformity significantly increased the likelihood of defection in the Prisoner’s Dilemma or inequitable task assignment. Implications of these findings are discussed. ACKNOWLEDGMENTS I owe great gratitude to the many people who have helped me, academically and otherwise, to reach this important academic milestone. First, I would like to give thanks to my committee chair and academic advisor, Isis Settles, who has been a consistent source of knowledge and encouragement over the course of this project. Her patience is truly legendary. I would also like to thank the other members of my committee, Nicole Buchanan and Joseph Cesario, for their consistently helpful and insightful comments and assistance through this process. Next, there are no words for the support and strength provided to me by my mother, Elizabeth Pearson. She inspires me, comforts me, teaches me, and shapes me into a better person each and every day, and without her I would have never achieved as much as I have. I must also thank my friends, Jackson Katz, Caroline Wise, Quincy Knapp, Kat Zubko, Bethany Singer-Biefsky, Chris Sole, and Lauren Solomon, among others. I owe more to them than I can ever say. Finally, I would like to thank my research assistants for their invaluable help with data collection on this project, Aurora Walchak-Arndt, Carly Mills, Cici Cheng, Mira Friedman, and Paige Erickson. iii TABLE OF CONTENTS LIST OF TABLES ………………………………………………………………………………vi LIST OF FIGURES ……………………………………………………………………………...ix Introduction ……………………………………………………………………………………….1 Social Identity Theory and Identity Threat …………………………..………...…………2 Defining Sexual Harassment vs. Gender Based Harassment ……………...……...5 Why People Harass ……………………………………………………………………….8 Power and Dominance ……………….…………………………………………...9 Gender Role Conformity ….……………………………………………………………..12 Expressions of Sexual and Gender Harassment …….…………………..…………….....14 Computer Harassment Paradigm …….………………………………………….15 Social Dilemmas and Cooperation ….………………………..………………….17 Incivility and Task Allocation ….……………….………………………………20 The Current Study ….……………………………………………………………………21 Hypotheses ………………………………………………………………………………………26 Primary ………………………………………………………………………………….26 Secondary ………………………………………………………………………………..27 Method …………………………………………………………………………………………..28 Participants ………………………………………………………………………………28 Procedure ………………………………………………………………………………..29 Measures ………………………………………………………………………………...33 Measures following identity threat manipulation ……………………………….33 Cognitive Ability ………………………………………………………..33 Gender Identity Salience ………………………………………………...33 Dependent measures following exposure to gender conformity stimuli ………..34 Manipulation Check and Partner Questions ….…………………..……..34 Computer Harassment Paradigm………………………………………...34 Social Dilemma Task …………………………………………………....35 Task Allotment …………………………………………………………..36 Measures after experiment……………………………………………………….37 Demographics………………………………..…………………………..37 Suspicion Check………………………………….……………………...37 Pilot Analyses………………………………………………………..…………………..37 Results ...…………………………………………………………………………………………39 Descriptives ...……………………………………………………………………………39 Hypothesis 1 ……………………………………………………………………………..46 Homophobic Jokes: 1a – 1c ..……………………………………………………46 Sexist Jokes: 1a – 1c …..………………………………………………………...48 Prisoner’s Dilemma: 1a – 1c ……………………………………..……………..49 iv Unfavorable Task Assignment: 1a – 1c ……………..…………………………..50 Hypothesis 2 …………………………………………………………………………….51 Gender Identity Threat …………………………………………………………..51 Partner Gender Conformity Threat ……………………………………………...59 Hypothesis 3 …………………………………………………………………………….66 Hypothesis 3a ……………………………………………………………………66 Hypothesis 3b ……………………………………………………………………66 Hypothesis 3c ……………………………………………………………………66 Discussion ……………………………………………………………………………………….68 Limitations and Future Directions ……………………………………………………....79 Conclusion ………………………………………………………………………………82 APPENDICES ……………………..……………………………………………………………84 APPENDIX A Instructions Given to Participants …………………………….………...85 APPENDIX B : Modified Anagram Task ………………………………………………86 APPENDIX C: “Who Am I/I Am…” Identity Assessment …………………………….88 APPENDIX D: Individual Self-Description……………………………………………..89 APPENDIX E: Gender Conformity/Nonconformity Stimuli …………………………...90 APPENDIX F : Partner Assessment …………………………………………...……….92 APPENDIX G: Computer Harassment Paradigm ………………………………………94 APPENDIX H: Social Dilemma ………………………………………………………...96 APPENDIX I: Task Allocation ………………………………………………………...102 APPENDIX J: Computer Harassment Paradigm Review ….…………………………..103 APPENDIX K: Demographic Information ……………………………………………110 REFERENCES ………………………………………………………………………………...112 v LIST OF TABLES Table 1. Means and standard deviations for demographic variables ...........................................................................................................................................39 Table 2. Means, standard deviations, and alphas for all main study variables ...........................................................................................................................................42 Table 3. Correlations between demographic variables and partner assessment and continuance variables ...........................................................................................................................................43 Table 4. Correlations between demographic variables, continuance variables, and outcome variables ...........................................................................................................................................44 Table 5. Correlations between partner assessment and outcome variables ...........................................................................................................................................45 Table 6. F statistics for all dependent variables by all conditions ...........................................................................................................................................46 Table 7. Means and standard deviations for homophobic jokes by all conditions ...........................................................................................................................................47 Table 8. Means and standard deviations for sexist jokes by all conditions ...........................................................................................................................................48 Table 9. Means and standard deviations for total defection in prisoner’s dilemma by all conditions. ...........................................................................................................................................49 Table 10. Means and standard deviations for unfavorable task assignment by all conditions ...........................................................................................................................................50 Table 11. Coefficients for mediation pathways of gender identity threat by mediation outcome variables ...........................................................................................................................................53 Table 12. Coefficients for mediation pathways of gender identity threat by partner femininity ...........................................................................................................................................54 Table 13. Coefficients for mediation pathways of gender identity threat by partner masculinity ...........................................................................................................................................54 vi Table 14. Coefficients for mediation pathways of gender identity threat by partner sexuality ...........................................................................................................................................55 Table 15. Indirect effect and confidence intervals of mediation analyses outcome variables by gender identity threat ...........................................................................................................................................57 Table 16. Indirect effect and confidence intervals of partner femininity by gender identity threat ...........................................................................................................................................57 Table 17. Indirect effect and confidence intervals of partner masculinity by gender identity threat ...........................................................................................................................................58 Table 18. Indirect effect and confidence intervals of partner sexuality by gender identity threat ...........................................................................................................................................58 Table 19. Coefficients for mediation pathways of partner gender conformity threat by mediation outcome variables ...........................................................................................................................................60 Table 20. Coefficients for mediation pathways of partner gender conformity threat by partner femininity ...........................................................................................................................................61 Table 21. Coefficients for mediation pathways of partner gender conformity threat by partner masculinity ...........................................................................................................................................61 Table 22. Coefficients for mediation pathways of partner gender conformity threat by partner sexuality ...........................................................................................................................................62 Table 23. Indirect effect and confidence intervals of mediation analyses outcome variables by partner gender conformity threat ...........................................................................................................................................64 Table 24. Indirect effect and confidence intervals of partner femininity by partner gender conformity threat ...........................................................................................................................................64 Table 25. Indirect effect and confidence intervals of partner masculinity by partner gender conformity threat ...........................................................................................................................................65 Table 26. Indirect effect and confidence intervals of partner sexuality by partner gender conformity threat ...........................................................................................................................................65 vii Table 27. Means and standard deviations of cognitive and identity salience measures. ...........................................................................................................................................67 viii LIST OF FIGURES Figure 1. Model of Hypothesis 3 …… ………………………………………………………….24 Figure 2. Model of Hypotheses 1 and 2 ...……………...………………………………………..25 ix Introduction Same-sex gender-based harassment is an issue of growing concern and occurs across a variety of contexts, such as in peer relationships during adolescence and adulthood, and in the workplace, especially those that are male-dominated (DuBois et al., 1998; Pryor, 1987). Though sexual harassment between men and women (i.e., opposite-sex sexual harassment) remains the most ubiquitous form of harassment, the prevalence and impact of same-sex harassment (i.e., male-male or female-female) is not as well understood. Recent findings concerning masculine identity threat provide support for examining its relation to harassment behavior. Specifically, given the importance of sustaining a culturally valued masculine identity in a patriarchal society, men are theorized to place great emphasis on the presentation of traditional masculinities (Funk & Werhun, 2011; Stockdale et al., 1999). For men who value traditional masculinity, the fear of their masculinity being threatened can serve as motivation to engage in sexual and gender harassment towards others (Hitlan et al., 2009). Research has provided support for the argument that threats to male identity increase the likelihood of sexual and gender-based harassment towards both gender deviant women and men (Funk & Werhun, 2011; Maass et al., 2005). This research has given rise to many questions, particularly, what drives within-sex gender harassment? And what role does gender identity threat play in the enactment of gender-based harassment behavior? The current study explored same-sex harassment under induced identity threat among male undergraduates. Cognitive ability and gender identity salience were measured as indicators of the extent of identity threat experienced. Furthermore, because gender nonconformity in a female target plays an important role in male expressions of sexual and gender-based harassment, the gender presentation of the 1 male target in this study was manipulated to present a range of gender conformity and nonconformity to see if similar processes occur with a male target. Importantly, this study is one of the few to experimentally assess gender-based harassment behavior. Employing the computer harassment paradigm as well as the Prisoner’s Dilemma model, I tested whether exposure to gender nonconformity resulted in more negative assessments, more incivility and reduced cooperation with a male target (i.e., male partner), especially when the participant is under conditions of identity threat. In this paper, I will review the literature concerning social identity threat as it relates to behavior. Next, I will define genderbased harassment as a subtype of the broader construct of sexual harassment. Following, I will also discuss the prevalence of opposite-sex and same-sex sexual harassment across contexts, the outcomes of harassment, the ways in which harassment is enacted, and the various reasons for engaging in harassment. Particular emphasis will be placed on harassment motivations relating to policing gender nonconformity. Social Identity Theory and Identity Threat Social identity theory (SIT; e.g., Tajfel, 1974) posits that belonging to a group serves as an important part of one’s self identity. This sense of belonging and identification comprises social identities, defined as conceptions of one’s self as a social entity belonging to a group. These groups may be either clearly delineated from others (such as groups based on gender, race, or class), or more abstract and ambiguous (such as multiracial or bisexual identities), but they must hold value and significance to one’s sense of self. Identification with a given social category serves the important function of substantiating needs for communion with others. It also provides norms and prototypes with which to identify, while ideally allowing for individuals to 2 fulfill needs of personal distinctiveness (Brewer, 1991; Ellemers et al., 2002; Tajfel, 1974; Tajfel & Turner, 1986). Although identities vary in personal significance, features of the social context are important determinants of which identities are cognitively accessible to the individual, and the social context in turn shapes behavior and perception within a given situation (Ellemers et al, 2002). Categorization only occurs to provide meaningful organization of social and situational stimuli (Ellemers et al., 2002), thus the social context is an important determinant for whether identities are accessible and applied cognitively. The social context, according to social identity theory is regarded as particularly powerful, and can be either a source of security for a given identity, or a source of threat (Ellemers et al., 2002; Tajfel & Turner, 1986). Context provides important feedback about one’s identity in relation to another’s. This can include the social position of one’s group relative to another’s, or more simple feedback, such as the gender or race of another person relative to oneself. Features or actions of an outgroup member can engender a threat to one’s own self or group (i.e., identity threat; Ellemers et al., 2002). Identity threat, particularly among those that identify highly with a given social category, can also result from more internal or perceptual concerns, such as a fear of a lack of acceptance in a valued group (Glick et al., 2007; Herek, 1986; Cramer, 1998). Gender identity threat in men for example, can be induced by negative feedback regarding their perceived sexuality or gender presentation. Men given false feedback that they seemed more homosexual were more likely to respond defensively by self-describing themselves as masculine, and describe a target figure as homosexual, than men who had not received false feedback about their sexuality (Cramer, 1998). Gender contrary feedback (i.e., telling the participant that he does not fit masculine gender norms) can also induce similar defense 3 mechanism activation and result in identity threat (Cramer, 1998; Glick et al., 2007; Hitlan et al., 2009). Thus, threats to one’s identity can engender a host of behavioral and cognitive effects, including engaging in out-group discrimination (Ellemers et al., 2002). Threats to masculine identity can arise from myriad sources, including having women present in predominantly male organizational environments, men presenting gender in a nontraditional manner, and/or men fearing that they have deviated from the norms of the masculine category (Ellemers et al., 2002; Funk & Werhun, 2011; Hitlan, et al, 2009; Maass et al., 2003). Funk and Werhun found that when male participants’ gender identity is threatened by a female experimenter (by being told that they “had a grip like a girl”) they performed significantly worse on anagram completion tasks and the Stroop measure of attentional selfregulation. Participants were also more likely to list masculine related category references in an open-ended self-description task (“Who Am I”) than those in the no harassment condition, demonstrating increased salience of their threatened male social identity. Maass et al. (2003) further demonstrated that identity threat can engender negative behavior, specifically harassment, in their analysis of the relationship between experiences of identity threat and gender-role harassment towards women. Male university students were subjected to identity threat and, following threat exposure, were given the opportunity to send pornographic material to a virtual female partner. Men who had experienced identity threat were more likely to harass the female partner, particularly those men who were highly identified with their masculinity. Herek (1986) argued that defensiveness and threat can arise from male insecurities about their personal adequacy in meeting gender-role demands. This sense of inadequacy, as well as more direct threats to masculinity (such as gender contrary feedback) can result in increased identification with gender related characteristics, and expressions of homophobia and 4 discrimination towards gender nonconforming others (Glick et al., 2007; Herek, 1986; Hitlan et al., 2009, Maass et al. 2003). These results are supported by Maass et al. (2003), who found that men who experienced identity threat were more likely to identify with masculine characteristics and attributes. Glick et al.’s (2007) analysis, men who were provided bogus feedback on their “feminine” personality were more likely to exhibit negative affect when exposed to gender nonconforming gay men than gender conforming gay men. This research demonstrates that identity threat, and specifically threats to an individual’s gender identity, can result in increased salience of the masculine category, and increased negative affect following exposure to gender nonconformity. Further research has shown that the presence of masculine identity threat in male participants can serve as a predictor of their likelihood to respond to gender nonconformity in others with gender policing and harassment. This includes harassing women in traditionally masculine fields (e.g., the military, police officers, etc.) or men who present gender in a nontraditional way (Funk & Werhun, 2011; Hitlan et al., 2009). In sum, limited evidence from current research suggests that males with threatened identities respond with increased salience of their masculine identity, sexual harassment of women, derogation of effeminate gay men, negative affect following exposure to gender nonconformity in other men, and reductions in cognitive ability and attentional self-regulation (Cramer, 1998; Dall’Aria & Maass, 1999; Ellemers et al., 2002; Funk & Werhun, 2011; Glick et al., 2007; Hitlan, 2009; Maass et al., 2003). The current study builds upon this foundational work by examining whether men who experience identity threat will also harass other men, especially those who are perceived as gender nonconforming. Defining Sexual Harassment vs. Gender Based Harassment. Sexual harassment has been defined broadly as harassing behavior that denigrates and frightens targets based upon their 5 sex; sexual harassment frequently extends over a period of time and results in many negative outcomes for victims (Fitzgerald, 1993). Sexual harassment classifications are commonly delineated into either legal or psychological frameworks (Fitzgerald, Gelfan, & Drasgow, 1995). Legal definitions, established under Title VII of the Civil Rights Act of 1964 (Grochin & Kleiner, 1998), generally separates sexual harassment behavior into two categories: quid pro quo and hostile environment harassment (E.E.O.C., 1980). Quid pro quo harassment is the act of making employment benefits and opportunities conditional upon the receipt of sexual favors or cooperation. Hostile environment harassment is the subjection of employees to severe and pervasive unwanted sexual behavior, because of their sex, that has the effect of substantially interfering with the work environment. Legal definitions were further expanded to include “unwelcome sexual advances, requests for sexual favors, and other verbal or physical conduct of a sexual nature” (Berdahl, 2007b, p. 641). More simply, Berdahl (2007b) argues that sexual harassment is commonly conceptualized as a form of sex-based harassment characterized by “behavior that derogates, demeans, or humiliates an individual based on that individual’s sex” (p. 644). Psychological classifications of opposite-sex sexual harassment favor the three dimensional model identified by Fitzgerald et al. (1995) in her theoretical conceptualization of sexual harassment: gender harassment, unwanted sexual attention, and sexual coercion. Gender harassment is characterized by lewd comments and negative remarks about the target’s sex or gender. Unwanted sexual attention is evidenced in behaviors such as unwanted touching or pressure for dates, and sexual coercion is sexual bribery and the use of threats. Psychological dimensions identified by Fitzgerald et al. (1993) map on to legal definitions in a number of ways. Quid pro quo harassment is most similar to the sexual coercion dimension identified by 6 Fitzgerald et al. (1995), as both concern the attempt to force sexual cooperation with threats or bribery. Hostile environment harassment shares a number of similarities with both the dimensions of gender harassment and unwanted sexual attention. The legal system first acknowledged the presence and impact of same-sex gender-based harassment in Oncale v. Sundowner Offshore Services, Inc. (1996). In their decision, the Supreme Court designated that unwelcome sexual conduct by someone of the same-sex as the intended target could be defined legally as sexual harassment. The court ruled that the sexual orientation of the perpetrator and/or target is ultimately irrelevant to produce a finding of sexual harassment. Lower courts further broadened the definition of the behaviors that fall under the designation of harassment. Whereas the courts initially defined sexual harassment in terms of opposite-sex harassment (generally towards female targets), later rulings would deem same-sex sexual harassment as a valid and recognizable form of harassment as well (Grant & Kleiner, 1998). Despite the growing legal history around same-sex sexual harassment, there has been a surprising deficit in psychological research (DuBois et al., 1998; Knapp & Kustis, 2000; Stockdale, Visio & Batra, 1999). The dimension of gender harassment, and hostile environment harassment, have been particularly applicable to the study of same-sex harassment as the corresponding behavior is less likely to be sexually coercive in nature (Gerrity, 2000; Willness et al., 2007). Leskinen, Cortina, and Kabat (2011) argue that it is organizational and legal practice to privilege sexual advance forms of sexual harassment (i.e., sexual coercion), at the same time as neglecting the important dimension of gender harassment. It is likely that this overemphasis in organizational and legal practices is also present in social psychological research. For this reason, the current study was designed to employ experimental methodology to assess gender harassment behavior towards 7 same-sex targets, and provide further knowledge into the relationship between male gender nonconformity and expressions of harassment. Why People Harass There are many potential reasons individuals engage in gender-based harassing behavior towards those of the same sex. Gerrity (2000) and others (Berdahl, 2007; Funk & Werhun, 2011; Hitlan et al., 2009) have posited that there is a common misconception that sexual harassment, and to a lesser extent gender harassment, is primarily motivated by sexual interest. Rather, theorists propose that sexual harassment is typically the result of the desire to express power and dominance over others (Cortina & Wasti, 2005). Recent laboratory analyses have also provided support for the argument that threats to one’s gender identity can also lead to harassment behavior (Funk & Werhun, 2011; Glick et al., 2007), as individuals seek to reinstate feelings of lower power and dominance. Legal frameworks also initially assumed that sexual harassment was based in sexual desire of the target by the perpetrator. For example, quid pro quo was originally defined as a subordinate woman experiencing sexual harassment from a male supervisor or boss; sexual harassment was viewed as an act necessarily motivated by the target’s sex and the harasser’s sexual desire. However, Berdahl’s (2007a) findings, as well as other research, dispel this misconception. Women in Berdahl’s study reported greater experiences of sexual harassment regardless of the extent to which they fit a gender or sexual ideal. Instead, women with more masculine personalities were substantially more likely to have experienced sexual harassment than feminine women, and this was not due to a greater likelihood to perceive acts as harassing. Similar work has found that those who defy gender norms or lack organizational power are most subject to sexual harassment (Stockdale et al., 1999; Uggen & Blackstone, 2004; Weiner et al., 8 2012).Thus, it seems that sexual interest and attraction rarely underlie sexual harassment, even in opposite-sex harassment. Power and Dominance. Analyses of sexual and gender harassment in organizational settings have emphasized the role of power and dominance as key motivational forces underpinning the expression of sexual harassment behaviors overall. Traditional sexual harassment research is characterized by an emphasis on power-based models that assume the association of power and harassment are automatic (Lucero et al., 2007). This means that research often presumes that the enactment of sexual harassment is necessarily related to power hierarchies within organizations. Cleveland and Kerst (1993) defined power in the context of sexual harassment as a varied and complex construct, with power resulting from societal, organizational, interpersonal, or individual factors. Men are more likely to hold positions of power and dominance, with women being less likely than men to be perceived as even having power in organizational or workplace settings (Begany & Milburn, 2002; Rospenda, Richman & Nawyn, 1998; Uggen & Blackstone, 2004). Because men have greater access to power than women in organizations, the sexual harassment literature has emphasized disparate access to power as a crucial feature of the male-female sexual harassment framework. Power hierarchies in the military have also been shown to be related to experiences and expressions of sexual harassment, with men of greater rank or organizational status being “protected” from many forms of harassment (Settles et al., 2012). Settles et al. (2012) found that the sexual harassment of men and women by same-sex peers and supervisors was relatively prevalent, with higher ranking individuals being less likely to be harassed. Further study of women’s experiences of sexual harassment in the military have also demonstrated that officers overall report less sexual harassment, including sexual and gender harassment, unwanted sexual 9 attention and sexual coercion, than enlisted women (Buchanan, Settles, Woods, 2008, p. 351). Ultimately these findings demonstrate that status can be protective against sexual harassment. Those who harass are usually older than their targets, married, and have reputations for sexually exploitative behavior, as well as an increased likelihood to hold higher level organizational positions than their targets (Lucero et al., 2007). However, power differences at the organizational level are not limited to supervisor-supervisee relationships (Cleveland & Kerst, 1993; Fineran, 2002; McMaster et al., 2002). Even among co-workers power differences can be present through informal networks, differential support, and the latitude of decision making allotted by the supervisor. These subtle sources of power are complex, and potentially less influential than more markedly differential power hierarchies. In instances where power differentials are present, as is often the case in male-female sexual harassment in the workplace, those with less power are more commonly subject to harassment (Uggen & Blackstone, 2004). However, examining more equitable power relations and their relation to harassment, such as harassment among peers rather than between bosses and subordinates, has significant benefits, particularly to the study of same-sex harassment (Lucero et al., 2007). In those instances where power is more or less equitable between harasser and target (i.e., peer harassment), sexual harassment could be evidence of a desire for greater organizational power as achieved through maligning those of a similar or higher status, or contrapower harassment (Cleveland & Kerst, 1993; DeSouza & Soldberg, 2004). Contrapower harassment is defined as harassment that occurs between a subordinate and their superior (e.g., the subordinate harasses the superior), and has gained prominence in recent years as research has demonstrated that harassment within organizations can occur outside of power dichotomies. 10 Uggen and Blackstone (2004) cited many theorists who posit that sexual harassment is a gendered expression of power and dominance over others. Common sociocultural systems that privilege heteronormative masculinity (characterized by the belief in heterosexuality as the norm and homosexuality as deviant) in the gendered workplace as well as age relations that give adult men greater rights than young people, comprise power arrangements that contribute to sexual harassment expression. However, these power arrangements can also be more intersectional and interpersonal. Expressions of sexual harassment are not limited to supervisor-subordinate relationships (Cleveland & Kerst, 1993; Pina, Gannon, & Saunders, 2009). Rospenda et al. (1998) contribute to this understanding in their analysis of the intersectional components of harassment expression. In their theoretical model, Rospenda et al. emphasized the role of gender, race, and class as salient aspects of workplace harassment. These features are critical to our understanding of both traditional (male-female, supervisor-subordinate) sexual harassment dynamics as well as nontraditional (male-male, female-female, female-male, subordinatesupervisor) dynamics. Nontraditional sexual harassment dynamics are particularly likely to be characterized by intersectional and interpersonal relationships rather than ones based in organizational hierarchy (Rospenda et al., 1998). Given the preponderance of findings that demonstrated the relationship between perceived power and likelihood to engage in harassment, the current study placed the participant in the role of “team leader” over the gender conforming/nonconforming confederate. It is from this position of perceived power that participants interacted with the confederate, given that research demonstrated higher power individuals were more likely to engage in sexual harassment. 11 Gender Role Conformity Given the importance of sustaining a culturally valued masculine identity in a patriarchal society, men are theorized to place great emphasis on the presentation of traditional masculinities (Funk & Werhun, 2011; Stockdale et al., 1999). This emphasis can result in both behavioral conformity in men’s own selves, and gender role policing in others. Many theories on gender role ideologies conceive of gender identity as achieved through a socialization process beginning at a very early age that defines men and women as fundamentally different. This results in behavioral prescriptions that promote different sets of behavioral norms depending on gender (Hamburger et al., 1996). Examples of male gender-typed characteristics and behaviors include dominance, aggression, and sexual agency, whereas female gender-typed behavior is characterized by passivity, submissiveness, and sexual objectification. Endorsement of traditional and restrictive gender ideologies has been shown to be related to expressions of sexual aggression, sexism, and opposite-sex gender harassment (Hamburger & Hegben, 1996; Herek, 1986; Leaper & Van, 2008; Pryor, 1995; Stockdale et al., 1999) suggesting a link between gender role norms and harassment. For those men who value traditional masculinity, the fear of their masculinity being threatened can serve as motivation to engage in sexual and gender harassment (Hitlan et al., 2009). Such threat, combined with the rejection of the feminine in themselves as well as others (Bem & Lenney, 1976; Glick et al., 2007; Pryor, Giedd, & Williams, 1995), can underlie motivations to police gender nonconformity. Policing gender is characterized by the rejection of nontraditional gender roles and gender nonconformity either by hyperconformity to gender roles or harassment and discrimination towards gender deviant people (Cramer, 1998; Maass et al., 2003; Weiner et al., 2012). 12 Many theorists have explored the underlying motivation of sexual harassment as a way to maintain traditional, hierarchical gender structures and roles (Konik & Cortina, 2008; Stockdale et al., 1999). Stockdale et al. (1999) argued that sexual harassment behaviors are necessarily defined by the extent to which the conduct is used to enforce or perpetuate these traditional views on gender, views comprised of “hypergender” norms and stereotypes. These “hypergender” norms are composed of a binary definition of hypermasculinity (a rigid definition of gender-role stereotype of masculine norms, such as endorsements of violence, objectification of women, and rejection of femininity) and hyperfemininity (a rigid female gender-role stereotype characterized by the belief that women are passive, sexual objects; Hamburger & Hogben, 1996). Endorsement of hypergender norms and stereotypes typify a rejection of gender nonconformity, such as homosexuality or the violation of traditional gender roles. Sexual harassment can then serve as a function of this rejection, with sexual or gender harassing behaviors reinforcing rigid gender norms and stereotypes that benefit traditionally masculine men at the same time as subjugating women and gender nonconforming men. Street et al. (2007)’s findings support this interpretation; military men in their sample were more likely to report experiencing forms of sexual harassment involving the enforcement of rigid masculine gender norms rather than unwanted sexual attention and sexual coercion, an important difference from their female participants. These results support the perspective that many forms sexual harassment behavior can be categorized as serving the motivation of perpetuating gender norms rather than serving the sexual interest of the perpetrator, especially in same-sex harassment encounters. Stockdale et al. (1999) and others (Berdahl, 2007a; Konik & Cortina, 2008) stress that sexual harassment also has the effect of perpetuating hypermasculine and heterosexist standards. 13 Heterosexism and the endorsement of heterosexist standards relate to the desire to police gender nonconformity. Genders are traditionally defined in heterosexist terms (i.e., defining males and females as naturally and ubiquitously heterosexual). Given this, homosexual or gender nonconforming individuals, particularly men, are judged to be in violation of heterosexist standards that underlie traditional views about gender (Konik & Cortina, 2008; Silverchanz et al., 2008, Street et al., 2007). This violation makes them subject to gender policing, often through the use of sexual and gender harassment. Theories of gender identity threat (Funk & Werhun, 2011; Glick et al, 2007) substantiate this perspective, as men whose masculinity has been threatened are more likely to view gender nonconforming men more negatively and report reduced affectivity following exposure to effeminate versus traditionally masculine gay men. In sum, based upon the current research, the argument can be made that men who experience a threat to their masculine identity also experience a loss of power (Herek, 1986). To reclaim this power, men may be motivated to engage in sexual and gender-based harassment as a means of asserting their masculine identity (Berdahl, 2007a; Konik & Cortina, 2008; Stockdale et al., 1999). Men who are gender nonconforming are especially likely to be targets of this harassment because the underlying motivation for sexual and gender-based harassment is to police traditional gender norms (Konik & Cortina, 2008; Silverchanz et al., 2008; Street et al., 2007). These processes may be especially likely to occur when the perpetrator is a man in a position of power over his target. Expressions of Sexual and Gender Harassment Expressions of gender harassment in this study were assessed with a series of behavioral and perceptual measures. The behavioral measures included the computer harassment paradigm, a Prisoner’s Dilemma paradigm, and a task allocation assignment where participants were 14 provided with the opportunity to act with civility and allot tasks equitably or act with incivility and allot tasks inequitably. These measures provide a unique addition to the field of harassment research, which is noted for its lack of experimental analyses and behavioral outcomes (Dall’Aria & Maass, 1999; Mitchell et al., 2004; Pina, Gannon, Saunders, 2009; Pryor, Giedd, Williams, 1995; Siebler et al., 2008). Computer Harassment Paradigm. Sexual harassment behaviors range from relatively benign forms, such as telling sexist jokes at the work place, all the way up to sexual blackmail and sexual aggression (Dall’Area & Maass, 1999). The computer harassment paradigm was created to reflect this variance in harassment behavior. Siebler, Sabelus, and Bohner (2008) presented a computerized paradigm simulating online harassment, an increasingly common form of sexual harassment, to assess sexual harassment behavior in the laboratory. Given that the majority of perpetrators of sexual harassment are men, and the majority of victims of sexual harassment are women, their study used male participants and female targets. However, the proposed paradigm can be adapted for other gender combination beyond that of the male-female dynamic, as it will be in the current analysis. The procedure of the computer harassment paradigm was as follows: male participants were made to believe they were participating in a computer chat in which they had to exchange images with a (fictitious) female chat partner. In each trial, subjects could choose from neutral images and a selection of pornographic images. Sexual harassment behavior was operationalized within the paradigm as the act of choosing to send one or more of the pornographic images despite knowing that doing so is inappropriate. The paradigm has advantages over correlational research in that it represents an experimental method by which researchers can directly examine harassment behavior, manipulate theoretically interesting variables, and examine causal factors 15 that may impact harassment. Siebler, Sabelus, and Bohner (2008) were able to successfully create a validated refined version of the computer harassment paradigm by using less blatant stimulus materials (e.g., sexist jokes rather than pornographic vignettes or images) and removing cues to the offensive nature of the materials within the context of the experimental situation. Pornographic images were replaced with sexist jokes (always targeting women). Sending the sexist jokes rather than neutral material was defined as constituting gender harassment. Siebler et al. manipulated the target characteristics of the female chat partner, and demonstrated that the partner who was described as holding feminist attitudes was more harassed than the female chat partner described as holding traditional attitudes. The researchers posited that females who deviate from traditional attitudes are viewed as posing a threat to male dominance, thus making them more likely to be victims of harassment. This interpretation is consistent with a social identity interpretation. Importantly, the computer harassment paradigm marks an important progression in behavioral analyses of sexual and gender-based harassment behavior. Typically, laboratory (in vivo) analyses of sexual harassment have been limited by both the ethical constraints that proscribe exposing unsuspecting research participants, as well as research confederates, to harassing behavior, as well as the simple difficulty of engendering harassment behavior in subjects in the lab. Employing this methodology, subjects are able to express harassment behavior in a lab setting towards a target they believe to be real, without having to expose a participant or research confederate to actual harassment. The computer harassment paradigm was introduced to overcome problems of experimental harassment assessment by providing a highly realistic harassment setting online that can be manipulated by experimenters (Maass et al., 2003). 16 The current research used the computer harassment paradigm to measure the responses of male participants towards gender nonconforming male interaction partners. The paradigm is beneficial, as compared to correlational studies, because variables can be systematically varied and it provides insight into their causal effects on harassment. This was particularly relevant in the current study, as the gender conformity of the interaction partner as well as the harassment materials were manipulated in order to examine variability in harassment responses. Social Dilemmas and Cooperation. The primary goal in this study was to examine what drives within-sex harassment. Harassment behavior was operationalized in this study not only as behavior and comments that deride an individual based on their gender (as will be measured with the computer harassment paradigm portion of the study), but also as a reduced interest in cooperating with or trusting the interaction partner based upon their gender presentation. This behavior is meant to reflect real-world behavior and responses to gender nonconformity outside of laboratory settings. To assess reduced cooperation and trust, a social dilemma model was used. Liebrand (1983) argued that social dilemmas have value because, during life, many important decisions occur in settings in which individuals are interdependent, with one’s outcomes and another’s being tied together. The interdependence of the individuals results in a situation where an actor’s decision affects both the actor’s outcomes and those of another person. Social dilemmas seek to replicate this interdependence. Dawes (1975) defined social dilemmas as situations in which each person involved has an available dominating strategy (i.e., one that yields the best payoff in all circumstances for the actor). However, if all actors in a social dilemma collectively choose the dominating strategy then the resulting outcome for all parties is negative and yields the least utility for everyone. Thus, uniform defection (betraying the other 17 person to receive a better payoff at that person’s expense), where all parties betray one another, is undesirable. Cooperation within social dilemmas is often influenced by the status of the interaction partner as an in-group versus out-group member (Ando, 1999; Kollock, 1998; Simpson, 2006). Social identity, or “we” feeling, has been found to increase cooperative behavior in social dilemmas (Simpson, 2006). When social identities are made salient, cooperation in social dilemmas is influenced by in-group members’ perceptions of the shared and mutual interchangeable nature of theirs and their group’s interests. Social identification shifts can alter what constitutes self-interest (one of the foremost goals in social dilemma tasks) in social dilemmas (Simpson, 2006). The process by which self-interest becomes subsumed beneath greater cooperative motivations is rooted in many potential causes. One such reason is that outcomes between group members are viewed as interchangeable when the partner is viewed as an in-group member. Further, there may be a greater expectation that in-group members will cooperate more than out-group members in social dilemmas (Simpson, 2006) Ultimately, although the process remains unclear, research has demonstrated that salience of a shared social identity (Brewer, 1979, 1981) can lessen the perceived distance between oneself and other group members, leading to greater weight being placed on joint (collective) over individual gains. The role of in-group status is particularly interesting as the current study seeks to explore the cooperative tendencies of male participants towards gender conforming/nonconforming samesex others under induced identity threat. The current study used the two-person prisoner’s dilemma (PD), perhaps the most commonly investigated social dilemma, to assess cooperation towards the interaction partner. The dilemma was characterized by two players confronting two options, ‘cooperate’ and 18 ‘defect,’ resulting in four potential individual outcomes that the players’ joint actions can generate: mutual cooperation, mutual defection, unilateral defection, unilateral cooperation (Simpson, 2006). The payoffs in PD are ordered such that unilateral defection > mutual cooperation > mutual defection > unilateral cooperation. The best possible outcome for the actor is to defect while their partner cooperates, and the second best outcome is mutual cooperation, followed by mutual defection. The worst scenario is the one in which the actor cooperates, but their partner defects (Kollock, 1998). Under the current model, participants had the social category of gender made salient. The participants, given the salient social category and the deviance of the gender nonconforming male from the prototype of that category, should have view the gender nonconforming interaction partner as an out-group member. Differential perceptions of the interaction partner as similar to the in-group prototype potentially mediated cooperation in the social dilemma task, such that participants exhibit greater likelihood of defection when their partner is gender nonconforming, than when the partner is gender conforming. To asses this mediated hypothesis, measures of partner liking, perceived similarity of the interaction partner to themselves and the prototypical male were included in the final study as mediators for harassment behavior and cooperation. These measures of in-group perception were based upon Tajfel et al’s (1974) theories regarding self-categorization as well as the findings of Glick et al. (2007) concerning men’s rejection of effeminate and gender nonconforming men following induced identity threat. To assess the extent to which the partner was perceived as a member of the in-group, assessment of the partner’s perceived similarity to the in-group prototype were measured (Tajfel & Turner, 1986). These measures also served as indicators of the extent to which the partner was perceived as conforming or deviating from traditional gender norms and characteristics (Glick et al., 2007). 19 Incivility and Task Allocation. Current advances in antidiscrimination laws and sexual harassment suits have served to make blatant employment discrimination based on gender, race, and sexual orientation (in most settings) illegal or, at the very least, non-normative. These changes in the acceptability of expressions of discrimination in the last four decades has resulted in “modern” expressions of discrimination, which often take the form of incivility and generalized hostility. Perhaps most importantly, Cortina (2008) states that changing norms regarding discrimination have potentially resulted in a system where “incivility, sexual harassment, and racial harassment are, at times, one and the same” (p. 57). Thus, noncooperation, hostility, and exclusion may be common expressions of gender harassment. Increasingly, research on gender harassment behaviors in the workplace have focused on expressions of everyday acts of incivility, or general incivility, as elements of discrimination and harassment expression that may be expressions of gender harassment. Cortina et al. (2001) specifically cited workplace incivility as a common element of psychological mistreatment that, although often unexplored by researchers, can have a pernicious effect on targets. Behaviors associated with general incivility include those characteristically associated with rude and discourteous behavior, including displaying a lack of regard for others and even psychological aggression. Cortina et al.’s analysis of incivility viewed it as behavior whose intentionality or purpose is unclear, making their interpretation difficult. Cortina’s (2008) later research defined contemporary expressions of “modern” incivility in organizations as a method of communicating sexism or racism in organizational settings without being explicitly discriminatory. These subtle forms of incivility serve as a more insidious representation of covert discrimination and prejudice. 20 The task allotment assignment included in this study served as a measure of incivility. Participants were provided with the opportunity to favor themselves over their partner by assigning themselves more desirable tasks, or to cooperate and distribute tasks in an equitable manner. This measure ideally provided the necessary subtlety to allow for the expression of less hostile and out-right forms of gender-based harassment and discrimination. As Cortina et al. (2001) noted, “modern” incivility can be best characterized by less explicit forms of discrimination, represented in everyday behavior. Assigning and completing tasks within groups is a common feature of a student’s educational experience. Thus, this assignment provided a familiar context in which to express incivility and discrimination. Participants were told that they had the opportunity to act as the “leader” for themselves and their partner by deciding who completed a series of task in the final section of the study, after identity threat has been induced and they had been exposed to their partner (who was either gender conforming or nonconforming). The tasks varied in desirability, and incivility was determined by the number of undesirable tasks the participants allocated for their partner. This study sought to examine the likelihood for inequitable versus equitable division of tasks (that vary in perceived desirability),depending on the whether the participant had experienced masculine identity threat and the gender conformity or nonconformity of the interaction partner. The Current Study Previous research on same-sex gender and sexual harassment has focused on examinations of the incidence, prevalence, and outcomes associated with harassment in various settings. However, experimental analyses of even traditional sexual harassment dynamics (menwomen) are few. Unsurprisingly, there are even fewer studies addressing same-sex harassment, particularly those concerning gender presentation as a key factor. There is difficulty inherent in any attempt to measure individual harassing and discriminatory behavior. Even in anonymous 21 surveys, people are reluctant to report they have sexually harassed someone, in both opposite-sex and same-sex contexts (McMaster et al. 2002; Pryor, 1987). Given this, the current study sought to employ methods that indirectly measured discrimination, reduced cooperation, and harassment responses to perceived gender nonconformity in a same-sex other. The goal of the proposed research was to empirically examine how gender identity threat and partner gender conformity influence responses to a same-sex partner (i.e., perceptions of similarity and liking, as well as cooperation, harassment, and incivility). This proposed study used a 2 (identity threat: no threat, threat) x 2 (gender conformity: conforming, gender nonconforming) between-subjects design. Identity threat was induced through having a male participant test their handgrip strength in the lab and either receive negative feedback from a female researcher that they have a grip “like a girl” (identity threat condition), or not receive any feedback on their grip (no threat condition). Following this, participants read a self-description from an “interaction partner” that provided exposure to the gender conformity or nonconformity stimuli. Participants were “introduced” to the interaction partner early in the study, and while the partner’s sex is definitively stated as male, their gender presentation was more variable. The selfdescriptions were characterized by either masculine (gender conforming) or feminine (gender nonconforming) typed interests and behavior. I sought to expand on previous research analyzing identity threat and same-sex harassment1 (Berdahl, 2007b; DuBois et al., 1998; Funk & Werhun, 2011; Pryor, 1995; 1 Participant personality and attitudes, as indicated by scores on measures that have been found to correlate strongly with opposite-sex harassment, were assessed online prior to the main portion of the study; however, they are not a focus of the thesis predictions. These measures included the Ambivalent Sexism Inventory (Glick & Fiske, 1996), the Male Role Norms Inventory-Revised (Levant & Richmond, 2007), the Social Dominance Orientation scale (Pratto, Siddanius, Stallworth & Malle, 1994), the Authoritarianism questionnaire (Cribbs & Austin, 2010) and the Likelihood to Sexually Harass scale (Pryor, 1995). 22 Stockdale et al., 2004; Waldo, Berdahl, & Fizgerald, 1998) in four ways. First, I attempted to replicate the findings established by Funk and Werhun (2011) regarding the impact of identity threat on cognition and attentional self-regulation by mirroring their experimental design (see Figure 1). Second, I examined the relationship between identity threat and harassment by measuring whether men who were told they had a “grip like a girl” were more likely to send homophobic and sexist jokes instead of ‘normal jokes’, express defection in a Prisoner’s Dilemma paradigm, and assign undesirable tasks to an interaction partner, compared to men in the no-identity threat condition. Third, I sought to determine whether men’s likelihood to engage in same-sex gender-based harassment was influenced by the gender presentation of their target. To accomplish this, I manipulated the gender conformity of the interaction partner to present two distinct gender presentations: a gender conforming masculine description, and a gender nonconforming feminine description. Last, I explored the mutual effect of identity threat and gender nonconformity on gender-based harassment by comparing the responses of those in both the identity threat and gender nonconformity conditions to all other groups. I hypothesized that this interaction produced the most negative responses towards a gender nonconforming interaction partner, resulting in the highest rates of harassment, cooperation, and incivility (see Figure 2). 23 - H3a Attentional SelfRegulation Control Condition - H3b Cognitive Ability Identity Threat Condition + H3c Identity Salience Figure 1. Model of Hypothesis 3. Note. Path labels represent the corresponding hypothesis and predicted direction of the relationship. 24 + H1ab Control Condition Harassment - H1ab Identity Threat Condition Cooperation + H1c + H1ab Gender Nonconformity Condition Incivility + H2 Control Condition Partner Assessment Figure 2. Model of Hypotheses 1 and 2. Note. Path labels represent the corresponding hypothesis and predicted direction of the relationship. 25 Hypotheses Primary 1. It is predicted that there will be main effects of identity threat and partner conformity, as well as an interaction between the two. a. Men in the identity threat condition will engage in more gender harassment (i.e., sending offensive jokes, non-cooperation in the Prisoner’s Dilemma, and more unfavorable partner task assignment) toward their partner, regardless of gender nonconformity, compared to men in the no identity threat condition. b. There will be a main effect of partner conformity condition, such that men in the gender nonconformity partner condition will engage in more gender harassment (i.e., sending offensive jokes, non-cooperation in the Prisoner’s Dilemma, and more unfavorable partner task assignment) toward their partner than men in the gender conformity partner condition. c. There will be an interaction between identity threat condition and partner conformity condition. Specifically, compared to when exposed to a gender conforming male, men in the identity threat condition exposed to a gender nonconforming partner, will engage in more gender harassment (i.e., sending offensive jokes, non-cooperation in the Prisoner’s Dilemma, and more unfavorable partner task assignment) toward their partner. 2. The relationship between the main effects of identity threat and gender nonconformity on the outcome measures (harassment, cooperation, and incivility) will be mediated by participant perception of the interaction partner (i.e., liking, similarity to self, and similarity to prototype). 26 Secondary 3. Findings from the Funk and Werhun (2011) experiment will replicate such that men in the identity threat condition will perform significantly worse on tests cognitive ability and attentional self-regulation, and will report increased salience of their gender identity than men in the control condition, reflecting the presence of gender identity threat. a. Men in the identity threat condition will perform significantly worse on the Stroop Task (a measure of attentional self-regulation reflecting the presence of gender identity threat) than men in the control condition. b. Men in the identity threat condition will perform significantly worse on a test of cognitive ability (solving anagrams) than men in the control condition, reflecting the presence of gender identity threat. c. Men in the identity threat condition will report more sum total masculine categories and adjectives in the open-ended self-description (“Who Am I”) measure compared to men in the no threat condition, reflecting identity salience. 27 Method Participants In total, 281 male undergraduate students were recruited through the Psychology Department’s Human Participation in Research website at Michigan State University and received 2 HPR credits for their participation. After removing participants from the sample (see below), 197 participants remained in the final sample. The breakdown of participants by condition was as follows: no identity threat, no gender conformity threat, n = 46; no identity threat, gender conformity threat, n = 66; identity threat, no gender conformity threat, n = 40; identity threat, gender identity threat, n = 45. The majority of participants were White (n = 158), and a much smaller percentage of the sample were Black (n = 16), Asian (n = 13), and Latino (n = 6). Several manipulation checks were included to ensure that participants had read the interaction partner’s self-description and thus understood the gender conformity threat manipulation. When participants were asked to report what their partner’s stated gender was, eight participants reported that they believed their partner to be female, and eight refused to respond. All 16 were excluded from the analyses. Additionally, a question asked participants to report what they really thought the study’s purpose was to see if the deception was successful. These responses were coded by the researcher for level of suspicion. Seven participants identified that they knew they were interacting with a computer rather than an actual partner and were removed from the analyses. A further thirty-one participants posited that the intention of the study was to examine responses to and presentations of various sexual orientations. Given this suspicion, their responses to the outcome measures were analyzed, and due to their marked difference from the majority of the sample these participants were removed. Finally, 21 participants reported that 28 they were not born in the United States, and were removed from the analysis due to potential differences in the sociocultural norms associated with gender conformity and harassment. Thus, a total of 84 participants were removed from the study. Procedure Upon entering the laboratory participants were told they were participating in a study of team work, specifically the different ways that people interact when they are placed into various kinds of teams, including pairs and trios (See Appendix A). Participants were randomly placed in either an identity threat or no identity threat group, determined after they log into the survey software and consent to participation in the lab experiment. They were assigned into the identity threat conditions, characterized by whether the top portion of their introduction to the study had either a “1” or a “2” placed unobtrusively in the corner. The research assistant in the lab approached the participant after they read their introduction, and provided their exposure to the identity threat paradigm, i.e., completing a handgrip task. Given that the intent of this analysis was to, at least in part, replicate the Funk and Werhun (2011) gender identity threat design, the research assistants in this analysis were all female. Funk and Werhun posited that male participants that are the target of gender-based criticism (being told they “have a grip like a girl”) from women would result in the greatest inducement of gender identity threat. Participants were told that recent research supports the argument that physical strength can relate to behavior in team tasks, so their strength is going to be measured prior to beginning the interaction phase of the experiment. Participants placed in group one (who have a “1” in the corner of their introduction) tested their strength on the handgrip provided by the research assistant and were told they have scored below the “average” for other men and that they have a 29 “grip like a girl.” Participants in the second group also completed the handgrip task, but were provided with no feedback on their scores. Participants then completed the first round of measures, including a measure of their cognitive ability assessed by the time it took to complete 24 anagrams (See Appendix B). Total scores reflect the total time to complete 24 anagrams. They then completed a short “Who Am I” identity measure to assess the salience of their male identity (See Appendix C). Participants were then told the researchers wanted to know how an individuals’ personality, attitudes, and liking of each other influence their ability to complete tasks as a team. At the beginning of this task participants completed their own self-description, using Qualtrics survey software. The self-description included information about their work history, goals, selfsummary, hobbies, as well as six things they could not live without. Participants created a username rather than indicating their real name to preserve confidentiality. Following this, participants received their “partner’s self-description” (actually one of two self-descriptions created for the study), which served as the gender conformity experimental stimuli (See Appendix E). After completing their self-description, the participant was randomly assigned a gender conforming (control), or gender nonconforming (experimental) description of an interaction partner. The self-description of the partner as gender conforming or nonconforming was created in several ways. In the self-description narrative, masculine adjectives derived from the Bem Sex Role Inventory (BSRI; Bem, 1978) were used for the conforming, and feminine adjectives were used for the nonconforming partner. Additional information on partner gender conformity/nonconformity was created through the presence or absence of masculine-type or feminine-type hobbies and employment descriptions in their work history goals (business and economics versus social work and education). The self-descriptions 30 provided participants’ exposure to the primary manipulation; therefore participants were told that it was very important they read over the profile closely. Following exposure to the experimental or control material (partner self-description), participants completed an assessment of their partner, as well as the manipulation check, characterized by questions measuring participant awareness of the information through basic recall of some of the factual elements in the self-descriptions. Participants were reminded that their interaction partner is completing similar tasks to better achieve an honest depiction of teamwork, and that their ratings of their partner were not shared with their partner. Partner assessment included questions assessing their general initial liking of and affiliation with their assigned partner, their perceived similarity of their partner to themselves and to the broader MSU “ideal” (or prototypical) student of their gender. Following the introduction with the interaction partner, they completed the behavioral measures of harassment, cooperation, and incivility. These included the computer harassment paradigm, which involved the participants being told they are going to be given the opportunity to send “ice-breaker” jokes to their interaction partner. There were three classes of jokes; homophobic, sexist, and ‘normal.’ They were given the option to pick from a list of 25 possible jokes, 5 of which were sexist, 5 were homophobic, and 15 which were “normal” jokes. These jokes were based on previous computer harassment paradigm analyses (Bahns & Branscombe, 2011; Dall’Aria & Maass, 1999; Mitchell et al., 2004; Siebler et al., 2008). The participant must choose no more than ten jokes to send to their interaction partner. Of primary interest was whether or not the participant decided to send homophobic and/or sexist jokes to the partner. This was the first behavioral measure of gender-based harassment. The sum total of sexist and homophobic jokes chosen provided the quantitative data for this measure. 31 In the next part of the study, which also took place on computers in the lab, participants were able to express cooperation or defection toward their partner through a prisoner’s dilemma task. Participants were given a cover story to explain the dynamics of the task; a “cyberwizard” has locked the participant and their interaction partner in a “diabolical” game with a series of turns, the number of which is unknown to the participant. On each turn of the game, the participant and the “partner” must choose either to cooperate or “compete” (defect) with the interaction partner, after which they will receive coins depending their choices. If both decided to cooperate, they will receive three gold coins. However, if they defected and the partner cooperated, they will receive 5 gold coins. There are many potential likelihoods from which they could choose, and an average will be determined each round that reflects their gains and losses overall. The participants were told that if this average drops below a certain, undisclosed number, they will suffer an “Inconceivably Foul Fate” meant to motivate them to achieve more gains than losses. These instructions are standard and are taken directly from the source website. They were told to try on each and every turn to maximize either their and their partner’s outcome, or their own income (See Appendix G). Participants then completed a task allotment activity as a measure of incivility and discrimination. The task allotment assignment assessed the extent to which perceived gender nonconformity in a same-sex partner impacted the assignment of desirable versus undesirable tasks. Participants were told they are the designated “team leader” and thus had to assign, at their discretion, who would complete five of 10 different clerical, online tasks later in the study. Half of the tasks were more desirable and half were less desirable. The desirability of these tasks has been determined through pilot testing of twenty possible tasks (See Table 2). Although the 32 number of tasks assigned to the participant and the partner are the same, the participants were able to determine who received the more desirable vs. undesirable tasks. Participants were then told that the study has ended and that the clerical tasks would not be completed. They will then report on demographic information and suspicion, after which they were thanked and debriefed. Importantly, when completing their suspicion check at the end of the study, participants were asked to indicate their perception of their partner’s sexual orientation, to assess the extent to which harassing responses toward the partner were predicated on the assumption that the partner was a sexual minority. Measures Measures following identity threat manipulation. Cognitive Ability. The measure of cognitive ability was derived from an analysis by Smith et al. (2006) regarding timed completion of anagrams of varying difficulty and its relation to expressions of self-efficacy and future task performance. The anagrams were modified for the current study by being shortened from the original design. Cognitive ability was measured following exposure to the gender harassment or no gender harassment condition. Participants completed as many solvable anagrams as they could from a list of 24 (See Appendix B; Funk & Werhun, 2011; Smith et al., 2006). Total scores for cognitive ability reflected the sum total number of anagrams solved in five minutes. Gender Identity Salience. Participants completed an open-ended 10-item “I am…/Who am I…” (See Appendix C) word completion task to assess masculine identity threat following exposure to the gender prime. The presence or absence of masculine related social categories and self-descriptions was used to assess the extent of gender role identity salience. Greater numbers 33 of gender identity related categories equated to greater masculine identity salience. The sum of references to male categories were calculated. Dependent measures following exposure to gender conformity experimental stimuli. Manipulation Check and Partner Questions. After reading the self-description and completing the above measures of identity threat, participants completed a series of questions to determine if they correctly understood the gendered information provided in the self-description, with questions including “What are some of your partner’s stated interests?,” and “What are three characteristics of your partner?” Following exposure to the manipulation, participants completed a series of partner assessment questions (See Appendix F). The twenty-one questions were divided into three subcategories: perceived partner similarity to themselves; perceived partner similarity to the prototype; and perceived partner liking. The first subcategory of perceived partner similarity to the participant includes items such as, “I think my partner seems very similar to me,” and “I can see many differences between myself and my work partner.” Similarity of the partner to the prototypical student includes items such as “My work partner is very similar to the average MSU student.” All questions will be rated on a scale from 1 (very untrue) to 5 (very true). Responses were reverse coded as needed and averaged to create a mean of each subcategory. Computer Harassment Paradigm. The computer harassment paradigm is a laboratory analogue to study sexually, and gender, harassing behavior in the laboratory (See Appendix G; Bahns & Branscombe, 2011; Dall’Aria & Maass, 1999; Mitchell et al., 2004). Participants were given the opportunity to send a maximum of ten sexually offensive or homophobic jokes to the gender conforming/nonconforming confederate, though participants did not know how these jokes will be perceived. The primary stimulus was a joke list/questionnaire containing 25 jokes 34 (five homophobic jokes, fifteen clean jokes, and five sexually offensive jokes). Participants were told that they had been given the opportunity to send jokes to their interaction partner as an icebreaker. They were given an option to “send” the jokes to the partner, and they received no cues from the interaction partner as to whether or not they found the jokes humorous or offensive. They received five “normal” jokes after sending their own jokes to the partner to increase the likelihood that participants believed they are interacting with a real partner. Total scores of harassment were measured with the total number of homophobic and/or sexist jokes the participant decided to send to the interaction partner. Those who decided to send the sexist and/or homophobic jokes rather than ‘normal’ jokes were expressing harassment behavior as operationalized by the computer harassment paradigm. The number of sexist and/or homophobic jokes sent overall was measured, with higher numbers indicating greater harassment. Social Dilemma Task. Social dilemmas are defined as situations that pose a tension between individual and collective interests (See Appendix H; Ando, 1999; Poppe, 2005; Simpson, 2006). The participant completed a Prisoner’s Dilemma task meant to assess cooperative tendencies towards a gender conforming or nonconforming same-sex other following gender identity threat inducement. The dilemma was presented as a decision making task, with the direction that participants had to arrive at the best possible decision in a series of scenarios, and this decision could benefit only them or both themselves and their partner as a collective. Either choice was acceptable, and they were told their decisions on each round will not be made known to the partner until the end of the prisoner’s dilemma task. There were several rounds, and their choice to defect continuously resulted in fewer coins and a reduced average overall. The interaction partner was set to Tit-for-Tat, and thus their responses mirrored the choices made by the partner. They initially cooperated, and if the partner defects repeatedly 35 they also defected on all rounds. There were multiple options for the participants; if they chose to defect, they received five coins in the first round, and then 1 coin in the ensuing rounds. If they chose to cooperate, they received five coins in each round. There were a total of five rounds, and cooperation versus defection tendencies were determined by summing their choices to cooperate or “compete” (i.e., defect) across the trials. The purpose of the ‘coins’ and their sum total was to reflect the reward system of the Prisoner’s Dilemma paradigm to increase participant motivation to take the task seriously. Higher numbers indicate greater defection across the trials. Task Allotment. Incivility towards the partner was measured with the extent to which the participant disproportionately assigned undesirable vs. desirable tasks to their partner (See Appendix I). Participants were told they will complete a final portion of the study wherein they and their partner would complete clerical and work-related tasks online. Participants were designated the “team leader” and had the responsibility to assign specific tasks to either themselves or their partner. There were ten total tasks to allocate of varying desirability: five undesirable items and five desirable items (based on the results of pilot testing; see Table 2). The participants allocated the tasks equally with regards to quantity (five for themselves, five for their partner) however participants were able to choose whether they gave themselves or their partner desirable or undesirable tasks (and the number of tasks of each type). The measure was scored according to the mean number of undesirable to desirable tasks participants assigned to themselves as compared to their partner. Higher numbers indicate higher sum total unfavorable tasks assigned. 36 Measures after experiment. Demographics. Participants were asked to indicate their race, gender, age, year in school, current relationship status, sexual orientation, religiosity, and socioeconomic status (See Appendix K). Suspicion Check. After completing all of the measures, participants were asked to respond online to an open-ended question about what they thought the purpose of the study was, what the study was about, and if at any time they thought the study or questions were unusual. They were also asked, prior to the debriefing, what they thought their partner’s sexual orientation was, on a scale from 1 (completely heterosexual) to 5 (completely homosexual). Pilot Analyses Pilot studies were run to determine the validity of 1) the gender conformity and nonconformity stimuli, and 2) the perceived desirability of the tasks for the task allotment. The self-descriptions were created by including masculine or feminine adjectives (Bem, 1978) and by providing masculine or feminine hobbies and employment experiences. A total of eighty-eight male participants taken from the Human Subjects Pool of students from the psychology department reviewed either the gender conforming or gender nonconforming selfdescription and then rated the individual described on a series of measures. Participants rated the perceived similarity of the person in the self-description to most men and most women. The items included the BSRI (Bem Sex Role Inventory; Bem, 1978), the partner similarity measures from the final design (Appendix F), as well as “To what extent do you feel your partner is… Very masculine, Like most men, Very feminine, Like most women” on a scale of 1 (strongly disagree) to 5 (Agree). As shown in Table 1, as predicted, the gender conforming stimuli was viewed as significantly more similar to the participant and the in-group prototype, more 37 competent, less warm, more masculine and less feminine compared to the gender nonconforming stimuli. The only measure on which the two self-descriptions did not differ was the measure of perceived sexual orientation (“How would you describe the sexuality of the person described?” on a scale of 1 (Completely Homosexual) to 5 (Completely Heterosexual). The desirability of the specific tasks comprising the task allotment activity was also measured. Participants were asked to read about twenty tasks (see Table 2) and were instructed to report the extent to which they would be interested in completing each task on a scale of 1 (very disinterested) to 5 (very interested). A descriptive analysis was run to assess differences in the mean desirability of each task, with the most desirable tasks (those with the highest means) being tasks 13, 9, 10, 19, and 12. The five least desirable tasks were tasks 5, 6, 3, 2, and 1, each with means below the median point on the scale (see Appendix I). A paired-samples t-test was conducted to compare overall mean differences between the combined means of the top five most desirable tasks and five least desirable tasks. There was a significant difference in the two combined means of the most desirable tasks (M = 3.33, SD = .69) and least desirable tasks (M = 2.58, SD = .73); t (222) = -15.84, p = .00. Significant differences were also found between the least desirable of the most desirable tasks (M = 3.10, SD = 1.06) and most desirable of the least desirable tasks (M = 2.61, SD = 1.12); t (219) = -5.05, p = .00. 38 Results Descriptives Demographic information for the sample is presented in Table 1. The men who participated in the experiment and were included in the final analyses ranged in age from 18 to 35, and were, on average, 19 years old. The average participant was a sophomore in college and described themselves as having grown up with “more than enough to get by” economically, and reported their current socioeconomic status as “having enough to get by but no extras.” The average religiosity within the sample fell between viewing religion as “somewhat important” and “neither important nor unimportant” in their lives. The stated political views of the participants fell between “somewhat liberal” and “moderate” on average, and the majority of the sample (80%) were White. Table 1 Means and standard deviations for demographic variables Demographic Variable Scale Range 18 - 35 Overall M (SD) 19.74 (1.82) 195 Year in School 1-5 2.01 (1.11) 197 SES Growing Up 1-6 4.24 (.93) 197 SES Now 1-6 3.87 (1.01) 197 Religiosity 1-7 3.81 (1.95) 197 Political Affiliation 1-7 3.81 (1.24) 197 Age n The overall sample means, standard deviations, and Cronbach’s alpha for the main study variables and their subscales are presented in Table 2. Two of the partner assessment items (Partner Femininity and Partner Masculinity) had unclear scale language in the midpoints of the scale so they were removed from the analyses. The scale midpoints were initially supposed to 39 state, for example in the partner femininity item, “neutral in femininity,” but in reality were displayed as “neither feminine nor masculine.” The two items had the same unintended language, and it was a concern that they would confuse participants by referencing masculinity in the femininity items, and femininity in the masculinity items. Thus, respondents who chose the midpoint on the variables were removed from the analyses. This resulted in a lower total sample size within these variables that was not reflected in the preceding partner assessment items. Further, there were 22 missing responses in the partner sexuality item. These participants completed the rest of the items but declined to respond as to their perception of the partner on a scale of 1 (completely homosexual) to 5(completely heterosexual), and thus were not removed from the sample overall. However, these sample differences are reflected in the analyses. Correlations between these demographic variables and the main study variables are presented in Tables 3, 4, and 5. Demographic variables that were associated with main study dependent variables were included as control variables in all of the main analyses. The variables meeting this criterion were childhood socioeconomic status and religion importance. Present day socioeconomic status and political affiliation were strongly correlated with socioeconomic status growing up and religion importance, respectively. I chose the more strongly correlated variables, socioeconomic status grow up and religion importance, to include as control variables. They were also included as covariates due to their theoretical significance. Conservative political beliefs are correlated with traditional gender role beliefs and sexual harassment proclivity (Begany & Milburn, 2002; Berkley & Watt, 2006). Socioeconomic status has been explored in the past for its relationship with sexual harassment likelihood, and was included in this analysis because its inclusion demonstrates a relationship between income and gender-based harassment (Begany & Milburn, 2002). 40 Political affiliation and childhood socioeconomic status were positively related to responses towards the partner, indicating that moderate and conservative political beliefs and higher childhood socioeconomic status were positively related to harassment, cooperation, and incivility responses. The scale anchors ranged from very liberal to very conservative for political affiliation, and from very poor, not enough to get by to extremely well-to-do for childhood socioeconomic status. Thus, the positive relationship these items share with specific outcomes, such as likelihood to send sexist jokes, is such that more conservative political beliefs, and higher childhood socioeconomic status, were related to sending more sexist jokes to the interaction partner. 41 Table 2 Means, standard deviations, and alphas for all main study variables Variable Scale Range a Overall M (SD) n 0 - 10 - 1.03 (1.36) 197 0 - 24 - 9.82 (3.98) 197 Partner Similarity to Self 1-5 .72 3.27 (.75) 194 Partner Similarity to Prototype 1-5 .83 3.88 (.45) 195 Partner Competence 1-4 .69 3.46 (.42) 197 Partner Warmth 1-4 .89 3.20 (.60) 196 Partner Approval 1-5 - 1.48 (.75) 197 Partner Perceived Femininity 1-5 - 1.55 (.96) 177 Partner Perceived Masculinity 1-5 - 3.54 (1.26) 171 Partner Perceived Sexuality 1-5 - 4.14 (1.11) 175 No - 8 196 Yes - 188 196 Cognitive Measures Sum Masculine References/ “Who Am I” Sum Anagrams Completed Partner Assessment Decision to Continue with Same Partner Computer Harassment Paradigm Homophobic Jokes Sent 0-5 - .41 (.63) 197 Sexist Jokes Sent 0-5 - .94 (1.09) 197 Neutral Jokes Sent 0-5 - 3.57 (1.34) 197 Prisoner’s Dilemma: Total Defections 0-1 - .19 (.28) 197 Sum Total Unfavorable Tasks Assigned 0-5 - 2.76 (.95) 197 42 Table 3 Correlations between demographic variables and partner assessment and continuance variables 2 1. Political Orientation 2. Religiosity 1 - 3 4 5 6 .17* - 3. SES Growing Up -.02 -.11 - 4. SES Now .12 .04 .59* - 5. Partner Similarity to Self 6. Partner Similarity to Prototype 7. Partner Competence 8. Partner Warmth -.06 .09 .03 .08 - -.13 -.03 -.10 -.05 .28* - -.08 .07 -.08 -.06 .22* .27* - -.07 .08 .02 .21* .08 -.12 -.14* - 9. Partner Approval .04 -.06 -.08 -.17* -.50* -.16* -.05 -.26* - 10. Partner Femininity 11. Partner Masculinity 12. Partner Sexual Orientation 13. Continue with Same Partner .05 -.06 .06 .10 -.30* -.31* -.39* .45* .11 - -.10 -.04 .01 .01 .19* .27* .48* -.42* -.03 -.65* - .04 .10 -.03 .02 .31* .14 .19* -.31* -.03 -.51* .58* - -.03 -.10 .06 .00 -.13 -.19* -.13 .06 .08 .18* -.21* -.24* Note. *p < .05 43 7 8 9 10 11 12 Table 4 Correlations between demographic variables, continuance variables, and outcome variables 1 2 3 4 5 6 7 8 9 10 1. Political Orientation 2. Religiosity .17* - 3. SES Growing Up -.02 -.11 - 4. SES Now .12 .04 .59* - 5. Continue with Same Partner 6. Masculinity Sum References 7. Anagram Sum Completed 8. Homophobic Jokes Sent 9. Sexist Jokes Sent -.03 -.10 .06 .00 - -.06 -.04 .03 .11 .17* - -.07 .12 -.09 -.00 -.07 .13 - .15* -.04 .01 -.07 -.01 -.01 -.09 - .06 -.03 -.10 -.14 .03 .02 .09 .01 - 10. Total Defection .16* .11 .02 .07 -.07 .14 -.06 .21* .15* - 11. Unfavorable Tasks Assigned Note. *p < .05 .00 .08 .16* -.04 -.12 -.11 .04 .05 .25** .10 44 Table 5 Correlations between partner assessment and outcome variables 1 1. Partner Similarity to Self 2. Partner Similarity to Prototype 3. Partner Competence 4. Partner Warmth 2 3 4 5 6 7 8 9 10 11 12 13 .28* - .22* .27* - .08 -.12 -.14* - 5. Partner Approval -.50* -.16* -.05 -.26* - 6. Partner Femininity -.30* -.31* -.39* .45* .11 - 7. Partner Masculinity .19* .27* .48* -.42* -.03 -.65* - 8. Partner Sexual Orientation 9. Masculinity Sum References 10. Anagram Sum Completed 11. Homophobic Jokes Sent 12. Sexist Jokes Sent .31* .14 .19* -.31* -.03 -.51* .58* - -.06 -.12 -.17* .12 .11 .20* -.06 -.16 - .08 .22* .01 -.00 -.01 -.09 .10 .04 .13 - -.10 -.07 -.01 -.10 .04 .12 -.02 .01 -.01 -.09 - .01 -.01 .08 -.21* .08 -.05 .04 .06 .02 .09 .01 - 13. Total Defection -.17* -.15* -.01 -.10 .03 .01 .02 .03 .14 -.06 .21* .15* - 13. Unfavorable Tasks Assigned -.05 -.02 .00 -.11 -.14* -.06 .01 .01 -.11 .04 .05 .25* .13 Note. *p < .05. 45 Hypothesis 1 It was predicted that there would be main effects of identity threat and partner conformity, as well as an interaction between the two, on responses toward the interaction partner. Four univariate analyses of covariance (ANCOVA) were run by dependent variable (homophobic jokes, sexist jokes, defection within the prisoner’s dilemma, and undesirable task allotment) including both independent variables (identity threat, gender conformity threat), and their interaction. Political orientation and socioeconomic status growing up were included as covariates. The overall ANCOVA statistics for all study variables are shown in Table 6. Table 6 F statistics for all dependent variables by all conditions Homophobic Jokes F p Political Orientation Sexist Jokes Defection F p F p Task Allotment F p 3.69* .05 .53 .47 5.03* .03 .01 .91 .11 .74 1.87 .17 .09 .76 4.70* .03 Gender Identity Threat 4.79* .03 .42 .52 Condition Partner Conformity .15 .70 8.67* .00 Condition Threat Condition X .23 .63 .94 .33 Conformity Condition Note. df = 1, 197 for all variables. *p < .05, **p < .01. .37 .54 2.16 .14 1.48 .23 .30 .58 .12 .73 .40 .53 SES Growing Up Homophobic Jokes: Hypothesis 1a – 1c. Means and standard deviations for the ANCOVA for the number of homophobic jokes sent to the interaction partner are shown in Table 7. The results indicated that there was a significant effect by gender identity threat condition (being told by a female confederate that they “had a grip like a girl” versus no feedback on their grip) on the number of homophobic jokes sent to the partner F(1, 197) = 4.79, MSE = 1.85, p < .05. Specifically, those in the gender identity threat condition sent more 46 homophobic jokes to the partner compared to those in the no gender identity threat condition. However, the effect by partner gender conformity condition (having a gender conforming versus nonconforming partner) on the number of homophobic jokes sent to the partner F(1, 197) = .06, MSE = .15, p =.70, was not significant. Further, contrary to predictions, there was not a significant effect by interaction (gender identity threat x partner gender conformity) on the number of homophobic jokes sent to the partner, F(1, 197) = .23, MSE = .09, p = .63. Thus, Hypothesis 1a was supported; however the partner gender conformity and interaction hypotheses (1b and 1c) were not supported. There was no significant difference in the number of homophobic jokes sent to the partner when the interaction partner was gender nonconforming versus conforming, and the interaction between the gender identity threat and partner gender conformity conditions did not yield significant differences in the number of homophobic jokes sent. Table 7 Means and standard deviations for homophobic jokes by all conditions No Gender Identity Threat M (SD) .33 (.56) Gender Identity Threat M(SD) .48 (.64) M(SD) .40 (.60) Gender Nonconforming Partner .32 (.64) .56 (.66) .42 (.65) Overall .32 (.60)a .53 (.65)b .41 (.63) Gender Conforming Partner Overall Note. Cell means with the different superscripts within rows are significantly different at p < .05. 47 Sexist Jokes: Hypothesis 1a – 1c. For sexist jokes as the dependent outcome (see Table 8), the ANCOVA indicated that there was not a significant effect by gender identity threat condition on the number of sexist jokes sent to the partner F(1, 197) = .42, MSE = .48, p = .52. However, there was a significant effect by partner gender conformity condition (having a gender conforming versus nonconforming partner) on the number of sexist jokes sent to the partner F(1, 197) = 8.67, MSE = 1.06, p < .00. Specifically participants assigned a gender nonconforming partner were more likely to send sexist jokes than those who were exposed to a gender conforming partner. Finally, the interaction between gender identity threat condition and gender conformity condition was not related to the number of sexist jokes sent to the partner F(1, 197) = .94, MSE = 1.06, p =.33. Table 8 Means and standard deviations for sexist jokes by all conditions No Gender Identity Threat M(SD) .63 (.95) Gender Identity Threat M(SD) .70 (.99) M(SD) .66 (.97)a Gender Nonconforming Partner 1.24 (1.15) 1.02 (1.10) 1.15 (1.13)b Overall . .87 (1.06) .94 (1.09) Gender Conforming Partner .99 (1.11) Overall Note. Cell means with the different superscripts within columns are significantly different at p < .05. 48 Prisoner’s Dilemma Defections: Hypothesis 1a – 1c. The results of the ANCOVA using defection in the prisoner’s dilemma paradigm as the dependent variable are presented in Table 9. As shown, neither gender identity threat, F(1, 197) = .37, MSE = .03, p = .54, nor partner gender conformity, F(1, 197) = 1.48, MSE = .12, p = .23, significantly impacted the likelihood of defection within the prisoner’s dilemma. Additionally, the interaction between the gender identity threat and partner gender conformity conditions was not significantly related with likelihood to defect, F(1, 197) = .12, MSE = .01, p = .73. Thus, Hypothesis 1 was not supported for defections. Table 9 Means and standard deviations for total defection in prisoner’s dilemma by all conditions No Gender Identity Threat M(SD) .18 (.29) Gender Identity Threat M(SD) .14 (.27) M(SD) .16 (.28) Gender Nonconforming Partner .22 (.28) .22 (.28) .22 (.28) Overall .20 (.28) .18 (.28) .19 (.28) Gender Conforming Partner 49 Overall Unfavorable Task Assignment: Hypothesis 1a – 1c. Using unfavorable tasks assignment as the dependent outcome (see Table 10), the ANCOVA indicated that there was no significant difference between the gender identity threat conditions, F(1, 197) = 2.16, MSE = 1.91, or the gender conformity conditions, F(1, 197) = .30, MSE = .27, p = .58. Similarly, the interaction of conditions were not significant, F(1, 197) = .40, MSE = .36, p = .53. Thus, Hypothesis 1 was not supported for task assignments. Table 10 Means and standard deviations for unfavorable task assignment by all conditions No Gender Identity Threat M(SD) 2.87 (.98) Gender Identity Threat M(SD) 2.55 (.88) M(SD) 2.72 (.94) Gender Nonconforming Partner 2.85 (.93) 2.71 (.99) 2.79 (.95) Overall 2.86 (.95) 2.64 (.94) 2.76 (.95) Gender Conforming Partner Overall Overall, Hypothesis 1 was only partially supported. On most of the dependent variables, participants did not differ in their responses by threat condition or their interaction. The only statistically significant differences were in the number of homophobic jokes sent to the partner in the gender identity threat condition, and the number of sexist jokes sent to the partner in the partner gender conformity condition. In both instances, those in the threat group (gender identity threat and gender nonconforming partner, respectively) sent significantly more offensive jokes than those in the no threat conditions. 50 Hypothesis 2 Hypothesis 2 examined the mediating role of the partner assessment variables in the relationship between gender identity threat, partner gender nonconformity, and their interaction on expressions of harassment, cooperation, and incivility toward the interaction partner. This hypothesis was tested using the Hayes and Preacher (2014) mediation modeling MEDIATE macro, an alternative to the PROCESS macro, which allowed for the use of multicategorical independent variables and multiple mediators. In the mediation analyses, condition was the independent variable, the harassment variables were included as dependent variables, and political affiliation and socioeconomic status growing up were controlled. The mediation models were tested separately for gender identity threat and partner gender conformity and each of the four dependent variables (homophobic and sexist jokes sent, total defection in the prisoner’s dilemma, and the assignation of undesirable tasks). Partner similarity to self, similarity to the prototypical student at MSU, partner competence, partner warmth, and partner approval were grouped together as multiple mediators within each mediation analysis as multiple mediators. Because of the missing participants for partner femininity, partner masculinity, and partner sexuality (which had 22 missing cases), these three variables were each examined with separate single mediation models by the independent variables. Because the mediation analysis uses pairwise deletion, including the variables with missing data with those variables without missing data would have lowered the sample size for all the mediational analyses, thereby reducing analytic power. Gender Identity Threat. I predicted that the partner assessment variables would mediate the relationship between experiencing gender identity threat and expressions of harassment and incivility towards the interaction partner, regardless of gender conformity. The mediation pathways for gender identity threat by dependent variables through the partner assessment 51 variables are presented in Tables 11 – 14. As shown in Table 11 and Table 14 gender identity threat was significantly related to only one of the dependent variables – homophobic jokes, which is consistent with the findings of Hypothesis 1a. Gender identity threat was not significantly related to any of the mediators. However, two of the mediators, partner warmth and partner approval, were significantly related to two of the harassment and incivility outcome measures; partner warmth was negatively related to the number of sexist jokes sent to the interaction partner, as well as the likelihood to allocate undesirable tasks. Partner approval was also negatively related to the total number of unfavorable tasks assigned to the interaction partner. 52 Table 11 Coefficients for mediation pathways of gender identity threat by mediation outcome variables Dependent Variables Mediator Homophobic Jokes b t .22* 2.34 b -- t -- Gender Identity Threat Partner Similarity to Self .15 1.34 -- Partner Similarity to Self -- -- Gender Identity Threat Partner Similarity to Prototype Partner Similarity to Prototype -.06 Gender Identity Threat Partner Competence Sexist Jokes Total Defection Unfavorable Tasks b t -.08 -.60 b -.10 t -.64 b .15 t 1.34 -- -- -- -- -- -- -- -.10 -1.45 .06 .44 -.06 -2.03 -.84 -2.02 -.84 -- -- -- -- -- -- -- -- -- -- -.02 -.22 -.15 -.80 .02 -.10 -1.57 -.45 -.10 -1.57 -- -- -- -- -- -- -- -- Partner Competence -- -- .07 -1.25 .11 .54 -.10 -1.57 .02 .09 Gender Identity Threat  Partner Warmth .11 1.26 -- -- -- -- -- -- -- -- Partner Warmth -- -- -.10 -.38 -.34* -2.45 .11 1.26 -.26* -2.23 Gender Identity Threat Partner Approval .00 .03 -- -- -- -- -- -- -- -- Partner Approval -- -- -.03 .07 .04 .31 .00 .03 -.34* -3.22 Gender Identity Threat Note. df = 3, 187. *p < .05. 53 Table 12 Coefficients for mediation pathways of gender identity threat by partner femininity Dependent Variables Mediator Homophobic Jokes b t .17 1.81 b -- t -- Gender Identity Threat  Partner Femininity .27 1.79 -- Partner Femininity -- -- .06 Gender Identity Threat Sexist Jokes Total Defection Unfavorable Tasks b t -.23 -1.68 b -.13 t -.80 b -.02 t -.53 -- -- -- -- -- -- -- 1.27 -.04 -.43 .00 .18 -.05 -.71 Note. df = 4, 172. *p < .05. Table 13 Coefficients for mediation pathways of gender identity threat by partner masculinity Dependent Variables Mediator Gender Identity Threat Gender Identity Threat  Partner Masculinity Partner Masculinity b t Homophobic Jokes b t -- -- .15 -.24 -1.24 -- -- Sexist Jokes Total Defection Unfavorable Tasks b t b t b t 1.55 -.17 -1.00 -.01 -.32 -.24 -1.75 -- -- -- -- -- -- -- -- .00 .09 .03 .48 .01 .45 -.01 -.10 Note. df = 4, 166. *p < .05. 54 Table 14 Coefficients for mediation pathways of gender identity threat by partner sexuality Dependent Variables Mediator Homophobic Sexist Jokes Total Defection Jokes b t b t b t b t Unfavorable Tasks b t Gender Identity Threat Gender Identity Threat  Partner Sexuality Partner Sexuality -- -- .25 2.65* -.11 -.62 -.02 .04 -.17 -1.11 -.32 -1.86 -- -- -- -- -- -- -- -- -- -- .03 .61 .05 .70 .00 .16 .07 .97 Note. df = 4, 170. *p < .05. 55 Despite these significant pathways, overall, none of the partner assessment variables significantly mediated the effect of gender identity threat. The multiple mediator analysis (Table 15), which included partner similarity to self, partner similarity to the prototypical male student, partner competence, partner warmth, and partner approval, yielded no significant mediation of the relationship between the gender identity threat conditions and harassment outcomes. Further, perceived partner femininity (Table 16), partner masculinity (Table 17) and partner sexuality (Table 18) did not significantly mediate the relationship between gender identity threat and the harassment and incivility measures. 56 Table 15 Indirect effect and confidence intervals of mediation analyses outcome variables by gender identity threat Outcome Variables Mediator Homophobic Jokes Sexist Jokes Total Defection Unfavorable Tasks Assigned Effect Confidence Size Interval -.03 -.13, .01 Effect Size -.02 Confidence Interval -.08, .00 Effect Size .01 Confidence Interval -.02, .08 Effect Size -.01 Confidence Interval -.04, .00 .00 -.01, .03 .01 -.01, .08 .04 -.00, .03 .00 -.02, .06 -.01 -.05, .01 -.01 -.08, .02 -.00 -.01, .01 -.00 -.05, .04 Partner Warmth -.01 -.06, .00 -.04 -.13, .02 -.01 -.03, .00 -.03 -.11, .01 Partner Approval -.00 -.02, .02 .00 -.04, .04 -.00 -.01, .01 -.00 -.10, .08 Partner Similarity to Self Partner Similarity to Prototype Partner Competence Note. df = 3, 173. * = confidence interval does not contain zero. Table 16 Indirect effect and confidence intervals of partner femininity by gender identity threat Outcome Variables Mediator Partner Femininity Homophobic Jokes Sexist Jokes Total Defection Effect Size Confidence Interval Effect Size Confidence Interval Effect Size Confidence Interval .02 -.01, .07 -.01 -.07, .03 .00 -.01, .02 Note. df = 3, 173. * = confidence interval does not contain zero. 57 Unfavorable Tasks Assigned Effect Confidence Size Interval -.01 -.09, .01 Table 17 Indirect effect and confidence intervals of partner masculinity by gender identity threat Outcome Variables Mediator Partner Masculinity Homophobic Jokes Effect Size -.00 Confidence Interval -.03, .02 Sexist Jokes Effect Size -.01 Confidence Interval -.08, .02 Total Defection Effect Size -.00 Confidence Interval -.02, .01 Unfavorable Tasks Assigned Effect Confidence Size Interval .00 -.03, .05 Note. df = 3, 167 * = confidence interval does not contain zero. Table 18 Indirect effect and confidence intervals of partner sexuality by gender identity threat Outcome Variables Mediator Partner Sexuality Homophobic Jokes Sexist Jokes Total Defection Unfavorable Tasks Assigned Effect Size Confidence Interval Effect Size Confidence Interval Effect Size Confidence Interval Effect Size Confidence Interval -.00 -.06, .02 -.01 -.10, .03 -.00 -.02, .01 -.00 -.05, .04 Note. df = 3, 149. * = confidence interval does not contain zero. 58 Partner Gender Conformity Threat. I hypothesized that the partner assessment variables would mediate the relationship between partner gender conformity and expressions of harassment and incivility. The mediation pathway coefficients for partner gender conformity by the dependent variables through the multiple partner assessment measures are presented in Table 19. As demonstrated in the table, partner gender conformity was significantly related to all of the multiple mediators (partner similarity to self, partner similarity to the prototype, partner competence, and partner warmth) except one (partner approval). Interestingly, the relationship between partner gender conformity and perceived partner warmth was negative, and positive for partner competence, indicating that the gender nonconforming partner was described as less warm and more competent compared to the gender conforming partner. This is in contrast to the expected relationship, given that more feminine self-descriptions are commonly associated with perceptions of decreased competence and increased warmth. Partner similarity to self, partner warmth, and partner approval were also significantly negatively related to unfavorable task assignment; as perceptions of the partner as similar, warm, and approved increased, unfavorable task allotment decreased. The pathways for the individual mediation analyses by partner gender conformity are similar in that all demonstrate a significant relationship between partner gender conformity and sexist jokes, which is consistent with the findings of Hypothesis 1b. Partner femininity (Table 20), partner masculinity (Table 21), and partner sexuality (Table 22) were also significantly related to partner gender conformity, with partner femininity having a negative relationship with the independent variable. 59 Table 19 Coefficients for mediation pathways of partner gender conformity threat by mediation outcome variables Mediator Homophobic Jokes b t -.12 -.90 t 1.61 b .08 t 1.25 -- -- -- -- -- -- -- -.07 -.97 .02 .16 -.07 -2.25 -.22* -2.09 4.02 -- -- -- -- -- -- -- -- -- -- -.02 -.21 -.19 .19 -.08 -1.58 -.06 -.37 .38* 6.92 -- -- -- -- -- -- -- -- -- -- .08 .66 -.02 -.11 -.01 -.09 .05 .27 -.77* -11.34 -- -- -- -- -- -- -- -- Partner Warmth -- -- -.14 -1.39 -.16 -.90 -.02 -.38 -.30* -1.94 Partner Gender Conformity Partner Approval .11 2.03 -- -- -- -- -- -- -- -- Partner Approval -- -- -.01 .07 .03 .25 -.04 -1.25 -.35* -3.28 Partner Gender Conformity Partner Similarity to Self Partner Similarity to Self Partner Gender Conformity Partner Similarity to Prototype Partner Similarity to Prototype Partner Gender Conformity Partner Competence Partner Competence Partner Gender Conformity  Partner Warmth t -- .23* 2.03 -- -- -- .25* Unfavorable Tasks b t -.06 -.30 b .38 Partner Gender Conformity b -- Dependent Variables Sexist Jokes Total Defection Note. df = 3, 187. *p < .05. 60 Table 20 Coefficients for mediation pathways of partner gender conformity threat by partner femininity Dependent Variables Mediator Homophobic Sexist Jokes Total Defection Jokes b t b t b t b t Unfavorable Tasks b t Partner Gender Conformity -- -- .11 .91 .57* 2.80 .04 .66 .00 .03 Partner Gender Conformity  Partner Femininity -1.20* -9.66 -- -- -- -- -- -- -- -- -- -- .10 1.76 .12 1.21 .01 .47 -.06 -.74 Partner Femininity Note. df = 4, 172. *p < .05. Table 21 Coefficients for mediation pathways of partner gender conformity threat by partner masculinity Mediator b t Homophobic Jokes b t Partner Gender Conformity -- -- -.01 Partner Gender Conformity  Partner Masculinity 1.72* 11.52 -- -- Partner Masculinity Dependent Variables Sexist Jokes Total Defection Unfavorable Tasks b t b t b t -.07 .76* 3.40 .03 .43 .15 .76 -- -- -- -- -- -- -- -- .00 .01 -.16 -1.82 .00 .08 -.03 -.46 Note. df = 4, 166. *p < .05. 61 Table 22 Coefficients for mediation pathways of partner gender conformity threat by partner sexuality Dependent Variables Mediator b t Homophobic Jokes b t Partner Gender Conformity -- -- .03 Partner Gender Conformity  Partner Sexuality .97* 5.91 -- -- Partner Sexuality Sexist Jokes Total Defection Unfavorable Tasks b t b t b t .28 .65* 3.34 .05 .87 .06 .34 -- -- -- -- -- -- -- -- -.01 -.10 -.07 -.82 -.00 -.08 .00 .03 Note. df = 4, 170. *p < .0 62 Though there were many significant pathway coefficients between partner gender conformity and the mediators, the mediation output indicated that there were only a few significant mediation effects. Specifically, only perceived partner similarity to self (Table 23) was found to significantly mediate responses towards the interaction partner in the gender conformity condition. For total defection within the Prisoner’s Dilemma paradigm and unfavorable task assignment, perceived partner similarity to oneself mediated outcomes by partner gender conformity. Coefficients for the mediation paths were negative which indicated that when the partner was perceived as similar to themselves, participants were both less likely to defect in the Prisoner’s Dilemma and less likely to assign unfavorable tasks to the partner. The remaining assessment variables, including partner femininity (Table 24), partner masculinity (Table 25), and partner sexuality (Table 26), did not significantly mediate the relationship between partner gender conformity and the harassment and incivility measures. Thus, Hypothesis 2 was only partially supported. The only significant mediation of harassment and incivility responses were by perceived similarity of the partner to oneself on the relationship between partner gender conformity and likelihood to defect and assign unfavorable tasks to the interaction partner. All other mediators, in both threat conditions, did not significantly mediate the relationship between the independent and dependent variables. 63 Table 23 Indirect effect and confidence intervals of mediation analyses outcome variables by partner gender conformity threat Outcome Variables Mediator Homophobic Jokes Sexist Jokes Total Defection Unfavorable Tasks Assigned Effect Confidence Size Interval -.05* -.17, -0.00 Effect Size -.02 Confidence Interval -.08, .01 Effect Size -.00 Confidence Interval -.05, .08 Effect Size -.02* Confidence Interval -.05, -.001 Partner Similarity to Prototype Partner Competence -.01 -.07, .05 -.05 -.16, .04 -.02 -.05, .00 -.02 -.12, .07 .03 -.06, .13 -.01 -.18, .16 -.00 -.04, .04 .02 -.12, .16 Partner Warmth .11 -.05, .31 .12 -.15, .39 .01 -.07, .09 .23 -.01, .45 Partner Approval -.00 -.04, .02 .00 -.03, .07 -.00 -.03, .00 -.04 -.15, .02 Partner Similarity to Self Note. df = 3, 187. * = confidence interval does not contain zero. Table 24 Indirect effect and confidence intervals of partner femininity by partner gender conformity threat Outcome Variables Mediator Partner Femininity Homophobic Jokes Effect Size -.13 Confidence Interval -.30, .01 Sexist Jokes Effect Size -.15 Confidence Interval -.41, .07 Note. df = 3, 173. * = confidence interval does not contain zero. 64 Total Defection Effect Size -.02 Confidence Interval -.09, .45 Unfavorable Tasks Assigned Effect Confidence Size Interval .08 -.13, .26 Table 25 Indirect effect and confidence intervals of partner masculinity by partner gender conformity threat Outcome Variables Mediator Partner Masculinity Homophobic Jokes Effect Size .00 Confidence Interval -20, .21 Sexist Jokes Effect Size -.27 Total Defection Confidence Interval -.59, .01 Effect Size .00 Confidence Interval -.08, .08 Unfavorable Tasks Assigned Effect Confidence Size Interval -.06 -.32, .20 Note. df = 3, 167. * = confidence interval does not contain zero. Table 26 Indirect effect and confidence intervals of partner sexuality by partner gender conformity threat Outcome Variables Mediator Partner Sexuality Homophobic Jokes Effect Size -.00 Confidence Interval -.12, .10 Sexist Jokes Effect Size -.07 Confidence Interval -.33, .09 Note. df = 3, 149. * = confidence interval does not contain zero. 65 Total Defection Effect Size -.00 Confidence Interval -.05, .03 Unfavorable Tasks Assigned Effect Confidence Size Interval .00 -.13, .15 Hypothesis 3 Finally, the purpose of Hypothesis 3 was to examine the replication of Funk and Werhun’s (2011) findings that men placed in identity threat conditions performed significantly worse on tests cognitive ability and reported increased salience of their gender identity than men in the control condition, reflecting the presence of gender identity threat. The means and standard deviations are presented in Table 27. Hypothesis 3a. Given constraints of time and difficulties in software application, the measure of self-regulation was not included as part of the model replication. Hypothesis 3b. Hypothesis 3b predicted that men in the identity threat condition would perform significantly worse on a test of cognitive ability (solving anagrams) than men in the control condition. This decrease in performance was hypothesized to reflect the presence of identity salience. To test this hypothesis, an ANCOVA was run using gender identity threat as the independent variable, total number of anagrams correctly solved was the dependent variable, and political affiliation and socioeconomic status growing up were included as covariates. The ANCOVA indicated that there was not a statistically significant difference in anagram completion between the threat conditions, F(1, 197) = .02, MSE = .03, p = .90. Thus, hypothesis 3b was not supported. Hypothesis 3c. Hypothesis 3c predicted that men in the identity threat condition would report more sum total masculine categories and adjectives in the open-ended self-description (“Who Am I”) measure compared to men in the no threat condition, reflecting identity salience. As in the previous hypothesis, an ACNOVA was employed to analyze the relationship between gender identity threat and the salience of masculinity. The ACNOVA indicated that there was not a significant difference in the sum total of masculine categories referenced between the two threat conditions, F(1, 197) = .47, MSE = 7.44, p = .49. 66 Thus, Hypothesis 3 was not supported. Those in the gender identity threat condition did not solve significantly fewer anagrams in the cognitive ability measure, and they were not more likely to self-report masculine categories in the open-ended “Who Am I” task, compared to those in the no gender identity threat condition. Table 27 Means and standard deviations of cognitive and identity salience measures. Variables No Gender Gender Identity Identity Threat Threat M (SD) M (SD) Sum Total Anagrams 9.99 (.38) 9.60 (.43) Completed Gender Identity Salience (“Who Am I”) 1.02 (.13) 67 1.05 (.15) F .47 p .49 .02 .90 Discussion Advances in the legal rights and cultural visibility of homosexuals and gender nonconforming people have yielded increasing awareness of their high rates of harassment. In the current study, I sought to replicate and extend the gender harassment and gender identity threat literature by: 1) examining the impact of gender identity threat and partner gender nonconformity on expressions of harassment and incivility by men in same-sex virtual encounters, 2) identifying important mediators of the relationship between identity and gender conformity threat and harassment behavior within same-sex interactions, and 3) attempting to replicate the findings of Funk and Werhun (2011) regarding the impact of gender identity threat on men’s cognitive functioning and identity salience. My analyses examined how experiencing gender identity threat, induced among participants by a female confederate telling them they “had a grip like a girl,” and being exposed to a gender nonconforming versus conforming interaction partner, were related to harassment and incivility behaviors among men. These analyses also examined the mediating role of several important partner perceptual items including perceived partner similarity, competence, warmth, approval, femininity, masculinity, and sexuality. I also sought to replicate Funk and Werhun’s findings that experiences of gender identity threat negatively impact cognition, as measured by the number of correctly solved anagrams within 5 minutes from a list of 24, and result in heightened salience of the masculine social category, as measured by the sum total number of references to masculinity in an open-ended “Who Am I” task. Thus, my theory links men’s harassment behavior, as measured by traditional sexual harassment measures (Siebler, Sabelus, & Bohner, 2008), towards same-sex others with salience of the masculine category, and posits that same-sex gender-based harassment results, at least in part, as a way for men to affirm their 68 masculinity after experiencing gender identity threat (Funk & Werhun, 2011). Overall, the r esults of my analyses only partially supported my predictions. These results and possible explanations for the presence and absence of significant findings are discussed below. In the first hypothesis, I predicted main effects by gender identity threat and partner gender conformity conditions, as well as an interaction between the two conditions. For the first main effect I predicted that men who had experienced gender identity threat would engage in more gender harassment and incivility, operationalized as sending more sexist and homophobic jokes to the partner, defecting in the Prisoner’s Dilemma paradigm, and assigning undesirable tasks to the partner, compared to men who had not experienced gender identity threat. This hypothesis was predicated on sexual and gender harassment theories that demonstrate that threats to masculine identity increase the likelihood of sexual and gender-based harassment towards both gender deviant women and men (Funk & Werhun, 2011; Maass et al., 2005). This prediction was further supported by Social Identity Theory, which posits that belonging to a group serves as an important part of one’s sense of self and fulfill needs of personal distinctiveness (Brewer, 1991; Ellemers et al., 2002; Tajfel, 1974; Tajfel & Turner, 1986). Threats to a man’s sense of masculinity, which can be induced via negative feedback regarding a their perceived sexuality or gender presentation, as well as their general ability to fulfill the masculine gender role, have been shown to result in heightened salience of their gender identity (Cramer, 1998; Hitlan et al., 2009) and even expressions of out-group discrimination (Ellemers et al., 2002). Thus, if threatening a man’s sense of masculinity induces a sense of gender identity threat, it should correspond to increased expressions of harassment and incivility towards the interaction partner regardless of their gender conformity, compared to men who did not experience gender identity threat. However, the results of my analyses only partially 69 supported this prediction. The only significant difference in the harassment and incivility behavior of men who experienced gender identity threat, compared to those who did not, was in the number of homophobic jokes sent to the interaction partner. Men in the gender identity threat condition were significantly more likely to send homophobic jokes to the interaction partner than men who did not experience gender identity threat. Research has shown that masculine identity threat can result in the need to reaffirm one’s masculinity by harassing and discriminating against women and gender nonconforming men (Herek, 1986; Hitlan et al., 2009; Maass et al., 2003). Maass et al. found that when men who had experienced gender identity threat were more likely to harass the female interaction partner, especially among those men who were highly identified with their gender identity. Further, Funk and Werhun (2011) proposed that gender identity threat threatened male participants self-esteem and gender status, and their subsequent responses were an attempt to restore their injured male and masculine identities. Finally, Glick et al. (2007) demonstrated that men who were provided untrue feedback on their “feminine” personality were more likely to exhibit negative affect when exposed to gender nonconforming gay men than gender conforming gay men, and express homophobic behaviors towards them when provided the opportunity. Thus, the increased likelihood to send homophobic jokes to the partner following experiences of gender identity threat could reflect the desire to affirm their masculine identity. However, it is more difficult to explain the lack of significant difference in the other harassment and incivility measures. It could perhaps be due to the fact that the other measures did not provide sufficient opportunity for the participant to affirm their gender identity. Masculinity affirmation within the context of this study might have been achievable only through derogating the interaction partner, rather than expressing incivility in the Prisoner’s Dilemma or task 70 allotment. This may have been due to the fact that criticism about a man’s sexuality is particularly offensive to masculine norms. Indeed, Herek (1986) argued that to be a “man” in contemporary society is, by definition, to be homophobic. He defined this as a need to be hostile to homosexual people and gay men in particular, and posited that hostility toward gay people serves to enhance masculine identity. Thus, I propose that sending homophobic jokes, rather than sexist jokes, served the very specific psychological purpose of enhancing a threatened masculine identity. A more obvious explanation points to the attitudes and beliefs of the participants. The increase in homophobic jokes sent following gender identity threat could itself simply be reflective of the homophobia of the participants. Those with defensive attitudes, such as those inspired by gender identity threat, manifest greater conformity to what they view as genderappropriate characteristics (Herek, 1986). Herek also found that men who express homophobia due to underlying defensive attitudes are likely being driven by anxieties about personal adequacy in meeting gender-role demands. Within the present study, gender identity threat likely served to ignite fears of masculine gender role adequacy within participants, resulting in expressions of homophobia among men who likely highly valued their gender-role conformity. Theoretically, the increased likelihood to send homophobic jokes following gender identity threat could have been driven by this effect. Within the first hypothesis I also predicted that men who were exposed to a gender nonconforming interaction partner would engage in more gender harassment and incivility towards their partner, compared to men who were assigned a gender conforming partner. I predicted this based on theories of sexual harassment that argue that sexual harassment arises from the desire to police gender nonconformity and perpetuate traditional gender roles (Konik & Cortina, 2008; Maass et al., 2003; Stockdale et al., 1999). Research shows that gender 71 nonconforming or feminist women are subject to greater harassment within the computer harassment paradigm (Siebler, Sabelus, & Bohner, 2008). From this standpoint, I predicted that a gender nonconforming male interaction partner would be the recipient of more sexist and homophobic jokes, increased defection in the prisoner’s dilemma, and be allotted tasks more unfavorably, compared to the gender conforming male interaction partner. Contrary to these predictions, only the number of sexist jokes sent to the interaction partner significantly differed between the conditions. My analyses confirmed that gender nonconforming men were more likely to be sent sexist jokes than their gender conforming counterparts. However, on all other measures of harassment and incivility, there was no significant difference in the treatment of gender conforming and nonconforming interaction partners. In previous implementations of the computer harassment paradigm, only sexist jokes and imagery were used to measure harassment responses towards female targets (Maass et al., 2003; Siebler, Sabelus, & Bohner, 2008). It is possible, due to the highly feminine nature of the adjectives and characteristics used within the gender nonconforming partner’s self-description, the impetus to send sexist jokes could have been driven by a conditioned response to femininity in an interaction partner. However, the findings of Maass et al. demonstrated that it was gender nonconformity in the female partner that motivated the harassment in their analyses employing the computer harassment paradigm. Given this, I argue that it was the partner’s gender nonconformity, and not simply the femininity of their description, that drove the significant difference in the number of sexist jokes sent to the interaction partner. These results potentially provide a new understanding into same-sex responses towards gender nonconforming men. The significant difference in sending sexist jokes could be related to the fact that sexist responses may be the common trait underlying negative responses to gender nonconforming people, 72 regardless of gender. Specifically, expressions of sexism, at least against gender nonconforming women, have been theorized to be rooted in a desire to perpetuate traditional gender roles (Stockdale, Visio, & Batra, 1999). Thus, rather than conceiving of sexist behavior as solely targeted towards women who deviate from prescribed gender roles, men who deviate from masculine gender norms may also be subject to sexist harassment for the same underlying cause; to assert traditional gender role expectations. The sending of sexist jokes to the gender nonconforming interaction partner could serve not only as a backlash against the partner’s perceived gender deviance, but also to affirm and perpetuate traditional gender roles. It is possible that the behaviors participants deemed most harmful to the gender nonconforming partner were those also associated with harassing and denigrating women. One potential explanation for the lack of significant difference in the number of homophobic jokes sent was that the current prescriptive norms discourage homophobia, especially towards individuals who are perceived as sexual minorities. Responses to male gender role transgressions are motivated at least in part by implicit assumptions that the gender deviance is symptomatic of homosexuality (McCreary, 1994). Likewise, participants could have responded cooperatively in the Prisoner’s Dilemma and equitably in the task allotment because they were, in part, driven by a desire to not harass based upon sexuality. Thus, they did not expressed behaviors (sending homophobic jokes) that could have been identified as homophobic. While those participants who identified within their suspicion check that they perceived the study as being about homosexuality were removed, those participants who did not state outright their suspicion remained within the sample. Thus, the participant’s responses could have been motivated by a bias that was not measured directly. The other potential issues underlying the lack of significant difference in the harassment and incivility measures, with the exception of sexist 73 jokes, could have been due to the same reasons discussed in the preceding paragraphs. Social desirability effects could have impacted responses on the other measures due to perceptions of being “caught” violating social norms. Finally, I also hypothesized that, for those men who both experienced gender identity threat and were assigned a gender nonconforming partner, expressions of gender harassment would be the greatest, compared to all other interaction groups. The combined effects of gender identity threat and partner gender nonconformity was posited to result in the strongest negative response to the interaction partner, however the results did not support this hypothesis. The interaction conditions did not significantly differ in their responses to the interaction partner. At least part of this is possibly due to the fact that the gender identity threat manipulation was not sufficiently strong enough to induce significant differences on any of the harassment and incivility measures, besides the number of sexist jokes. Further, more than one participant was removed for suspicion regarding whether their interaction partner was actually real, undermining their likelihood to respond authentically. Another possible reason for the lack of interaction effect, particularly in the presence of significant main effects within the computer harassment paradigm, is that the physiological arousal induced was not increased by double threat any more than single threat. Perhaps once threat emerges, additional threat does not impact participant feelings or behavior. In future analyses, it might be wisest to include a singular threat manipulation, which could result in a stronger overall main effect when not conflicted by the presence of additional threat manipulations. My second hypothesis posited that the relationship between the main effects of identity threat and gender nonconformity on expressions of gender harassment and incivility would be mediated by participant perceptions of the interaction partner, as measured by a series of partner 74 assessment variables. Partner assessment variables included partner similarity to self, similarity to the prototypical student at MSU, partner competence, partner warmth, partner approval, and perceived partner masculinity, femininity, and sexuality. While the main effects of gender identity threat and partner gender nonconformity were in some instances significantly related to two of the outcome measures (homophobic and sexist jokes, respectively), as detailed in the preceding paragraph’s discussion of Hypothesis 1a and 1b, significant mediation pathways were few. Gender identity threat was not significantly related to any of the mediators; however the mediating variables of partner approval and partner warmth were significantly related to the number of sexist jokes sent to the interaction partner, as well as the likelihood of inequitable task allotment. Thus, the relationship between gender identity threat and the harassment and incivility outcome measures was not mediated by the partner assessment variables. The lack of significant findings by gender identity threat likely reflects a lack of connection between gender identity threat and subsequent perceptions of the interaction partner. The lack of mediation by partner assessment, despite the significant relationship between partner approval and partner warmth and the number of sexist jokes sent and tasks allotted, reflects a fundamental disconnect between gender identity threat and perceptions of in-group members. This is contrary to expectations as it was predicted that gender identity threat would result in gender identity salience that would in turn shape perceptions of the interaction partner, regardless of gender conformity/nonconformity of the partner. Thus, perceptions of the partner did not explain why identity threat resulted in more homophobic jokes being sent. Social identity theory supports the argument that responses were motivated by lower self-esteem, engendered by threats to their sense of group identity. 75 Partner gender nonconformity yielded many significant relationships with the mediators, while also sharing a significant relationship with the number of sexist jokes sent to the partner. Partner similarity to self and to the prototypical male student, as well as perceived partner competence, warmth, masculinity, femininity, and sexuality were all significantly related to the independent variable. Further, a few of the mediators were themselves related to the harassment and incivility variables. Partner similarity to self and partner warmth were significantly negatively related to unfavorable task assignment, as was partner approval. However, only a fraction of these significant relationships resulted in significant mediation by the assessment variables of the relationship between partner gender nonconformity and expressions of harassment and incivility. Only perceived partner similarity to self significantly mediated responses to the interaction partner, and demonstrated a negative impact on the likelihood of defection within the Prisoner’s Dilemma and the allocation of unfavorable tasks to the partner. Specifically, participants in the partner gender conformity condition, who perceived their partner as similar to themselves, were both less likely to defect in the Prisoner’s Dilemma and less likely to assign unfavorable tasks to the partner. Thus, my second hypothesis was only partially supported by the finding that partner similarity to self significantly mediated the relationship between partner gender nonconformity and defection and unfavorable task assignment. While almost all of the assessment variables were related to the independent variable, only partner similarity significantly mediated responses. Clearly, perception of the partner as similar to the participant was a stronger determinant for responses than the other measures of partner perception. Sexual and gender harassment of men by heterosexual men is often motivated by the desire to reinforce heterosexist and hypermasculine gender roles (Stockdale, Visio, & Batra, 1999). Unsurprisingly, a partner 76 who was deemed more similar to themselves was less likely to be the target of increased defection and unfavorable task assignment. The importance of perceived similarity in the interaction partner is further supported by self-categorization theory. Self-categorization theory argues that self-categorization depersonalizes attitudes towards fellow in-group members in terms of the in-group prototype (Hogg, Hardie, & Reynolds, 1995). The gender conforming partner was seen as more similar to themselves, reflecting a sense of identification that shielded them from harassment and incivility responses. It is my belief that at least one reason for the lack of significant mediation by the partner assessment variables is that they were not relevant to harassment or incivility intention. Results show that the mean response on the partner assessment variables hovered around the midpoint of the scale, perhaps indicating participants were overall primarily neutral in their perception of the partner. It is possible there are alternative processes at work that I did not include in this study, and that perhaps do not relate to perceptions of the partner. A potential determinant for harassment and incivility expressions towards a gender nonconforming interaction partner is not the individual’s perception of the partner as warm, competent, masculine, feminine, or homosexual, but rather their own more static personality and attitudes. In the majority of studies of sexual and gender harassment that exist, measurements of participant gender identity and centrality (Maass et al., 2003; Funk & Werhun, 2011), Likelihood to Sexually Harass (Lee, Gizzarone, Ashton, 2003; Pina, Gannon, & Saunders, 2009; Pryor, 1987; Pryor, Giedd, & Williams, 1995), and hostile/benevolent sexism (Swim, Mallett, & Strangor, 2004; ) endorsement are included as covariates, predictors, or moderators within the analyses. These measures provide an invaluable insight into the belief systems of participants that often informs their harassment likelihood. It was the original intention of this study to include a series of 77 validated identity and attitudinal measures as covariates or even mediators within the methodology, however given the sources of threat and the behaviors examined it was deemed that the measures were too suspicion inducing to be included in the final analyses. In the absence of these measures I can only speculate as to the potential interrelationships between them and the outcome variables. My final hypothesis predicted that the findings of Funk and Werhun (2011) would replicate such that, following experiences of gender identity threat, participants would correctly solve significantly fewer anagrams and reference a higher number of masculine categories in the open ended “Who Am I” prompt, compared to men who had not experienced gender identity threat. I predicted this based upon the fact that Funk and Werhun have shown that gender identity threat results in increased salience of the masculine category and reductions in mental acuity due cognitive load. This prediction was not supported, and the means between the two groups did not reflect Funk and Werhun’s initial findings. Following gender identity threat, participants did not reflect cognitive load or heightened gender identity salience within the “Who Am I.” Thus, overall I failed to replicate the effects for gender identity threat found by Funk and Werhun. The most obvious explanation for these results is that the manipulation failed to induce a sense of gender identity threat in participants, or that the sense of threat that was induced was not strong enough to result in cognitive load and gender identity salience. The sense of threat was dependent upon the criticism that participants “had a grip like a girl,” and variations in the way this was framed and stated could have resulted in the absence of a sense of threat. Similarly, threat could have been induced, but not at sufficient strength, despite the presence of significant differences in the number of homophobic jokes sent. Had the manipulation been stronger, 78 perhaps there would have been measurable differences on the defection and task allotment measures as well. For one, my research assistants initially reported discomfort being so critical, and required additional training on how to correctly present the threat condition. There is also the fact that students that participated in the experiment were familiar with experimental methodology and the use of deception such that they may not have internalized the feedback as sufficiently threat inducing. Unfortunately time constraints did not allow for pilot studies of the threat manipulation, which might have provided early insight into its limitations. Finally, there is the potential that the measures utilized to detect threat were not sufficient. While the means for the cognitive ability and identity salience were not significant, they are in the direction of the hypothesis. Measuring identity salience is difficult, especially when attempting to measure identity outside of the confines of validated group identity measures such as the TAT (Cramer, 1998) or gender identity measures (Glick et al., 2007; Leaper & Van, 2008; McCreary, 1994). Unlike other identity measures, the “Who Am I” variable is not as validated or consistently employed as an identity salience measure. Limitations and Future Directions One of the biggest limitations of the study was the Funk and Werhun (2011) replication. The methodology of the current study design was heavily influenced by the Funk and Werhun study methodology, including the use of female confederates to induce gender identity threat within the male participants by having them state, to those randomly assigned to the gender identity threat condition, that they “had a grip like a girl.” However, there are many potential issues with the Funk and Werhun findings. Their overall study means were very small, as was the study sample size, and the significant between group findings failed to replicate in the present 79 study. Given the findings of the present analysis and the failure of the model to replicate, I have endeavored to critically examine the Funk and Werhun findings. Arguably, given that the means and sample size were so small, it is highly likely that the results of the Funk and Werhun study were simply noise. Thus, the findings failed to replicate within my study, despite the presence of a larger sample size, because the findings of the Funk and Werhun analyses were noise and not actually representative of the outcomes of identity threat. Similarly, the outcomes for sexist and homophobic jokes in the current study could also be the result of noise. While it is my belief that the significant between group outcomes on the sexist and homophobic jokes were due to the strength of the manipulations, it is necessary to note that they could in fact be the product of noise. A total of 14 between group analyses were run, and only 2 of these analyses resulted in significant outcomes. Further, the power of the present study was below that intended in the initial study proposal due to a substantial loss of participants. The potentiality for these responses being the product of noise is further supported by the fact that the overall participant assessment and responses to the partner on the harassment, cooperation, and incivility measures were positive, despite the presence of gender nonconformity in the interaction partner. The generally positive perception and responses to the gender nonconforming partner are contrary to the hypotheses, but reflect the broader trend within samesex harassment that the majority of men do not suffer same-sex gender based harassment. Despite these arguments and outcomes, I argue that it is necessary to keep exploring the question of same-sex harassment behavior to assess whether there is actually an effect. An additional limitation of the present study is the absence of strong behavioral measures of harassment expressions. The only measure included in the present analysis with extensive history of experimental implementation in sexual and gender based research was the computer 80 harassment paradigm. Perhaps it is not surprising then that this was the sole measure to yield significant differences between the conditions. The current study did however provide new understanding as to the utility of the computer harassment paradigm beyond male-female sexual harassment research. The inclusion of homophobic jokes within the methodology of the harassment paradigm, and the findings that gender identity threat results in an increase in homophobic jokes being sent to the partner, are new and important additions to the harassment literature. However, it is important to note that the content of the homophobic jokes included in the design solely referenced gay men, and the sexist jokes solely referenced women. Given the content of these jokes, it is necessary to note that including homophobic jokes whose content relates to lesbians, and sexist jokes whose content relates to men, could have resulted in different findings. In the future, including homophobic and sexist jokes with a range of content could provide further insight into the mechanisms of gender based harassment. Overall, the absence of significant differences by condition on decisions within the Prisoner’s Dilemma and task allotment assignment speaks to the necessity for validated behavioral measures of harassment proclivity. Similar to the issue of a lack of behavioral measures of harassment behavior, an additional limitation of this study was the use of weak gender identity salience measures. The gender identity salience measures included in this study are not as valid as more traditional measures of gender identity centrality. To attain the best possible representation of the complexity of identity processes, wholly discarding the “Who Am I” would be unwise, but using it in conjunction with stronger measures of identity salience would provide the best insight into how self-categorization is experienced. This is also an issue that could have been addressed with more in-depth pilot studies. Although some of the measures were tested using pilot studies prior 81 to the development of the final study design, the methodology of the current study could have been improved by additional pilot analyses that assessed the effectiveness of the identity salience measures to reflect identity threat. Another substantial limitation is that many participants expressed suspicion not only as to the intention of the study, but the very threat manipulations themselves. This undermined the strength of the manipulations and impaired accurate measurement of responses. One prominent issue was the presentation of the interaction partner and the “sending” and “receiving” of selfdescriptions. Given the limitations of the survey software, it was difficult to create a convincing false virtual interaction partner, especially given the technological acumen of the participants. Had the interaction seemed more realistic the computer harassment paradigm may have been more effective, with significant differences between the conditions arising in both sexist as well as homophobic jokes. The strength of the measure, and its implementation, were impeded by the limitation in the software, and with future implementations I would seek to better reflect the intention underlying the measure as an unobtrusive assessment of sexually harassing behavior (Siebler, Sabelus, & Bohner , 2008). Conclusion In summary, the results of the current study only partially supported the predictions that gender identity threat and partner gender conformity would result in increased harassment and incivility towards the interaction partner. Despite the lack of expected findings on some of the measures, the significant differences found between the number of sexist and homophobic jokes replicates previous studies by theorists within the sexual harassment literature employing the computer harassment paradigm. However, this conclusion is tempered by questions about the strength of the manipulations and the measurements of gender identity salience. Although there 82 were significant findings by the number of homophobic jokes sent to the partner in the gender identity threat condition, and the number of sexist jokes sent in the partner in the partner gender conformity condition, issues still remain in the implementation of social dilemmas and task allotment as measures of incivility and harassment intention. A major strength of the current study was the inclusion of behavioral indicators of gender-based and sexual harassment. Though gender identity threat and partner gender conformity did not result in the predicted differences on most of the harassment and incivility measures, the manipulations were sufficiently strong enough to result in differences within the computer harassment paradigm. Combined with previous research addressing the lack of behavioral and experimental methodology in sexual harassment research, the findings of this study support the importance of utilizing a variety of measures to assess harassment proclivity. 83 APPENDICES 84 APPENDIX A Instructions Given to Participants Thank you for participating in this examination of team work. Our goal is to assess the different ways that people interact when they are placed into pairs. Specifically, we want to know how individuals’ personality, attitudes, and liking of each other influence their ability to complete tasks as a team. You will be randomly assigned an interaction partner from the available pool of students in the same time slot that are currently in our second lab across campus. You and your partner will exchange brief self-descriptions so that you can get to know each other. You will also complete a few short surveys and measures that assess a few final attitudes and beliefs, as well as responses to your partner and your interest in working in pairs. Your partner will also complete these measures, however your responses will not be shown to one another at any time. After you complete the measures and read over your partner’s self-description, you will complete a series of questions about the content of your partner’s description. It is very important to read over the self-description closely and respond to the questions about how you view your partner honestly throughout the study. 85 APPENDIX B Modified Anagram Task Solvable Anagrams (Smith et al., 2006) Given the interests of this study, we would like to see how many anagrams you can complete accurately within a set time limit. Please answer as many as possible, correctly, within 5 minutes. The next few pages contain anagrams. To solve an anagram, rearrange the letters (d s e k) to spell a word (d e s k). Make sure you do not leave any anagram unsolved. Do your very best to avoid leaving any line blank. After five minutes have passed you will be automatically transferred to the next page. 1. Ohugt a. ought b. tough 2. umoth a. mouth 3. rcmie a. crime 4. etile a. elite 5. sartt a. start 6. mealr a. realm 7. lenai a. alien 8. atngi a. giant 9. Srhai a. hairs 10. lnove a. novel 11. roimn 86 a. minor 12. codul a. could b. cloud 13. simuc a. music 14. gicra a. cigar 15. filer a. rifle 16. egtsa a. stage b. gates 17. latys a. Salty 18. hoadr a. hoard 19. Asrug a. sugar 20. Rided a. Dried 21. Fnait a. Faint 22. Rthno a. North b. Thorn 23. ruotc a. Court 87 APPENDIX C “Who Am I/I Am….” Identity Assessment Please complete the following sentence with the best adjective, category, or role that describes you, to the best of your ability: I am _________________ 1. ________________________________________________ 2. ________________________________________________ 3. ________________________________________________ 4. ________________________________________________ 5. ________________________________________________ 6. ________________________________________________ 7. ________________________________________________ 8. ________________________________________________ 9. ________________________________________________ 10. ________________________________________________ 88 APPENDIX D Individual Self-Description Please provide information in the following blanks so that a self-description can be compiled and sent to your partner. After you complete it, you will be sent your partner's information. Please complete the form with as much honesty and information as possible. Username (1) Grade (2) Sex (3) Major (4) College/University (optional) (5) Study Abroad (optional) (6) Overall GPA (optional) (7) Career Objective: Personal Summary: Hobbies: Six things you could never live without: A recent job or internship you've held in the past year: Please be patient while your partner's information is retrieved..... Your self-description has been received! Press "next" to view your partner's self-description. Remember to pay close attention to the information contained as you will be questioned about it later. 89 APPENDIX E Gender Conformity/Nonconformity Stimuli Control Condition Stimuli Username: Sports_Fan Grade: Senior Sex: Male Double Major: Business, Economics College/University (optional): Michigan State University Study Abroad (optional): 1 semester in a law firm in Washington, D.C. Overall GPA: 3.24 Career Objective: I would be interested in employment in a competitive work environment that values drive, ambition, and performance. My ideal job would be in the finance management field, which would provide me with the opportunity to apply my skills to enhance my personal growth and success. Personal Summary: I live by myself, and I value self-reliance, independence, and standing up for my beliefs. I am forceful, reliable and analytical, which are skills I believe would be helpful in my chosen field. In service of my goals, I am willing to take risks, be a leader, and am able to make decisions easily. Hobbies: My hobbies include competitive sports, like football and hockey, and jogging. I can’t cook, but I enjoy eating good food and travelling. I like being active, healthy, and staying fit. Six Things I Could Never Do Without: Exercise, Rock Music, Coffee, Pancakes, Women, and MMA fighting On A Typical Friday Night: I’m working or drinking in a bar surrounded by my friends and watching a game Genesee County Land Bank, Flint MI (10/2012 - Current) Audit Intern Interviewed and hired branch managers, analysts and loans officers Coordinated branch activities Managed a network of 30 local branches including managers, staff, and analysts 90 Gender Nonconformity Stimuli Username: Music_Lover Grade: Senior Sex: Male Double Major: Child Psychology, Education College/University (optional): Michigan State University Study Abroad (optional): 1 semester in a hospital in Jamaica. Overall GPA: 3.24 Career Objective: I would like employment in a work environment that values compassion, honesty, and the well-being of others. My ideal job would be teaching kindergarten, which will provide me with the opportunity to apply my skills to enhance the well-being of others. Personal Summary: I live with my parents, and I dislike conflict. I am generally a happy person, and am sensitive to the needs of others. I am gentle, friendly and I love children, which are skills I believe would be helpful in my chosen field. In service of my goals, I am willing to help others, be cheerful, and be kind and thoughtful to others. Hobbies: My hobbies include gardening, watching movies, and going to see plays and musicals. I enjoy cooking, eating good food and travelling. I like reading, being healthy, and eating vegetarian. Six Things I Could Never Do Without: Making people laugh, Show Tunes, Books, Friends and Family, and Cooking Robert Woods Elementary School, Flint MI (10/2012 - Current) Pre-Kindergarten Teacher Instructed 24 students in a full-day program, in all content areas Developed and implemented pre-academic curriculum focusing on reading, writing, mathematics, science and religion. Combined literature-based learning, with phonics and hands-on activities. 91 APPENDIX F Partner Assessment The following questions regard how you view your partner. The two of you will both be completing these questions, to give us a clear sense of how these measures might help us understand team work. Partner similarity to themselves 1. I think my work partner seems very similar to me. a. 1 (very untrue) to 5 (very true) 2. I would seek out someone like my interaction partner to spend time with in the real world. a. 1 (very untrue) to 5 (very true) 3. I think my interaction partner and I have a lot in common. a. 1 (very untrue) to 5 (very true) 4. I think my interaction partner and I are very different (R) a. 1 (very untrue) to 5 (very true) 5. I can see many differences between myself and my interaction partner (R) a. 1 (very untrue) to 5 (very true) 6. I would not recommend my interaction partner to someone else (R) a. 1 (very untrue) to 5 (very true) Partner similarity to the prototype 7. My interaction partner is very similar to the average MSU student. a. 1 (very untrue) to 5 (very true) 8. My interaction partner would fit in well at MSU. a. 1 (very untrue) to 5 (very true) 9. My interaction partner would be out of place at MSU. (R) a. 1 (very untrue) to 5 (very true) 10. I can’t imagine meeting someone like my interaction partner on an average day at MSU. (R) 92 a. 1 (very untrue) to 5 (very true) 11. I wish there were more people like my interaction partner at MSU a. 1 (very untrue) to 5 (very true) 12. Most MSU students would like my interaction partner. a. 1 (very untrue) to 5 (very true) 93 APPENDIX G Computer Harassment Paradigm Now that you have read over your partner’s description, we would like to provide you with the opportunity to send them five ice breaker jokes. Please select a TOTAL of five jokes from the following list. We have provided as many kinds of jokes as possible, so you can express yourself honestly. Choose whatever jokes you would most like to send, and while you must send five jokes, it is your choice which ones you send. Please choose five jokes to introduce yourself to your partner. Please read over the following twenty five jokes closely and decide on FIVE TOTAL jokes to send, and for all other choices select Sexist jokes 1. What’s the smartest thing to ever come out of a woman’s mouth? Einstein’s penis. 2. What’s the difference between a woman and a computer? Only a computer will accept a three and a half inch floppy 3. What do electric train and women’s breasts have in common? They were both intended for children but it’s the fathers who play with them. 4. How is a woman like a condom? Both of them spend more time in your wallet than on your dick. 5. Why do women rub their eyes when they wake up? Because they don't have balls to scratch. Homophobic jokes 6. What do you call a gay cowboy? A Jolly Rancher 7. What do you call a homosexual dentist? Tooth Fairy 8. How do gay gangsters do a drive by? They throw skittles at you and say “Taste the rainbow, bitches!” 9. What do you get when you cross a gay man and a horse? A Unicorn. 10. What does a gay horse eat? HAAAAAYYYYYY. ‘Normal’ jokes 11. What part of a man’s body should be never moved when he is dancing with a woman? His bowels. 12. What’s worse than spiders on your piano? Crabs on your organ! 94 13. Why are dogs better than kids? When you get sick of your dog, you can have it put to sleep. 14. Did you hear about the little kid whose mother caught him jerking off in the bathroom? She told him to stop because he’d go blind and he asked her if he could keep going till he needed glasses. 15. What happened when two lepers played poker? One threw his hand in, the other laughed his head off. 16. What do you get when you play a country song backwards? You get your car back, your job back, and you stop drinking. 17. What does an accountant do when he’s constipated? He works it out with a pencil. 18. What’s the difference between an oral thermometer and a rectal thermometer? The taste. 19. What's green and eats nuts? Syphilis. 20. What does a nosey pepper do? Get jalapeño business. 21. How do you kill a circus? Go for the juggler... 22. Nurse: "Doctor, there's an invisible man in the waiting room." Doctor: "Tell him I can't see him." 23. Have you heard about the corduroy pillow? I hear it's making headlines 24. I went to the zoo the other day but they only had one dog. It was a shih tzu. 25. What do you call an alligator in a vest? An investigator. Please be patient while your jokes are sent to your partner..... 95 APPENDIX H Social Dilemma (derived from http://serendip.brynmawr.edu/playground/pd.html) A fiendish cyberspace wizard has locked you and your partner into a diabolical game with the following rules: On each turn of the game, you and your partner must choose, without knowing the other's choice, between cooperating with each other and trying to take advantage of each other. Following every turn of the game, you and your partner will each receive a certain number of gold coins, the number depending for each of you on the choices made by both of you. Should you both decide to cooperate, you will each receive 3 gold coins. If one of you has decided to cooperate but the other has opted for competition, the successful competitor will receive 5 gold coins and the unsuccessful cooperator none. Finally, if you both decide to outdo one another, each of you will receive a single gold coin. The game will continue until the wizard tires of it, and includes this additional stipulation: Your chances of surviving to play the next turn are closely related to the average number of coins you have received on previous turns. If this average on a given turn drops below a critical number (chosen in an unknown way by the wizard), some unknown but Inconceivably Foul Fate will befall you. The same is true for your partner, of course, but since neither of you knows the critical number, neither of you has any choice but to try on each and every turn to maximize either yours and your partner’s outcome, or your own income. The game is afoot. What should you do? Ready to Play? Your partner has already chosen a first move. Please make your move: Cooperate, or Compete 96 Participants will complete a total of 5 rounds, though they will be unaware of the number of rounds. If they cooperate + cooperate: for each cooperation they will gain three coins and the average for each participant in each round will remain at three. Cooperation: After the last round, the score is: You have 3 gold coins, and Your average number of coins is 3.00 . Your partner has 3 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 1. Keep going. Please make your next move: Cooperation: After the last round, the score is: You have 6 gold coins, and Your average number of coins is 3.00 . Your partner has 6 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 2. Keep going. Please make your next move. Cooperation: After the last round, the score is: You have 9 gold coins, and Your average number of coins is 3.00 . Your partner has 9 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 3. Keep going. 97 Cooperation: After the last round, the score is: You have 12 gold coins, and Your average number of coins is 3.00 . Your partner has 12 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 4. Keep going. Cooperation: After the last round, the score is: You have 15 gold coins, and Your average number of coins is 3.00 . Your partner has 15 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 5. The wizard has called an end to the task! You and your partner have both achieved 15 gold coins! Congratulations for achieving such a high score! If they cooperate + defect: for the first defection they will gain five coins (as the partner will cooperate). However, each successive round they defect will result in only a gain of one coin for each party and a decrease in the average number of coins for each round. Cooperation: After the last round, the score is: You have 3 gold coins, and Your average number of coins is 3.00 . Your partner has 3 gold coins. Your partner’s average number of coins is 3.00 . The number of rounds so far is 1. Keep going. Defection: After the last round, the score is: You have 8gold coins, and Your average number of coins is 4.00 Your partner has 3 gold coins. Your partner’s average number of coins is 1.50 . The number of rounds so far is 2. Keep going. 98 Defection: After the last round, the score is: You have 9gold coins, and Your average number of coins is 3.00 Your partner has 4 gold coins. Your partner’s average number of coins is 1.33 The number of rounds so far is 3. Keep going. Defection: After the last round, the score is: You have 10 gold coins, and Your average number of coins is 2.50 Your partner has 5 gold coins. Your partner’s average number of coins is 1.25 The number of rounds so far is 4. Keep going. Defection: After the last round, the score is: You have 11 gold coins, and Your average number of coins is 2.20 Your partner has 6 gold coins. Your partner’s average number of coins is 1.20 The number of rounds so far is 5. The wizard has called an end to the task! You have earned 11 gold coins! Congratulations for achieving such a high score! 99 If they defect + defect: for each defection after the first round the partner will respond with defection, yielding one coin for both each round, and a decrease in the average each successive round. Defection: After the last round, the score is: You have 5 gold coins, and Your average number of coins is 5.00 Your partner has 0 gold coins. Your partner’s average number of coins is 0.00 . The number of rounds so far is 1. Keep going. Defection: After the last round, the score is: You have 6 gold coins, and Your average number of coins is 3.00 . Your partner has 1 gold coins. Your partner’s average number of coins is 0.50 . The number of rounds so far is 2. Keep going. Defection: After the last round, the score is: You have 7 gold coins, and Your average number of coins is 2.33 . Your partner has 2 gold coins. Your partner’s average number of coins is 0.67 The number of rounds so far is 3. Keep going. Defection: After the last round, the score is: You have 8 gold coins, and Your average number of coins is 2.00 . Your partner has 3 gold coins. Your partner’s average number of coins is 0.75 . The number of rounds so far is 4. Keep going. 100 Defection: After the last round, the score is: You have 9 gold coins, and Your average number of coins is 1.80. Your partner has 4 gold coins. Your partner’s average number of coins is 0.80 . The number of rounds so far is 5. The wizard has called an end to the task! You have earned 9 gold coins! Congratulations for achieving such a high score! 101 APPENDIX I Task Allocation In the following section, you will be assigned as team leader and divide up the following online tasks to be completed with your partner at a later date. Your partner will not be aware of who designated the tasks, they will only see their assignments. Think carefully about which tasks you would like to complete, and those tasks not assigned to your partner you will have to complete. You MUST assign FIVE tasks to your partner and FIVE tasks to yourself. All tasks are to be completed online at a later date. Task # Task Description 1 Determine the relationships between individuals in a series of photographs. Product search: search the internet for the given products and arrive at an informed conclusion about their quality Find the corporate Twitter account for a company 2 3 4 5 6 7 Assign to Your Partner Find the basic information about a series of universities and colleges based on their school name, including its logo, location, whether it has online course, and then write a short overview. Given a url, find the company name and address Review 4 articles of no more than 300 words: edit a list title, article and look for errors in punctuation, spelling, verb tense and mechanics. Copy text from business card into a series of fields 9 Receipt data entry; enter the company, items, and prices from an image of a receipt. Complete an invoice of the company’s costs. Complete 5 accounting problems 10 Typing up a dictated letter 8 Assign to Yourself Note. The top five tasks were those listed as the most desirable, the bottom five tasks were those listed as least desirable. Their presentation will be randomized in the study. 102 APPENDIX J Computer Harassment Paradigm Review This study has been about teamwork and about responses to working pairs and trios. Given this, we would like to know how you would feel about continuing to the final stage of the study with your current partner. Please consider this seriously, whether or not the current partner you have is one you would like to interact with face-to-face in the lab. Now that you have thought about this critically, please answer the following question: 1. Would you like to continue to the fourth stage of the study with your current partner? 2. Would you like to be randomly assigned a new partner? Lastly, we would like to assess how funny and offensive you found the jokes we provided. Please respond honestly when prompted on the following questions: Please rate the following jokes in the extent to which you think they are FUNNY or OFFENSIVE to you, personally. Please answer honestly as we are still developing our jokes, and any feedback is welcome. Perceived Offensiveness and ‘Funniness’ of the Jokes Sexist jokes 1. What’s the smartest thing to ever come out of a woman’s mouth? Einstein’s penis. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 2. What’s the difference between a woman and a computer? Only a computer will accept a three and a half inch floppy ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive 103 ● 4 - very offensive 3. What do electric train and women’s breasts have in common? They were both intended for children but it’s the fathers who play with them. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 4. How is a woman like a condom? Both of them spend more time in your wallet than on your dick. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 5. Why do women rub their eyes when they wake up? Because they don't have balls to scratch. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive Homophobic jokes 6. What do you call a gay cowboy? A Jolly Rancher ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? 104 ● 1- not offensive ● 4 - very offensive 7. What do you call a homosexual dentist? Tooth Fairy ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 8. How do gay gangsters do a drive by? They throw skittles at you and say “Taste the rainbow, bitches!” ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 9. What do you get when you cross a gay man and a horse? A Unicorn. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 10. What does a gay horse eat? HAAAAAYYYYYY. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 105 ‘Normal’ jokes 11. What part of a man’s body should be never moved when he is dancing with a woman? His bowels. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 12. What’s worse than spiders on your piano? Crabs on your organ! ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 13. Why are dogs better than kids? When you get sick of your dog, you can have it put to sleep. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 14. Did you hear about the little kid whose mother caught him jerking off in the bathroom? She told him to stop because he’d go blind and he asked her if he could keep going till he needed glasses. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? 106 ● 1- not offensive ● 4 - very offensive 15. What happened when two lepers played poker? One threw his hand in, the other laughed his head off. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 16. What do you get when you play a country song backwards? You get your car back, your job back, and you stop drinking. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 17. What does an accountant do when he’s constipated? He works it out with a pencil. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? ● 1- not offensive ● 4 - very offensive 18. What’s the difference between an oral thermometer and a rectal thermometer? The taste. ▪ How funny is this joke? ● 1-not funny ● 4-really funny ▪ How offensive is this joke? 107 ● 1- not offensive ● 4 - very offensive 19. What's green and eats nuts? Syphilis. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 20. What does a nosey pepper do? Get jalapeño business. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 21. How do you kill a circus? Go for the juggler... i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 22. Nurse: "Doctor, there's an invisible man in the waiting room." Doctor: "Tell him I can't see him." i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 108 23. Have you heard about the corduroy pillow? I hear it's making headlines i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 24. I went to the zoo the other day but they only had one dog. It was a shih tzu. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 25. What do you call an alligator in a vest? An investigator. i. How funny is this joke? 1. 1-not funny 2. 4-really funny ii. How offensive is this joke? 1. 1- not offensive 2. 4 - very offensive 109 APPENDIX K Demographic Information We are interested in some information about who you are. Please answer the following questions. 1. What is your gender? _____ Female _____ Male 2. What is your age? __________ 3. What year are you in school? _____ 1st year _____ 2nd year _____ 3rd year _____ 4th year _____ 5th year _____ 6th year or higher 4. Were you born in the U.S.? _____ Yes _____ No a. If no, where were you born? _____________________________ b. At what age did you come to the U.S.? ________years of age 5. What is your racial group? _____ White / Caucasian _____ Black / African American _____ Asian or Pacific Islander _____ Hispanic / Latino / Latina _____ Native American / American Indian _____ Multiracial / Multiethnic (please describe) ________________________________ _____ Other (please describe) ______________________________________ 6. How important would you say your religion is in your life (circle one)? 1 2 3 4 5 6 7 Not at all Extremely Important Important 7. How would you describe your political views (circle one)? 1 2 3 4 5 6 7 Very Very Liberal Conservative 8. What is your sexual orientation? _____ Heterosexual _____ Gay / Lesbian 110 _____ Bisexual _____ Questioning (not sure) _____ Other (describe) __________________________________ 9. Are you currently employed? _____ Yes _____ No 10. How would you describe your economic resources? When you were growing up: Currently: _____ Very poor, not enough to get by _____ Very poor, not enough to get by _____ Barely enough to get by _____ Barely enough to get by _____ Had enough to get by but no extras _____ Had enough to get by but no extras _____ Had more than enough to get by _____ Had more than enough to get by _____ Well to do _____ Well to do _____ Extremely well to do _____ Extremely well to do 11. How feminine did your partner seem? a. 1 (not at all feminine) b. 5 (very feminine) 12. How masculine did your partner seem? a. 1 (not at all masculine) b. 5 (very masculine) 13. What do you think your partner's sexual orientation was? a. Completely homosexual b. Moderately Homosexual c. Neither Homosexual nor Heterosexual d. Moderately Heterosexual e. Completely Heterosexual 111 REFERENCES 112 REFERENCES Akrami, N., & Ekehammar, B. (2006). Right-wing authoritarianism and social dominance orientation: Their roots in big-five personality factors and facets. Journal of Individual Differences, 27(3), 1–10. Altemeyer, R. (1981). Right-wing authoritarianism. Winnipeg, Ontario, Canada: University of Manitoba Press. Ando, K. (1999). Social identification and a solution to social dilemmas. Asian Journal of Social Psychology, 2, 227 - 235. Bahns, A.J. & Branscombe, N.R. (2011). Effects of legitimizing discrimination against homosexuals on gay bashing. European Journal of Social Psychology, 41, 388 - 396. Begany, J. J., & Milburn, M. A. (2002). Psychological predictors of sexual harassment: Authoritarianism, hostile sexism, and rape myths. Psychology of Men & Masculinity, 3(2), 119–126. Bern, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting and Clinical Psychology, 42, 155-162. Bem, S.L. & Lenney, E. (1976). Sex typing and the avoidance of cross-sex behavior. Journal of Personality and Social Psychology, 33(1), 48 - 54. Berdahl, J. L. (2007a). Harassment based on sex: Protecting social status in the context of gender hierarchy. Academy of Management Review, 32(2), 641–658. Berdahl, J. L. (2007b). The sexual harassment of uppity women. The Journal of Applied Psychology, 92(2), 425–37. Buchanan, N. T., Settles, I. H., Woods, K. C. (2008). Comparing sexual harassment subtypes among Black and White women by military rank: Double jeopardy, the jezebel, and the cult of true womanhood. Psychology of Women Quarterly, 32,347 - 361. Cleveland, J., & Kerst, M. (1993). Sexual harassment and perceptions of power: An underarticulated relationship. Journal of Vocational Behavior, 42, 49 – 67. Cortina, L. M., & Wasti, S. A. (2005). Profiles in coping: responses to sexual harassment across persons, organizations, and cultures. The Journal of Applied Psychology, 90(1), 182–92. Cramer, P. (1998). Threat to gender representation: Identity and identification. Journal of Personality, 66(3), 332 - 364. Cribbs, S. E., & Austin, D. M. (2010). Enduring pictures in our heads: The continuance of authoritarianism and racial stereotyping. Journal of Black Studies, 42(3), 334–359. 113 DeSouza, E., & Solberg, J. (2004). Women’s and men's reactions to man-to-man sexual harassment: Does the sexual orientation of the victim matter? Sex Roles, 50(9/10), 623–639. DuBois, C., Knapp, D., Faley, R., & Kustis, G. (1998). An empirical examination of same-and other-gender sexual harassment in the workplace. Sex Roles, 39(9/10), 731 - 749. Ellemer,N., Spears, R., Doosje, B. (2002). Self and social identity. Annual Review of Psychology, 53, 161 - 187. Fineran, S. (2002). Sexual harassment between same-sex peers: Intersection of mental health, homophobia, and sexual violence in schools. Social Work, 47(1), 65 - 74. Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82(6), 878–902. Fitzgerald, L F. (1993). Sexual harassment. Violence against women in the workplace. The American psychologist, 48(10), 1070–6. Fitzgerald, L. F., Gelfand, M. J., & Drasgow, F. (1995). Measuring sexual harassment: Theoretical and psychometric advances. Basic and Applied Social Psychology, 17(4), 425– 445. Funk, L. C., & Werhun, C. D. (2011). “You’re such a girl!”: The psychological drain of the gender-role harassment of men. Sex Roles, 65(1-2), 13–22. Gerrity, D. A. (2000). Male university employees’ experiences of sexual harassment-related behaviors. Psychology of Men & Masculinity, 1(2), 140–151. Glick, P., Gangl, C., Gibb, S., Klumpner, S., & Weinberg, E. (2007). Defensive reactions to masculinity threat: More negative affect toward effeminate (but not masculine) gay men. Sex Roles, 57, 55 - 59. Grochin, G. A., & Kleiner, B. H. (1998). Same-sex sexual harassment. Equal Opportunities International, 17(3 - 5), 17 - 20. Hall, G.C.N. & Hirschman, R. (1994). The relationship between men’s sexual aggression inside and outside the laboratory. Journal of Consulting and Clinical Psychology, 62(2), 375 380. Hamburger, M., & Hogben, M. (1996). Assessing hypergender ideologies: Development and initial validation of a gender-neutral measure of adherence to extreme gender-role beliefs. Journal of Research in Personality, 30, 157–178. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York: The Guilford Press. 114 Herek, G. (1986). On heterosexual masculinity: Some psychical consequences of the social construction of gender and sexuality. American Behavioral Scientist. Hitlan, R., Pryor, J., Hesson-McInnis, M., & Olson, M. (2009). Antecedents of gender harassment: An analysis of person and situation factors. Sex roles, 61, 794 - 807. Kearney, L. K., Rochlen, A. B., & King, E. B. (2004). Male gender role conflict, sexual harassment tolerance, and the efficacy of a psychoeducative training program. Psychology of Men & Masculinity, 5(1), 72–82. Knapp, D., & Kustis, G. (2000). Same-sex sexual harassment: A legal assessment with implications for organizational policy. Employee Responsibilities and Rights Journal, 12(3). Kollock, P. (1998). Social dilemmas: The anatomy of cooperation. Annual Review of Sociology, 24, 183 - 214. Konik, J., & Cortina, L. M. (2008). Policing gender at work: Intersections of harassment based on sex and sexuality. Social Justice Research, 21(3), 313–337. Leaper, C., & Van, S. R. (2008). Masculinity ideology, covert sexism, and perceived gender typicality in relation to young men’s academic motivation and choices in college. Psychology of Men & Masculinity, 9(3), 139–153 Leskinen, E., Cortina, L. M., & Kabat, D. B. (2011). Gender harassment: broadening our understanding of sex-based harassment at work. Law and human behavior, 35(1), 25–39. Levant, R., & Richmond, K. (2007). A review of research on masculinity ideologies using the male role norms inventory. The Journal of Men’s Studies, 15(2), 130–146. Liebrand, W.B.G. (1983). A classification of social dilemma games. Simulation and Gaming, 14, 123 - 138. Lucero, M. A., Allen, R. E., & Middleton, K. L. (2006). Sexual Harassers: Behaviors, Motives, and Change Over Time. Sex Roles, 55(5-6), 331–343. doi:10.1007/s11199-006-9087-y Maass, A., Cadinu, M., Guarnieri, G., & Grasselli, A. (2003). Sexual harassment under social identity threat: the computer harassment paradigm. Journal of personality and social psychology, 85(5), 853–70. McMaster, L. E., Connolly, J., Pepler, D., & Craig, W. M. (2002). Peer to peer sexual harassment in early adolescence: a developmental perspective. Development and psychopathology, 14(1), 91–105. Mitchell, D., Hirschman, R., Angelone, D.J., & Lilly, R.S. (2004). A laboratory analogue for the study of peer sexual harassment. Psychology of Women Quarterly, 28, 194 - 206. 115 Pina, A., Gannon, T.A., & Saunders, B. (2009). An overview of the literature on sexual harassment: Perpetrator, theory, and treatment issues. Aggression and Violent Behavior, 14, 126 - 138. Poppe, M. (2005). The specificity of social dilemma situations. Journal of Economic Psychology, 26, 431 - 441. Pryor, J. B. (1987). Sexual harassment proclivities in men. Sex Roles, 17(5-6), 269 – 290. Pryor, J. B., Giedd, J. L., & Williams, K. B. (1995). A social psychological model for predicting sexual harassment. Journal of Social Issues, 51(1), 69–84. Rospenda, K.M., Richman, J.A., & Nawyn, S.J. (1998). Doing power: The confluence of gender, race, and class in contrapower sexual harassment. Gender & Society, 12, 40 - 60. Ruggieri, S., Bendixen, M., Gabriel, U., & Alsaker, F. (2013). Cyberball. Swiss Journal of Psychology, 72(2), 103–109. Runtz, M. G., & O’Donnell, C. W. (2003). Students’ perceptions of sexual harassment: Is it harassment only if the offender is a man and the victim is a woman? Journal of Applied Social Psychology, 33(5), 963–982. Settles, I. H., Buchanan, N. T., & Colar, B. K. (2012). The impact of race and rank on the sexual harassment of Black and White men in the U.S. military. Psychology of Men & Masculinity, 13(3), 256–263. Siebler, F., Sabelus, S., & Bohner, G. (2008). A refined computer harassment paradigm: Validation, and test of hypotheses about target characteristics. Psychology of Women Quarterly, 32, 22 - 35. Silverschanz, P., Cortina, L. M., Konik, J., & Magley, V. J. (2007). Slurs, Snubs, and Queer Jokes: Incidence and Impact of Heterosexist Harassment in Academia. Sex Roles, 58(3-4), 179–191. Simpson, B. (2006). Social identity and cooperation in social dilemmas. Rationality and Society, 18(4), 443 - 470. Stockdale, M. S., Gandolfo Berry, C., Schneider, R. W., & Cao, F. (2004). Perceptions of the sexual harassment of men. Psychology of Men & Masculinity, 5(2), 158–167. Stockdale, M. S., Visio, M., & Batra, L. (1999). The sexual harassment of men: Evidence for a broader theory of sexual harassment and sex discrimination. Psychology, Public Policy, and Law, 5(3), 630–664. Street, A. E., Gradus, J. L., Stafford, J., & Kelly, K. (2007). Gender differences in experiences of sexual harassment: Data from a male-dominated environment. Journal of consulting and clinical psychology, 75(3), 464–74. 116 Tajfel, H. (1974). Social identity and intergroup behavior. Social Science Information, 13, 65-93 Tajfel, H., Billig, M.G., Bundy, R.P., & Flament, C. (1978). Social categorization and intergroup behavior. European Journal of Social Psychology, 1(2), 149 - 178. Tajfel, H., & Turner, J.C. (1986). The social identity theory of intergroup behavior. Psychology of Intergroup Relations, Chapter 3: 7 - 24. Uggen, C., & Blackstone, A. (2004). Sexual harassment as a gendered expression of power. American Sociological Review, 69(1), 64–92. U.S., Merit Systems Protection Board. (1981). Sexual harassment in the federal workplace: Is it a problem? Washington, DC: U.S. Government Printing Office. U.S., Merit Systems Protection Board. (1988). Sexual harassment in the federal workplace: An update. Washington, DC: U.S. Government Printing Office. U.S., Merit Systems Protection Board. (1995). Sexual harassment in the federal workplace: Trends, progress, and continuing challenges. Washington, DC: U.S. Government Printing Office. U.S., Merit Systems Protection Board. (2004). Issues of merit. Washington, DC: U.S. Government Printing Office. Waldo, C. R., Berdahl, J. L., & Fitzgerald, L. F. (1998). Are men sexually harassed? If so, by whom? Law and human behavior, 22(1), 59–79. Willness, C. R., Steel, P., & Lee, K. (2007). a Meta-Analysis of the antecedents and consequences of workplace sexual harassment. Personnel Psychology, 60(1), 127–16. 117