Gender Nonconformity as a Target of Prejudice, Discrimination, and Violence Against LGB Individuals

Abstract

Research into antigay violence has been limited by a lack of attention to issues of gender presentation. Understanding gender nonconformity is important for addressing antigay prejudice and hate crimes. We assessed experiences of gender-nonconformity-related prejudice among 396 Black, Latino, and White lesbian, gay, and bisexual individuals recruited from diverse community venues in New York City. We assessed the prevalence and contexts of prejudice-related life events and everyday discrimination using both quantitative and qualitative approaches. Gender nonconformity had precipitated major prejudice events for 9% of the respondents and discrimination instances for 19%. Women were more likely than men to report gender-nonconformity-related discrimination but there were no differences by other demographic characteristics. In analysis of events narratives, we show that gender nonconformity prejudice is often intertwined with antigay prejudice. Our results demonstrate that both constructs should be included when addressing prejudice and hate crimes targeting lesbian, gay, bisexual, and transgender individuals and communities.

Researchers, community organizers, and law enforcement officials who are concerned with antilesbian, -gay, and -bisexual (LGB) prejudice and discrimination tend to conflate aspects of anti-LGB prejudice and discrimination that target sexual orientation with other aspects that target gender nonconformity. Because sexual orientation is the explicit target of anti-LGB prejudice, there can be a lack of attention to issues of gender nonconformity, gender norms, and gender presentations in anti-LGB prejudice and discrimination. Gender nonconformity is typically defined as some variation on the following:

Gender expression (or outward appearance) [that] does not follow traditional gender roles: “feminine boys,” “masculine girls,” and students who are androgynous, for example. It can also include students who look the way boys and girls are expected to look but participate in activities that are gender nonconforming, like a boy who does ballet. (Gay–Straight Alliance Network, 2004, p. 1)

The history and experiences of sexual minorities are tightly bound up with gender nonconformity and gender deviant behavior, and sexual minority subcultures in the 20th and 21st centuries have been both subverting and reifying normative genders (Chauncey, 1994;Foucault, 1978/1990; Garber, 1992; Meyerowitz, 2002; Nestle, 1992). From the huge drag balls in the Bowery, Village, and Harlem neighborhoods of New York City in the 1920s that attracted thousands of participants and spectators (Chauncey, 1994) to the butch drag artists of New York’s Greenwich Village in the 1940’s (Davis, 1992) to more contemporary gender-blurring performers such as modern day drag kings and queens (Halberstam & Volcano, 1999), gender nonconformity has long been a marker of homosexuality that has been both celebrated and stigmatized within sexual minority communities.

Simultaneously, gender nonconformity has been at the root of some forms of antigay discrimination and violence across multiple eras. Throughout the 20th century, gender deviance was, at times, the proximal motivation for discriminatory treatment, violent assault, or police brutality (Bulkin, 1992; Chauncey, 1994; Feinberg, 1993; Rivera, 2002). For example, in the early decades of the 20th century, “fairies and other homosexuals were widely recognized as social types in the streets of working-class neighborhoods” of New York, and, as such, “they were also regarded as easy marks by the gangs of youth who controlled much of the traffic on those streets” (Chauncey, 1994, p. 59). Same-sex desire and partnering is a characteristic that may be concealed in a process known as passing (Goffman, 1963), a process with particular mental health and other ramifications (Pachankis, 2007). Gender nonconformity differs in that it is typically a visible stigma, and thus a readily identifiable target for discrimination (Seidman, 2002). In addition, although homosexual behavior is gradually becoming more socially acceptable in certain forms (for example, in 2003 the U.S. Supreme Court ended the criminalization of sodomy), gender transgressive appearance or behavior remains widely stigmatized and targeted for violence (Bornstein, 1998; Lombardi, Wilchins, & Priesing, 2001; Wilchins, 2004).

In this article, we describe antigay prejudice related to perceived gender nonconformity. Using a mixed methodology combining a quantitative analysis of subgroup differences regarding gender-nonconformity-related discrimination and a qualitative analysis of respondent narratives, we show that gender nonconformity can play an important role in antigay prejudice and discrimination, violence, and hate crimes. We highlight two key issues: (a) the range of contexts, or social circumstances, in which gender-nonconformity-related prejudice occurs, and (b) the relationship between sexual orientation and gender presentation in regards to prejudice.

Gender Nonconformity in Research

Gender has long been an interest of those in LGB research, and gender nonconformity has often been embedded in work on prejudice targeting LGB populations (Balsam, Rothblum, & Beauchaine, 2005; Comstock, 1991; Franklin, 1998; Herek & Berrill, 1992). Yet assessment of the role of gender nonconformity in instances of prejudice has been more elusive. For example, Comstock (1991) discussedthe importance of sex-role stereotyping in motivating perpetrators of antigay violence and in determining the social position of gay and lesbian individuals, but in his survey of antigay/lesbian violence, he did not identify the ways in which sex-role violation or gender nonconformity played a part in antigay events. Herek, Gillis, and Cogan (1999) found that “hate crime victimization was higher for individuals reporting more openness about their sexual orientation” (p. 950)—a finding that could indicate that the participant is more visibly gender nonconforming, which is often equated with degree of openness or outness. Without disentangling gender nonconformity from sexuality here, we miss crucial information about the processes involved in prejudice and hate crimes.

A few approaches to gender nonconformity have focused on heterosexuals’ attitudes towards gay men and lesbians, based on whether or not they are gender nonconforming. Simon (1998) found, for example, that gay men and lesbians are “less liked” if they are perceived as unmasculine or unfeminine, respectively. Gender nonconformity as a separate construct is most frequently discussed in relation to children and lesbian, gay, bisexual, and transgender (LGBT) adolescents. Traditionally, social scientists studying childhood gender nonconformity were most interested in gender nonconformity as a predictor of homosexuality or psychosocial problems in adulthood (Bailey & Zucker, 1995; Dawood, Pillard, Horvath, Revelle, & Bailey, 2000; Landolt, Bartholomew, Saffrey, Oram, & Perlman, 2004; Weinrich, Atkinson, McCutchan, & Grant, 1995). However, in recent years a number of studies have begun to consider gender nonconformity in relation to prejudice, bullying and school climate (Bochenek & Brown, 2001; Cianciotto & Cahill, 2003; Gay–Straight Alliance Network, 2004; Harris Interactive & Gay, Lesbian, and Straight Education Network [GLSEN], 2005).

Although this article considers non-trans gender-identified LGB populations, prejudice against transgender populations must be mentioned as an important piece of the gender nonconformity picture. The growing body of research on transgender identities, health, and well-being both implicitly and explicitly deals with the role of gender nonconformity in experiences of discrimination and barriers to care (Lombardi et al., 2001; Park, 2002). The frequency and severity of violence targeting transgender individuals are often seen as evidence of the widespread prejudice against those visibly crossing gender norms (Sandfort, Melendez, & Diaz, 2007).

Prejudice and Discrimination as Social Stressors

Research on stigma, prejudice, and discrimination as social stressors has attempted to identify impacts of prejudice on both mental and physical health (Guyll, Matthews, & Bromberger, 2001; Kessler, Mickelson, & Williams, 1999; Krieger & Sidney, 1996; Meyer, 2003a; Schulz et al., 2000; Williams, Spencer, & Jackson, 1999). A growing body of research has suggested that “the psychological and physiological correlates and consequences of discrimination are similar to those of other psychosocial stressors” (Williams et al., 1999, p. 72). Meyer (2003a) posited that LGB individuals are exposed to such social stress and suggested a set of relationships between various levels of social stressors, minority identities, coping and social support, and mental health outcomes. Demonstrating the potency of antigay prejudice as a social stressor, Herek et al. (1999) have shown that antigay hate crimes can have greater negative psychological impact than other criminal victimization that does not target the LGB person’s sexual orientation. However, similar to research on antigay violence, gender nonconformity has been a missing or invisible component of stress research. Given the historical interconnectedness of sexual orientation and gender nonconformity, the visibility and targeting of gender nonconformity, and the potential impact of such prejudice on the life and health of LGB individuals and communities, the inclusion of gender nonconformity in research into sexual-orientation-based prejudice is essential.

METHOD

Sampling and Recruitment

The study results reported are from Project Stride, a study of stress, identity, and mental health in diverse LGB populations (Meyer, Frost, Narvaez, & Dietrich, 2006).

We used a venue-based sampling of LGB respondents. Venues were selected following ethnographic immersion into the communities of interest by the field director and outreach workers. Sampling venues were selected to ensure a wide diversity of cultural, political, ethnic, and sexual representation within the demographics of interest. Respondents were sampled in diverse New York City venues that catered primarily to LGB individuals (e.g., business establishments such as bookstores and cafes, social groups, and outdoor areas such as parks). Recruitment was conducted by outreach workers who approached potential study participants and invited them to participate in the study, described as concerning the health of “LGB communities.” To enrich the diversity of the sample by reaching respondents who were less likely to be found in public venues, we also used snowball sampling, in which respondents were given letters of invitation to pass along to potential respondents such as friends and colleagues. To reduce bias, venues were excluded from our venue-sampling frame if they were likely to over- or underrepresent people receiving support for mental health problems (e.g., 12-step programs, HIV/AIDS treatment facilities), or significant life events (e.g., organizations that provide services to people who have experienced domestic violence).

Between February 2004 and January 2005, 25 outreach workers visited a total of 274 venues in 32 different New York City zip codes. Recruiters first completed a brief screening form for each potential respondent that would determine eligibility for participation in the study. Respondents were eligible if they were 18–59 years old, New York City residents for 2 years or more, and self-identified as: (a) heterosexual or LGB; (b) nontransgender man or woman; and (c) White, Black/African-American, or Latino (respondents may have used other identity terms in referring to these social groups). We used quota sampling to ensure approximately equivalent numbers of respondents of similar age, across gender and race/ethnicity. Eligible respondents were contacted by trained research interviewers and invited to participate in a face-to-face interview. The cooperation rate for the study was 79% and the response rate was 60% (American Association for Public Opinion Research, 2005). Response and cooperation rates did not differ significantly by gender, race, and sexual orientation. Recruitment efforts were successful at reaching individuals who resided in diverse New York City neighborhoods and avoiding concentration in particular “gay neighborhoods” that is often characteristic of sampling LGB populations. Interviewed individuals resided in 128 different New York City zip codes and no more than 3.8% of the sample resided in any one zip code area.

Participants

Three hundred and ninety-six LGB respondents were interviewed in-person, using computer-assisted and paper-and-pencil instruments. The sample was designed to include equal numbers of White (34%, n = 134), Black/African-American (33%, n = 131), and Latino (33%, n = 131) respondents, as well as equal numbers of men and women (50%, n = 198 each). The mean age of the respondents was 32 (SD = 9). Nineteen percent had education equal to or less than a high school diploma (n = 97); 52% (n = 267) had a negative net worth (they owed more than their total assets (Conger et al., 2002); and 16% (n = 83) were unemployed. Some notable demographic differences existed among the subgroups in the sample defined by race/ethnicity, gender, and sexual orientation. The Latina lesbians/bisexual women were the least educated (30%, n = 20 having a high school diploma or less); and the Black lesbian/bisexual women had the highest instance of negative net worth (73%, n = 45). Interviews lasted a mean of 3.8 hr (SD = 55 min). Respondents were paid $80 for their participation in the study. The research protocol was reviewed by the Western Institutional Review Board. Respondents signed a written informed consent after the study procedure had been fully explained to them. As this study was part of a larger epidemiological study, the reader is invited to visit the full project Web site for further information about methodology (www.columbia.edu/~im15).

Measures

Self-Appraisal of Discrimination

The everyday discrimination measure (Cronbach’s α = 0.85) was modified from the 8-item instrument originally developed by Williams, Yu, Jackson, and Anderson (1997) and based on qualitative research with African-Americans. This instrument measured chronic, routine, and less overt experiences of unfair treatment. The eight items assessed experiences of: being treated with less courtesy, less respect, as if others are better, as if not smart, as if dishonest, receiving poorer service than others, being feared, and being called names or insulted. The scale was adapted so that it applied to all the minority groups in the study (i.e., gender, racial/ethnic, and sexual minority identities). The questions inquired as to how often these experiences occurred over respondents’ lifetimes on a four-point scale (1 = often through 4 = never). For each item, respondents were asked whether the experience was related to their sexual orientation, gender, race/ethnicity, physical appearance, or some other form of discrimination. Responses were coded so that higher scores reflected more everyday discrimination. Summed scores were divided by the number of items for the scale to obtain a mean total score for each participant. In addition, the modified everyday discrimination measure included a probe to record gender nonconformity as a target of discrimination among those who endorsed physical appearance (“What was it about your physical appearance that lead to. . . [type of discriminatory treatment]?”). This item was used to delineate the subset of respondents who reported discrimination related to gender nonconformity.

Life Events Assessment of Prejudic

Adapted from the Structured Event Probe and Narrative Rating (SEPARATE) scale (Dohrenwend, 2004; Dohrenwend, Raphael, Schwartz, Stueve, & Skodol, 1993), this instrument was used to evaluate general life events by assessing both subjective and objective stress-inducing properties of life events, including fatefulness (the extent to which an event occurred outside of the respondents’ control), threat to life, physical integrity, goals, and magnitude of change in usual activities that is likely to be brought about for an average person. The interview included an events checklist and probes about the number, dates, and types of events the participant experienced. A total of 43 possible life events were assessed; events assessed included, for example, job loss, separation from a partner, assault, major health problems, and childhood abuse events. All events receiving an affirmative response were probed using standardized guidelines; interviewers recorded short narrative summaries for each event. Each probe examined explicit evidence regarding the cause and context of the event (for example, the circumstances of a job loss) regardless of the participant’s attribution of the cause of the event.

After comprehensive data were collected, two independent raters rated the event on the dimensions previously outlined (such as fatefulness and threat to basic needs). The raters also classified the event as to whether or not prejudice was involved, and if so, the extent to which the event was motivated by prejudice (rated on a scale of 0 to 4, for which 0 indicated an event that was not at all motivated by prejudice and 4 indicated completely motivated by prejudice) and the type of prejudice that was involved was specified (i.e., gender, ethnic/racial, age, SES, religion, gender nonconformity, other physical appearance, disability, or sexual orientation status). In all instances, the average score of the two raters was analyzed. Cases in which the two raters were discrepant by more than 1 point were resolved by the research team in weekly rating meetings. Each event had the potential to be rated on 15 dimensions. Overall, 5,139 events were rated (i.e., 77,085 potential ratings). Of these potential ratings, 2.2% (1,705 ratings) were found to be discrepant and required consensus rating by the group. The most frequently discrepant ratings were source and magnitude.

To code an event as involving prejudice based on gender nonconformity, the raters looked for gender nonconformity in multiple domains: the respondent’s self-presentation (e.g., gender nonconventional clothing or mannerisms); the perpetrator’s communication (e.g., epithets targeting gender transgression); and other surrounding factors(e.g., the respondent socializing with others who were gender nonconforming). The presence of these gender nonconformity markers were based on self-report, rather than targeted interviewer probing. One respondent might have multiple events involving prejudice; in this case, each event would be counted separately.

Data Analysis

Two groups of individuals affected by prejudice due to gender nonconformity were identified using the 2 measures previously described: (a) 76 individuals who reported on one or more items of the everyday discrimination measure that they had been exposed to discrimination related to gender nonconformity, and (b) 35 individuals who reported one or more life events that were subsequently rated as involving gender nonconformity. We first conducted quantitative analyses to briefly describe these two subgroups. We compared demographic characteristics of the exposed groups (n = 76, n = 35, respectively) to the remainder of the LGB sample (n = 320, n = 361, respectively). We used odds ratios to look for differences based on gender, race/ethnicity, age group, educational attainment, and net worth between the exposed and unexposed groups (based on perception of discrimination and rater-assessed discrimination, respectively).

The heart of this study is our thematic analysis of the events narratives, which we conducted to explore the range of gender nonconformity prejudice and the contexts in which such prejudice occurs. Thematic analysis is defined broadly as a flexible, qualitative “method for identifying, analyzing, and reporting patterns (themes) within data” (Braun & Clarke, 2006, p. 79). Braun and Clarke argued that thematic analysis, although not often named explicitly, is frequently used in qualitative research because it offers a set of tools that can be used by a variety of epistemological frameworks. For this study, thematic analysis provided a way to work within a large and rich data set, but highlight a small selection of key themes.

We used a deductive, or theoretical, approach to the analysis (Braun & Clarke, 2006; Gilbert, 1993): that is, we used the minority stress model as applied to LGB populations as a theoretical base from which to define the themes of interest. Meyer’s (2003a) minority stress model hypothesizes that those from stigmatized social categories are exposed to excess stress as a result of their minority social position, and suggests a set of relationships between various levels of social stressors, minority identities, coping and social support, and mental health outcomes. The dimension of interest for this analysis was context: The situation in which gender-nonconformity-related prejudice events occurred; these situations can range, as we show, from anonymous incidents on the street to intimate violence within families. This dimension was selected because it provides essential groundwork for understanding the functioning of such prejudice in individuals’ and communities’ daily lives.

We followed a modified version of Braun and Clarke’s (2006) six phases of thematic analysis, using an iterative process of reviewing and coding (May, 2001). Using the codes of the independent raters, we isolated the event narratives that involved gender-nonconformity-related prejudice. We transcribed the handwritten interviewers’ summaries of each event, and read through the narratives multiple times to familiarize ourselves with the details of each response. On our second and third passes through the narratives, we began systematically coding features of the data related to gender nonconformity. We then grouped all relevant codes into our focal themes, including setting, perpetrator, and reason for targeting.

RESULTS

Prejudice Life Events and Everyday Discrimination

Group A: Self-Appraised Routine Discrimination

Of the 396 LGB respondents, 76 (19%) reported at least one occurrence of routine discrimination attributed to gender nonconformity on the everyday discrimination measure. Of them, 62 (82%) also reported at least one occurrence of discrimination due to sexual orientation. Among all 396 respondents, 295 (75%) reported at least one occurrence of discrimination due to their sexual orientation, and 115 (29%) reported neither.

The 76 respondents who reported gender nonconformity discrimination on the everyday discrimination measure did not differ significantly from the larger study sample in terms of their race/ethnicity, age group, educational attainment, and net worth (Table 1, column A; statistical analyses not shown). However, women were more likely than men to report discrimination instances due to gender nonconformity (OR = 4.14, 95%CI = [2.34–7.35], p < .001). In contrast, women were no more likely than men to report discrimination related to sexual orientation.

TABLE 1.

Demographic Differences in Exposure to Gender Nonconformity Prejudice

		A. Everyday Discrimination		B. Life Events
	All Respondents	Related to Gender Nonconformity		Prejudice Related tc Gender Nonconformi
Column	N	N	%	N	%
Total	396	76	19	35	9
Gender
Men	198	18	9	16	8
Women	198	58	29	19	10
Race/ethnicity
White	134	30	22	7	5
Black/African-American	131	25	19	13	10
Latino	131	21	16	15	11
Age Group
18–30	196	33	17	17	9
31–59	196	43	22	18	9
Education
< College degree	206	45	22	21	10
≥College degree	190	31	16	14	7
Net worth
Negative	171	31	11	18	11
Positive	212	45	21	17	8

Note. Negative net worth includes respondents reporting no assets or debt; positive net worth includes respondents reporting assets greater than zero.

Group B: Rater-Assessed Acute Life Eevents

In the assessment of prejudice-related life events, gender nonconformity was a cause of prejudice for 35 respondents (9% of the 396) who had, in sum, 43 separate events related to gender nonconformity. All gender nonconformity events were also assessed as involving sexual orientation-based prejudice. In the total sample, there were 133 respondents who experienced a prejudice event assessed by the raters to be related to sexual orientation. The group of 35 respondents who experienced a gender-nonconformity-related event represents about 27% of those experiencing prejudice related to sexual orientation. These 35 respondents exposed to life events related to gender nonconformity were not significantly different than the rest of the LGB sample in terms of gender, race/ethnicity, age group, education level, or net worth (Table 1, column B; statistical analyses not shown). All 35 individuals assessed with gender-nonconformity-related life events were among the 76 respondents who reported gender nonconformity as a target of discrimination at least once on everyday discrimination measure.

We next compared the type of life events and instances of discrimination that were due to gender nonconformity with those that were due to all other types of prejudice (sexual orientation, race/ethnicity, religion, etc.) to examine whether gender nonconformity is related to particular ways in which prejudice is enacted (Table 2). Some areas proved to be similar, for example, 42% of the gender-nonconformity-related life events involved violence, compared with 38% of all other prejudice-related life events. Additionally, 21% of the gender nonconformity life events were nonassault victimizations (harassment or vandalism), as compared with 20% of all prejudice events. However, in some areas gender nonconformity was distinct from other forms of prejudice events. For example, childhood peer assault (e.g., bullying and attacks at or near school) was associated with 12% of the gender nonconformity events but 7% of the total prejudice. Notably, the only two prejudice events of rape or childhood sexual assault both involved gender nonconformity-related prejudice.

TABLE 2.

Types of Life Events Related to Gender Nonconformity and All Prejudice Life Events

Event Type	% of All Prejudice Events(N = 260)	% of Gender Nonconformity Related Events(N= 43)
Assault or threat of violence	38.1%	41.9%
Nonassault victimization (harassment, vandalism)	20.4%	20.9%
Childhood peer assault	6.9%	11.6%
Theft or burglary	2.3%	4.7%
Homelessness	5.4%	4.7%
Crime against significant other	1.9%	4.7%
Childhood physical abuse	3.5%	2.3%
Dropping out of high school	2.7%	2.3%
Rape	0.4%	2.3%
Childhood sexual assault or abuse	0.4%	2.3%
Significant other raped	0.4%	2.3%
Unemployment 4 weeks or longer	5.0%	—
Other childhood events	1.9%	—
Convicted of a crime	1.5%	—
Charged with a crime, not convicted	0.8%	—
Arrested, not charged	1.1%	—
All other events	7.3%	—
Total	100.0%	100.0%

Thematic Analysis of Prejudice Events

We now examine the range of contexts in which gender-nonconformity-related prejudice events can occur, and, following the minority stress hypothesis, suggest that it is important to analyze the context and nature of such prejudice because these may have different psychosocial impacts.

We make distinctions between some of the events where gender-nonconformity-related prejudice is central to the event’s occurrence (e.g., an attack on the street that was explicitly motivated by gender nonconformity prejudice) and others where gender nonconformity is one of many factors involved. We also consider the connections between sexual orientation and gender nonconformity. Overall, these narratives provide a small window into the nature of prejudice against gender nonconformity across a range of situations.

Gender-nonconformity-related prejudice events took place in three broad types of settings: (a) a familial setting, or other home environment that is often assumed to be safe, such as the home or the car of a family-member; (b) a school or a work-place setting (including one event in a church); and (c) a public space such as a commercial or retail setting. For each setting described in this section, we also describe the perpetrators, as they are intricately linked, yet distinct from the setting. For the most part, perpetrators correspond to the settings: (a) home environment perpetrators were often family members or family friends, although a few events were perpetrated by neighbors; (b) school or work-place perpetrators were often peers and colleagues and less often school authority figures or employers; and (c) retail and public space perpetrators were often strangers who typically initiated the incident without prior contact with the respondent.

Home and Familial Spaces

Public forms of hate violence or harassment are the most visible to outside observers; events taking place in the home are rarely thought of in terms of prejudice or hate crime. We found that 8 of the 43 gender-nonconformity-related prejudice events (about 18%) took place in a respondent’s home and/or were perpetrated by a family acquaintance or member. In these events narratives, the familial or acquaintance settings were generally those in which some of the most serious events occurred. In a particularly brutal example, the respondent, a Latino man (age 43), reported that at age 13 he was raped, beaten unconscious, and left for dead by a family friend who picked him up after school. In this event, the respondent explained that the family friend “raped me because I was gay and to teach me what a faggot goes through.” Although the respondent was only 13, he was identified as “a faggot” by the family friend due to his gender nonconformity: “I was very feminine (hand gestures, walk, dress[ed in] tight pants).” In another home-based event, a Latino man (age 31) reported that up to age 12 he was excessively disciplined (compared with his siblings) and was humiliated by his mother and his four siblings. He was the only sibling who was gender nonconforming: He was “effeminate” and “liked to play with dolls,” and because of this his mother beat him with various objects. In another instance, he reported that his brother called him “faggot” and spit on him—an act that was condoned by their mother.

Other familial settings did not involve the same degree of violence and life threat, but were in other ways severe and life-changing, such as the two cases of respondents who became homeless at age 17 because they were kicked out by parents. In one instance, the respondent (a Latina woman, age 22) was “asked to leave” her house by her parents, who said, “‘We have a daughter, not a son, and we don’t want you in our home.’“ The respondent commented, “I wasn’t trying to hide my queerness from them—had a buzzed head, all friends female, they just sort of knew and assumed . . . ” She also noted that it was her father who most wanted her to leave; the mother stayed in contact with the respondent and eventually “coerced” the father into “letting me stay at home again.” In a similar event, a Latina woman (age 35) explained, “My dad kicked me out of my house [at age 17] because my stepmother didn’t like that I was gay. . . . Stepmother thought I would be a bad influence on her daughter—didn’t like the way I dressed—like a drag king.” In both of these events, the respondent’s self-presentation was considered a threat to the household. In the latter case, the stepmother also considered the respondent to be a “bad influence on her daughter”—a threat to the daughter’s sexuality and/or femininity. Also, both cases involved some degree of tension or a differential between two parents who had discordant responses to their child’s gender nonconformity; yet, at least initially, they chose to stand unified as parents, rather than one opting to support the child. In the latter event, the respondent did not reconcile sufficiently to return home, and instead reported seeing her father only on holidays. In the former event, the respondent eventually returned home although she and her father, who had been “really close growing up,” didn’t talk for 2 years after that, and she reported feeling pressured to “play by their rules,” even to the extent that she “got a boyfriend . . . and it was just a big mess.”

Neighbors as Perpetrators

The other form of home-environment-related events involved threats and abuse conducted not by family members, but by neighbors who vandalized or otherwise threatened the respondent’s place of residence. These events are linked to the family/home events in that both forms of events deeply threaten an individual’s sense of safety and stability in their own homes. However, neighbor-perpetrated prejudice is more akin to traditional notions of hate crimes and other forms of discrimination, and its psychological and behavioral impacts may be quite distinct from the effects of prejudiced violence or neglect by family members. Three of the eight home-based events were related to neighbors (about 7% of the 43 total gender nonconformity events).

A typical example of prejudiced intimidation by neighbors is the case of a Latino man (age 22) who reported that in the 4 months preceding the interview he had repeatedly found threatening notes and graffiti on his front door (“FAGIT” and “Be careful where you walk around here at night”). This respondent described himself as “very ‘out’ in my neighborhood and very fashion-oriented,” suggesting that physical cues of gender nonconformity were the dominant reason that he was targeted. In another situation of prolonged harassment and threat, a Latina woman (age 30) described the series of events that occurred when she and her partner, a mixed-race couple, moved to North Carolina to raise their 6-year-old daughter. The White neighbors, who “didn’t know my girlfriend was a woman til the summer,” engaged in vandalism and harassment of increasing severity that crosscut racism and homophobia, and implicated prejudice based on gender deviance. This included threatening their daughter and calling her “nigger,” breaking windows, calling out “spic” and “dyke,” running over their dog and killing it, and finally digging up their septic tank, resulting in the property being condemned and the respondent’s family being forced to move. According to the respondent, when these incidents were reported to law enforcement, the police claimed there “wasn’t proof” and called the situation a “neighborly dispute.”

These events, both those perpetrated by family and those perpetrated by known or unknown neighbors, indicate that for LGB people who are perceived as gender nonconforming, the home is, at times, not a place of security but the source and location of a host of social stressors.

School and Work Environments

The events narratives demonstrate that gender nonconformity also plays a role in many antigay incidents within schools and places of employment. Eight of the 43 events took place in school settings, and two in workplace settings. School settings are worth particular attention: Young people may not be “out” or even have formed a distinct nonheterosexual identity, but gender nonconformity can still be used by peers to tease, ostracize, or bully their gender-different classmates. One respondent, a Black woman (age 20), describes a junior high bully who “called me ‘butch,’ ‘lesbian,’ cuz I was aggressive and played sports. He’d push me sometimes but didn’t hurt me. I told the teachers but it created more conflict, so I stopped telling. Kept to myself in general.” Most of the other events in this category involved harassment at the high school level, leading at least one respondent to drop out of school at age 14. This respondent (a Black man, age 48) indicated that he “didn’t even approach school officials,” because he knew it would be futile and could potentially exacerbate the situation, similar to the response of the woman just quoted.

For some respondents, harassment and violence from their peers and other kids extended off-campus when the school day was over. In a particularly poignant narrative, a Latino man (age 27) described an attack after school at age 16 when he was walking home with a friend.

AT first they called us faggots and said, “We’re going to beat you up so you behave like men.” They then threatened us with a bat, came at us and held us down to the ground. They did not hit us or beat us . . . just held us down calling us faggots. They robbed us before leaving, took our jewelry. . . . We never told anyone about this—in a way I felt I deserved it for being gay. After this happened, I stopped wearing jewelry and didn’t walk to the train station; I took buses.

This narrative could also be categorized within the “public space” prejudice context, described in the following section. However, such events are most appropriately considered in school terms, given that the trip to and from school deeply impacts educational experience, and because “incidents occurring on public transit [or en route to public transit] can be a better indication than school-based incidents” of the experience of students who are victims of antigay violence (Anti-Violence Project, 2003, p. 2). Note, in particular, the language of gender regulation employed by the assailants: “We’re going to beat you up so you behave like men.”

One respondent, a Latina (Brazilian) woman (age 27), specifically reported sexual orientation and gender nonconformity prejudice at her workplace, where she was “extremely out on the job.” In the incident she described, she was new to the job and driving with male colleagues who were checking out women passing in the street and “asking if I like ‘that type’ and what I would do in bed with them—who’s the man and who’s the woman.” Like many of the prejudice events, this incident contains the intersection of multiple forms of prejudice, as these colleagues were also making comments “about Brazilian women, and how they’ll fuck you for five dollars,” all for the benefit of the respondent. She also noted that this was only one of multiple incidents, and that she subsequently became increasingly hateful toward colleagues and dreaded going to work. The respondent tried discussing the incidents with her boss, but the boss refused to honor her request not to work with those particular men anymore on the grounds that “they were good workers.”

Commercial Settings and Other Public Spaces

Over half of the events (23 of the 43) occurred in public spaces, such as a place of business or the street. These are the events that are most commonly thought of as hate crimes in both the popular imagination and the hate crimes literature. Sometimes such events take place in a known “gay” location—such as by a gay club, or in New York City’s Greenwich Village, where perpetrators purposefully go to seek LGB individuals with the intention to hurt them—and other times they take place in nongay spaces—such as a store, a train, or on the street where perpetrators happen to encounter an LGB person. In our respondents’ narratives, perpetrators of these events were almost always strangers, although in one case the perpetrator was a business owner with whom the respondent had multiple contacts. Public space was also, in one instance, the site of more institutional manifestations of prejudice, in the form of overt police harassment.

A number of events in this category involved physical assault on the street, such as the case of a Black man (age 41) who described walking down the street in New York City (not in a known gay neighborhood) when he was provoked and assaulted: “Two men made a remark like, ‘Catch the girlie comin’ down the street’ and called me faggot and homo. [They] called me these names and they started hitting me—I had scissors and stabbed one of the men in the arm and shoulder.” In another example, a White woman (age 24) was in a residential neighborhood (not in New York City) when two young men passed her and asked her how old she was, then asked if she “was a boy or a girl,” to which she said “girl” (she noted to the interviewer that she “looked more masculine at the time”). She then started to run away from the boys, who ran after her and punched her in the face, dislocating her jaw. The respondent had to go to the hospital and was physically unable to eat for a week. This event is atypical of the sample, in that there is no explicit overlap between homophobia and prejudice against gender nonconformity—only the respondent’s nonconforming gender presentation was implicated in the assault.

Another respondent, a White woman (age 20) reported an event in which she was wearing a shirt and tie on the subway. “Four big guys got on the train—started being rude—’look at that dyke’—said they would rape me, [saying] ‘We’ll show you what real men can do.” The respondent reported that no one else on the train said anything, and although she was not harmed in the incident, she became more “reserved with my clothing—didn’t want to outwardly appear gay—will cover up with a jacket or have someone with me.” This was not the only event in which threat of rape (explicit or implicit) was levied against someone who was gender transgressive. Another example came from the previously-discussed 22-year old Latina woman (whose parents kicked out of the home at age 17) who described a recent attack against her on a train platform near her apartment. In this event, five young men came up to the respondent and asked to borrow her cell phone:

And they figured I was a young guy and I didn’t speak because I didn’t want them to hear my voice. One pinned me up against a wall—I said something like “Back off,” and they heard my voice and were like, “You’re a fucking dyke”—and they started hitting me, trying to take my shirt off. . . . I was really scared. But then a train came and people got off and I ran away, onto the train.

Even though this respondent escaped further physical harm by running onto the train, she continued to be subjected to prejudice-based threat: Once on the train, the respondent “called a friend and then started getting harassed by another man who heard my story and he was like, ‘They should’ve fucked you up. . . . ” thus identifying himself as a new source of threat. This conclusion demonstrates that even making it to the relative safety of a more public space does not free the visibly gender nonconforming person from peril. The omnipresence of a prejudiced public eye meant that the respondent was not even allowed to recount her trauma to a friend without being further harassed for her gender presentation. The respondent explained this chain of events by noting, “I had just gotten a haircut, so I guess looked especially boyish.”

A final important event in this category demonstrates the potential for prejudice events in public space initiated by the police themselves. It also illustrates the historical continuity within this analysis of prejudice. This narrative describes events that took place over 3 decades ago, when a particular kind of police harassment in New York City was far more common. The respondent was 50 during the time of the interview, but age 16 during these events:

I would be hanging out with my friends and we were wearing chest bands to make our chests flat—wearing men’s clothing—very butch—holding hands. We looked like boys. This would happen routinely: After hanging out outside a club, cops would harass us—“You think you’re a man, dyke”—they would throw us against the wall, search us—my friends would get arrested for wearing men’s clothing. That was a law then—“male impersonation.” I was not arrested. Made me more courageous. If I was brave, I would have to deal with penalties. [I] did nothing differently.

Although these kinds of events are today less frequent, or at least less overt (and the anticross-dressing regulations are no longer in place in New York City), police harassment based on sexual orientation and gender presentation, as well as gender identity, still exists (Amnesty International USA, 2005).

DISCUSSION

Our findings illuminate multiple facets of the role that gender-nonconformity-related prejudice can play in the lives of LGB individuals. The prevalence of gender-nonconformity-related prejudice events was 9% using the rater-assessed measure and 19% from the self-appraised measure—a discrepancy that is probably related to the greater severity of incidents captured in the events category, as compared with the everyday discrimination category. The qualitative analysis of the event narratives demonstrated the significance, complexity, and severity of gender-nonconformity-related prejudice, and its range of manifestations across the lives of LGB individuals and communities. We focused on one element of the minority stress process that was observable in the narratives: the situational contexts from which social stressors emerged. In particular, the issue of prejudice and hate violence within the home warrants further discussion.

Violence at Home

Our analysis of the environmental contexts of prejudice events demonstrated that 8 of the 43 gender-nonconformity-related prejudice events took place in a respondent’s home and/or were perpetrated by a family member or acquaintance; this is an important element of prejudice experienced by LGBT persons, and particularly those who are gender nonconforming within a prejudiced family. In his survey examining violence against lesbians and gay men, Comstock (1991) also described the home as one of the primary settings in which antigay violence occurs. Comstock found that 26% of 157 victims of antiay/lesbian violence reported events that occurred in home settings (including the respondent’s own home, a parent’s home, or another’s home).

Garofalo (1991) studied police reports in New York City to compare racial bias and nonbias crimes (matched by type and day of occurrence). This study found that hate crimes were more likely to involve multiple offenders and victims, to occur between strangers and in public places, and to involve young male individuals (both victims and perpetrators). Although Garofalo’s findings do, to some extent, map onto the distribution of prejudice events collected here (e.g., 19, or almost 50%, of these events were perpetrated by male strangers in public places, but slightly more women than men in this study had experienced such events), this approach to focusing on dominant hate crimes patterns tends to overlook the difficulty of conceptualizing and measuring hate crimes that occur in the home, in schools, and in other places where greater safety than in the street or other public places is assumed.

Events in the home that are perpetrated by relatives or family members may resemble hate crimes in prejudice-motivation and in their severity or excessive brutality, but are unlikely to be reported as hate crimes. For example, consider the case of the respondent who was beaten because of playing with dolls and perceived as not being sufficiently masculine in childhood. Such a beating would likely not be reported as a hate crime (by the respondent himself, still a child at the time, or by a witness), although it might fit the criteria. The Anti-Violence Project of New York City collects and reports data related to domestic violence among LGBT populations. In their most recent report they found that, of the 423 new domestic violence cases they opened in 2003, 9% were perpetrated by relatives or family members (excluding partners); of the 486 in 2004, 12% were perpetrated by family members. Overall, only 36–38% of domestic violence incidents were reported to the police over those 2 years (Anti-Violence Project, 2005); no comparable estimates could be found of the proportion of all hate violence incidents that are reported to the police.

Unlike many other oppressed minority groups, young LGB individuals typically are not raised by other sexual minorities and are often raised in environments that promote antigay attitudes (Meyer & Dean, 1998). In that sense, the race/ethnicity model of prejudice, discrimination, and hate crimes is not ideal for application to some LGB contexts. This has implications for all forms of prejudice experienced by LGB individuals. In a study of LGB youths, D’Augelli (2006) found that 12% of his sample of 542 young people reported that their mothers had verbally abused them because they were LGB, and 7% reported that their fathers verbally abused them because they were LGB; there were also reports by young men and women of physical abuse by mothers and fathers because of their LGB identity. Balsam et al. (2005) found that, within the same family, LGB siblings may be at greater risk for victimization than their heterosexual siblings. This kind of intrafamilial, prejudice-based violence can result in adverse mental health consequences: in D’Augelli’s study, parental negative reactions to their child’s LGB identity were associated with lowered self-esteem and more negative mental health outcomes in the young person. Additionally, for LGB individuals growing up in an intensely prejudiced family structure, physical health and the security of basic needs such as shelter are also at risk (Hyde, 2005; Savin-Williams, 1994). Indeed, we highlighted cases from this study in which teens were forced out of their homes by parents who could not accept their child’s sexual orientation and/or gender presentation.

Gender nonconformity plays an important role in this kind of familial or home-based violence against youth. As in the narrative of the woman whose stepmother did not like the way she dressed “like a drag king,” gender nonconformity can be taken as a marker of homosexuality by family members, even if the child does not yet identify as LGBT. Some young people (e.g., the boy who only wants to sew and decorate, or the girl who won’t grow out of her “tomboy phase”) may be persecuted by family members for gender nonconformity because they are assumed to be homosexual, even though they ultimately develop as heterosexual (Harris Interactive & GLSEN, 2005). Because of the potentially damaging health and developmental consequences outlined above, the issues faced by youth who are gender nonconforming have rightfully merited particular attention in LGBT research and advocacy work. This work is essential, and should continue to build on the message that gender-nonconformity-based harassment and violence are detrimental to all students who don’t fit into the dominant gender ideals, and that such prejudice creates a school climate that is not conducive to learning or health (Russell, 2006).

Inclusion of Gender Nonconformity in the Minority Stress Model

Based on our findings, we recommend that researchers working with the minority stress model explicitly incorporate gender-nonconformity-related prejudice as a stressor for sexual minorities. The importance of paying attention to gender nonconformity in close relation to, or overlapping with, sexual minority status cannot be overstated. As our research has shown, there is an implicit link between gender nonconformity and sexual orientation. This link means that those exposed to prejudice events due to their nonconforming gender presentations are likely also being targeted as sexual minorities, regardless of the individual’s actual sexual or gender identity.

Study Limitations and Conclusions

In this study, we faced a number of conceptual and measurement limitations. The modified everyday discrimination scale, in particular, may have been limited by its reliance on subjective appraisal of discrimination and self-report. Some research has shown that minority group members may minimize discrimination (both in terms of perception and articulation) to protect self-esteem, maintain a sense of control, and protect against both recrimination and reinforcement of derogatory stereotypes (Crocker & Major, 1989; Krieger & Sidney, 1996; Ruggiero & Taylor, 1997).

We found an unexpected gender discrepancy: In the self-appraised measure, women reported significantly more gender-nonconformity-related discrimination than men, but this was not the case in the rater-assessed measure. We believe that this discrepancy reflects measurement limitations associated with the self-appraised measure of chronic discrimination, as well as conceptual challenges posed by the question of gender nonconformity. Research into anti-LGB violence has suggested that men report higher levels of victimization than women (Bon-tempo & D’Augelli, 2002; D’Augelli, 2006; Herek et al., 1999). The gender gap that we found for gender nonconformity suggests that gender-nonconformity-related discrimination is perceived and reported differently than antigay discrimination. Two measurement explanations are plausible, both forms of reporting bias, and both with implications for the identification and reduction of gender nonconformity-related discrimination.

First, women may be more likely to perceive and report gender-related discrimination due to the raised gender consciousness that the feminist movement of the 1960s and 1970s provided (Rosen, 2000). Men may not use the same framework for thinking about gender and power relations as women (Seidman, 2002), and gay men might, instead, conceptualize discrimination solely in terms of sexual orientation or other targets. Second, gay men may also be less likely to self-report such discrimination because gender nonconformity is still highly stigmatized even within many communities of gay men (Garber, 1992; Nestle, 1992; Taywaditep, 2001). Femininity in all men is socially unacceptable, but for sexual minority men, femininity places them in an even more socially subjugated position (Sandfort, 2005). The social pressure to be masculine makes it likely that gay men would actively or unconsciously avoid self-reporting gender nonconformity as a measure of psychosocial protection (Ruggiero & Taylor, 1997).

This limitation of the self-appraised measure reinforces the importance of externally-assessed measures of prejudice where gender is concerned. The SEPARATE measure used to gather life events narratives represents an attempt to move away from the biases of perceived discrimination measures; however, this measure is also limited, to some extent, by its reliance on self-report. Reporting bias could be present if certain groups (or certain types of events) are less likely than others to incorporate details into their narratives that signify gender nonconformity, causing an underclassification of that subgroup (or that type of event) in the event narratives (Meyer, 2003b).

Another limitation is the use of a nonrandom sample, which does not allow us to generalize the reported prevalences to the larger community. Nevertheless, the patterns we recorded and the comparisons among subgroups do not depend on external validity and are instructive. Future research using a random population survey should ascertain more accurate prevalence estimates of events related to gender nonconformity.

This study is also limited because it lacked an objective assessment of an individual’s gender nonconformity. We used implied gender nonconformity as identified via respondents’ reporting of prejudice-related life events and everyday discrimination. We cannot, however, assess the extent to which respondents were gender nonconforming. It can be claimed that in prejudice, discrimination, and violence, the perpetrator’s perception of gender nonconformity, rather than an objective assessment, is what matters. Still, it would be important to understand aspects of gender nonconformity that are associated with prejudice and violence. In the past, instruments for measuring gender nonconformity have focused on narrowly-defined behavioral traits or biological explanations of gender deviance in relation to sexual orientation (Lippa, 2000). Work to develop more nuanced measures of gender nonconformity-based prejudice is needed. Three key questions must be considered in this future measurement: (a) To what extent should gender nonconformity be based on self-report alone? (b) What are the appropriate ways to measure nonconformity with relative objectivity, or without relying on self-identification? and (c) How should one measure the extent to which gender nonconformity was an immediate motivator for a discriminatory or hate event, as well as the ways it was used within a prejudice event (e.g., through epithets)?

We studied nontransgendered LGB individuals. Future research into minority stress as experienced by transgender and transsexual populations is essential to fully understand the role of gender nonconformity within the minority stress model. In addition, a qualitative exploration of gender nonconforming heterosexuals’ experiences of prejudice and discrimination is also needed to flesh out gender nonconformity as a minority stressor. As we have argued, a full consideration of gender nonconformity should span the entire spectrum of gender possibilities.

We have demonstrated the importance of gender-nonconformity-related prejudice in the lives of LGB individuals and communities. Researchers have begun to recognize gender nonconformity as a field of inquiry in contemporary LGB research, but research on LGB health and anti-LGB discrimination has not sufficiently investigated gender nonconformity’s role. We argued that this form of prejudice must be disaggregated from sexual orientation if researchers are to expand their understanding of the social stressors impacting sexual minorities and their communities. Gender nonconformity is tied to histories of both abuse and resilience, as well as to the relatively new blossoming of transgender politics, and as such it has sometimes been a controversial or painful subject to discuss among LGBT and feminist communities. Although the links between sexual orientation and gender nonconformity may still be contentious, in this article, we hope to contribute to that ongoing and necessary conversation.