Introduction

Historically, lesbians, gays, and bisexuals (LGBs) in the United States have faced insensitive, degrading, and often hostile social environments (Herek 2009). Clearly, the “minority stress” literature shows that these practices can harm the well-being and mental health of LGB individuals (Meyer 2003). Responses to heteronormativity and heterosexism are multiple, depending on situational, contextual, political, and individual factors. As individuals, some sexual minorities may try to deny or hide their sexual identity from themselves and others (Gortmaker and Brown 2006). Some sexual minorities may personally accept their sexual identity but resign themselves to quietly enduring social injustices. However, many choose not to remain politically acquiescent, as many sexual minorities attempt to contest the basis of their subordination (Taylor et al. 2009). This may occur in several ways: sexual minorities may enact hidden resistances that covertly defy heterosexual privilege. They may also individually dispute heterosexist comments or use legal means to correct inequities, just as they may initiate or join social movements that contest unfair institutional practices (Miller et al. 1981).

While there is a growing literature on the goals, tactics, and outcomes of the gay and lesbian rights movement (e.g., Jenness 1995; Kane 2003; Wald et al. 1996), there are far fewer empirical studies on the reasons why North Americans join the gay and lesbian rights movement (e.g., Swank and Fahs 2011; Taylor et al. 2009; Waldner 2001). While the limited numbers of studies on gay and lesbian rights offer impressive insights, they often overlook the ways racial factors can inspire or inhibit participation in this American social movement (all but one of the cited LGB studies in this paper are based on samples in the United States). This negligence of racial factors is far from ideal. Studies often suggest that racial backgrounds influence the political behaviors of heterosexual Americans (Bobo and Gilliam 1990; Leighley and Vedlitz 1999; Schussman and Soule 2005) and issues of racial divisions and White privilege is prevalent in segments of gay and lesbian communities (Balsam et al. 2011; Fingerhut et al. 2005; Ward 2008).

To examine the race-gender blind spots of earlier research on gay and rights activism, we draw upon Collins’ (1990) work on intersectionality along with other sophisticated conceptions of intersectional identities (Riseman 2004). At its core, intersectionality argues that individuals reside in multiple systems of stratification (e.g., race, gender, sexuality, ability, and so on). People are constantly challenged to simultaneously respond to the numerous privileges and constraints that are bestowed on the various social roles they occupy. Through an intricate process of competing hierarchies, obligations, and experiences, these multiple locations inform and modify one another in ways that create many variations within a single stratum of people. Due to a “matrix of oppressions” in a racist, sexist, classist, and heterosexist society, affluent White heterosexual women generally have different power sources, life experiences, obligations, and expectations than poor Black lesbians, just as gay men might differently experience their own sexual identity compared to lesbian women and bisexual men and women. Succinctly summarizing intersectionality, Stewart and McDermott (2004) explained, “(a) no social group is homogenous, (b) people must be located in terms of social structures that capture the power relations implied by those structures, and (c) there are unique, non-additive effects of identifying with more than one social group” (pp. 531–532). Using an intersectionality framework nuances and improves understandings of gay and lesbian rights activism, as neglecting the effects of shared social statuses can lead “to incomplete, or possibly incorrect, conclusions” about the social world (Ovadia 2001, p. 342).

This paper uses intersectional analysis to improve upon early explanatory studies of LGB activism in the United States. First, this study uses logistic regressions to delineate the ways that gender and racial identifications are related to four types of LGB activism (voting, petition signing, protesting, and civil disobedience). While gender factors have been commonly studied in quantitative studies of LGB activism, the role of racial background has been almost entirely ignored. Moreover, by recognizing that sexual minorities can belong to more than one stigmatized populations, this paper also examines how race and gender intersections can alter the political activities of sexual minorities. Second, the paper addresses the ways that certain attitudes can foster gay and lesbian rights activism among gay and lesbian subpopulations. This is important because the salience of different framing practices can vary along gender and race lines.

Literature Review

Membership in stigmatized groups generally leads to greater levels activism on the behalf of that group. For example, gays and lesbians are more likely to join LGB rights movements than heterosexuals (Rollins and Hirsch 2003; Swank and Fahs 2011) and racial minorities belong to antiracism movements at higher rates than Whites (Bobo and Gilliam 1990; Schussman and Soule 2005). When addressing issues of multiple identities, many have argued that political activism may be patterned among the social divisions that rest within an aggrieved population (Simien 2007; Stewart and McDermott 2004).

Deprivation theories suggest that a desire for social change is strongest among people who have more than one stigmatized identity. In support of this claim, some studies have found that Black women are more supportive of feminism (Cook and Wilcox 1992; Tolleson-Rinehart 1992) and more politically active than White women (Cole and Sabik 2010; Manza and Brooks 1998). Conversely, resource theories suggest that dual oppressions can block any activist tendencies among women and racial minorities (Burns et al. 1997). Studies of this sort contend that political activism occurs less frequently among women who are unemployed, less formally educated, and Asian American or Latina (Coffé and Bolzendahl 2010; Duncan 1999; Manza and Brooks 1998; Ferree 1980; Tolleson-Rinehart 1992). Accordingly, this current study explores the “deprivation” and “resource” theories by discerning the ways that single or multiple race and gender identities influence the political activities of gays and lesbians in the United States.

Gender, Race, and Gay-Lesbian Rights Activism

The influence of gender on political participation is far from definitive. American samples that were collected in the 1950s and early 1960s suggest that women were slightly less inclined to be politically active than men (Barkan et al. 1995; Wallace and Jenkins 1995). Conversely, newer studies have suggested that this gender gap disappeared or had even reversed in the years that followed the second wave of the women’s movement (Bobo and Gilliam 1990; Harder and Krosnick 2008; Leighley and Nagler 1992; Paulsen 1994; Hritzuk and Park 2000). To complicate the topic even more, some studies suggest that gender plays a different role for each type of political expression. For example, a recent study suggested that heterosexual women were more likely to vote in elections and sign petitions but were less likely to write a politician or join a protest than heterosexual men (Coffé and Bolzendahl 2010).

The role of gender is unclear in studies of gay and lesbian political engagement. Some studies found gender differences for electoral activism, as lesbians wore more political buttons (Herek et al. 2010), while gay men wrote more letters to politicians (Herek et al. 2010) and made larger financial contributions to political candidates (Herek et al. 2010; Swank and Fahs 2011). Another study added that gay men wanted to go to gay and lesbian right protests slightly more than lesbian women (Lewis et al. 2011). Other studies found that the frequency of political activism was roughly the same with gays and lesbians (Jennings and Andersen 2003; Rollins and Hirsch 2003; Taylor et al. 2009; Waldner 2001). Thus, there is some doubt as to whether gays or lesbians have shown any greater group tendencies toward gay and lesbian rights activism.

Empirical studies rarely express certainty about how race influences gay and lesbian rights activism among sexual minorities. Studies show that race influenced the political activism of heterosexuals, particularly in type of activism performed (Hutchings and Valentino 2004). Compared to people of color, Whites more often voted, made campaign contributions, or volunteered for elected officials (Bobo and Gilliam 1990; Harder and Krosnick 2008; Leighley and Vedlitz 1999). Further, Blacks and, to a lesser degree, Latino(a)s had a greater tendency to turn to protest activities than did Whites (Paulsen 1994; Schussman and Soule 2005). Theoretically, the racist legacy of blocked opportunities in electoral realms has made social movements preferred vehicles of social change for Blacks and Latino(a)s.

Although these studies of heterosexual populations can be informative, the impact of race on LGB activism may have different meanings for heterosexual and sexual minority populations. Some studies suggest that sexual minority Blacks, Latino(a)s, and Asian Americans probably face greater stigma and hostility toward their sexual orientations compared to Whites. For example, surveys often find elevated levels of homonegativity among Blacks, Asian Americans, and Latino(a)s (Lewis 2003; Loftus 2001; Schulte and Battle 2004) while Black lesbians face more resentment when they “come out” to family members and friends (Grov et al. 2006; Mezey 2008). Additionally, LGB Blacks and Latino(a)s may suffer from a higher “sexual orientation” hate crime victimization rate than White gays and lesbians (Dunbar 2006; Meyer et al. 2008). Finally, one study concluded that race in crucial to “ally activism” because Euro American heterosexuals were more likely than Latino or African American heterosexuals to want to join gay and lesbian rights rally (Lewis et al. 2011).

To complicate the interplay between race, sexuality, and activism even further, White racism against people of color—both heterosexual and sexual minority—also impacted research findings. One older study found that gay White men were less racially prejudiced than heterosexual White men (Beran et al. 1992). Conversely, some studies find that sexual minorities of color are suspicious of overt and covert forms of racial prejudice among White gays and lesbians (Fingerhut et al. 2005; Masequesmay 2003). Recent qualitative studies noted that sexual minorities of color often encountered racial insensitivity in gay and lesbian restaurants and bars (Adams and Kimmel 1997) and Black college students are annoyed that White gay men often minimized the challenges of belonging to several marginalized populations (Goode-Cross and Good 2009). Similarly, quantitative studies also find that racial minorities of color face higher levels of disapproval and economic discrimination than White sexual minorities (Meyer et al. 2008), that darker skinned Latino gay men are treated worse by gay men than their light skinned counterparts (Ibeñez et al. 2009) and that Asian, Black, and Latino(a) sexual minorities are bothered by White sexual minorities who oppose interracial relationships (Balsam et al. 2011).

Such racial problems may also blend into the practices of gay and lesbian rights organizations. In fact, gays and lesbians of color claimed they often feel like a “token” racial minority at LGB events (Balsam et al. 2011), worry that LGBT activism is a betrayal to their race (Moore 2010), and fear that most wings of the gay and lesbian rights movement minimized, or sometimes reenacted, the practices of racism (Alimahomed 2010; Levitsky 2007; Ward 2008).

The consequences of perceived racism on activism for gay and lesbian rights are far from clear. Some studies conclude that sexual minorities of color are less inclined than White sexual minorities to participate in LGB groups (Barrett and Pollack 2005; Meyer, et al. 2008; Taylor et al. 2009). However, two other studies found that the “dual oppressions” of being a racial and sexual minority may not diminish one’s commitment to end heterosexism. Levitsky (2007) argued that such criticism by activists of color did not always diminish their commitment to gay and lesbian rights, as many sexual minorities of color prioritized raising the racial consciousness of White gays and lesbians. Lastly, White (2006) found that lesbian and gay African Americans are more likely join a feminist protest than heterosexuals from the same race.

Belief Systems as Collective Action Frames

Frames are generally conceived as cultural tools or schemas that provide “tacit theories about what exists, what happens, and what matters” (Gitlin 1980, p. 6). While frames help with the classification and organization of incoming stimuli, they also serve a political role of sanctioning or challenging conventional norms. Conservative frames acquire the consent of the less powerful by portraying the social order as proper, normal, and inevitable. Through the internalization of these frames, power elites get people to subscribe to values, ideals, and self-definitions that bind them to their location in the prevailing power structure. While conservative frames prioritize deference to conventional standards, collective action frames do the exact opposite, as they motivate people into joining collective efforts that publicly seek social change.

Movement theorists have identified four dimensions of frames that inspire activism (Ashmore et al. 2004; Duncan 1999; Kelly and Breilinger 1995; Klandermans 1997; Miller et al. 1981). First, collective action frames initially render some societal norms as wrong, unacceptable, and unjust. Second, frames identify the causes of the injustice. By providing a diagnostic function, frames are etiologies that explain why problems exist and assign levels of blame or capability to different entities. Third, frames also convince bystanders that they should use political tactics to stop these violations. These prognostic aspects of frames usually emphasize the urgency of political action and a sense that challenges from less powerful constituencies can force concessions from a reluctant target (this confidence in movement tactics is sometimes called “agency” or a “sense of collective efficacy”). Finally, frames must provide a collective identity among the aggrieved. In doing so, collective identities establish social boundaries of “us” and “them” by specifying who belongs to the righteous in-group of the mistreated and who exemplifies the antagonistic wrongdoers who must be challenged. These collective identities often contest and refute societal claims that members of their group are inferior, worthless, sick, or maladjusted. Instead, collective action frames offer narratives about the virtues of similar people and they suggest that their group is illegitimately threatened, deprived, or treated badly.

Studies on gay activism have mostly addressed injustice frames through personal experiences of discrimination and prejudice (rather than recognition of general heterosexist patterns in general). In many cases, sexual minorities may become aware of heterosexism by observing or hearing about the mistreatment of other LGB people (Evans and Herriott 2004). In other cases, enduring personal experiences of discrimination can delegitimize conventional norms and lead to an oppositional consciousness that challenges the status quo. The circumscribed, face-to-face nature of experienced discrimination makes it more proximal and salient than institutionalized forms of biases. This immediacy may create impulses to challenge this hardship, yet these impulses may be curbed or suppressed because heterosexism insists that sexual minorities should be passive, accommodating, silent, or self-hating. This combination of suppression and discrimination can foster a number of detrimental coping mechanisms such as disengagement (Wilson and Yoshikawa 2004), withdrawal or avoidance (Thompson 2006), greater suicide and drug risk (Mays and Cochran 2001), dangerous sexual practices (Wilson and Yoshikawa 2004), and stigmatizing others (Swim and Thomas 2006).

Although surviving discrimination can lead to negative consequences, cases of first-hand discrimination can encourage gay and lesbian activism (Jennings and Andersen 2003; Hyers 2007; Taylor et al. 2009; Waldner 2001). Two studies on AIDS activism found that gay men were more likely to protest governmental policies when they were demeaned by the medical professionals (Jennings and Andersen 2003; Tester 2004). Also, gays and lesbians who dealt with sexual and verbal harassment, or discrimination in housing and employment, were more likely to join radical gay rights groups (Friedman and Leaper 2010; Rollins and Hirsch 2003; Simon et al. 1998; Waldner 2001).

When concentrating on the “prognostic” tasks of collective action frames, the role of power interpretations in political activism is far from settled. In a number of studies, perceptions of personal efficacy (Hritzuk and Park 2000; Leighley & Vedlitz 1999), and/or collective efficacy, can be crucial to activism (Bobo and Gilliam 1990; Barkan et al. 1995; Coffé and Bolzendahl 2010). Conversely, some studies insist that a sense of efficacy may have little to do with participation in the women’s movement (Kelly and Breilinger 1995), liberal activism (Schussman and Soule 2005), antinuclear protests (Opp 1990), and youth movements (Paulsen 1994).

Efficacy perceptions may be related to the protesting inclinations of sexual minorities (Jennings and Andersen 2003; Jones 2002). For example, Swank and Fahs (2011) noted that gays and lesbians joined more protests when they thought that people like themselves could influence governmental policies and Jennings and Andersen (2003) suggested that gay men were more likely to join AIDS advocacy groups when they thought they had a good understanding of political issues. However, perceptions of political efficacy were irrelevant to how often married gays and lesbians joined gay right groups or attended a political demonstration (Taylor et al. 2009), and the only quantitative study of gay and lesbian electoral activism noted that “power expectancies” worked in the exact opposite ways as predicted (Waldner 2001). That is, sexual minorities were more likely to join a gay and lesbian political campaign when they thought the government was unresponsive to gay or lesbian demands.

To address the next task of “shaping collective identities,” the empirical literature has concentrated on relationships between interpretations of the self and attitudes toward other sexual minorities. Advocacy on behalf of oneself and others is often interwoven with issues of closeness to oppressed groups and moral obligations to work for social change. To address moral obligations, gays and lesbians were more likely to attend political demonstrations when they saw themselves as someone who challenges unjust laws (Swank and Fahs 2011). When discussing solidarity with the sexual minority community, three studies suggest that sexual minorities who liked and respected other sexual minorities joined more protests (Friedman and Leaper 2010; Gould 2002; Simon et al. 1998). Also, sexual minorities wanted to be more politically active when they fully embraced their sexuality identity (Konik and Stewart 2004), called themselves “queer” (Rollins and Hirsch 2003), and liked being a part of the gay and lesbian community (Konik and Stewart 2004). Similarly, hiding one’s sexual identity stunted the political involvement of gays and lesbians (Gortmaker and Brown 2006; Lewis et al. 2011; Waldner 2001).

Issues of concealing one’s identity and “passing-as-straight” are inevitably linked to gay and lesbian rights activism. Within a movement that tries to gain the recognition for new social identities, the very act of being “out” challenges the veracity of compulsorily heterosexuality (Bernstein 1997). The public revealing of sexual minority identities breaks the veil of silence and disrupts the myth that everyone is exclusively heterosexual. Public acknowledgements of an individual’s sexual identity are often crafted through a complicated set of disclosures practices. Some sexual minorities may feel comfortable revealing their sexual identity in most settings while others may only hint at their sexual identity to a few confidants or nobody at all. Disclosures of sexuality are often strategic, but gays and lesbians who frequently reveal their sexual identity are probably more inclined to publicly challenge heterosexism.

On top of embracing a sexual identity and bonding with a stigmatized group, collective identities can relate to how people display a desired or idealized self, or how they conceptualize leading a “principled life” (Kelly and Breilinger 1995; Oliver 1984; Opp 1990). For people who internalize activist identities, political engagement can become an opportunity to express key moral convictions and to act upon obligations of reciprocity, fairness, and concern for the common good. In highlighting the importance of activist self concepts, Hyers (2007) study noted that lesbians challenge homophobic comments more often when they embraced the activist norms of “standing up for what’s right” and defending the rights of subordinated groups.

From this research into the effects of social statuses on gay and lesbian activism, several hypotheses were generated. In light of previous research we tentatively hypothesized that: Hypothesis 1) Euro American gays and lesbians will engage in more electoral activism than African American, Asian American, Native American or Latino(a) gays and lesbians; and Hypothesis 2) Euro American gays and lesbians will do less protesting than African American, Asian American, Native American or Latino(a) gays and lesbians. Based on the framing literatures, these predictions were tested: Hypothesis 3) LGB activism will be more common among gays and lesbians who are exposed to more discrimination, divulge their sexuality more frequently, have internalized activist identities, and envision greater collective efficacy for the LGB community.

Method

Participants

This online study drew upon a sample of 285 self-identified gays and lesbians from throughout the United States (December 2007). Online surveys are often the best option when studying sexual minority populations (Moradi et al. 2009; Riggle et al. 2005) for several reasons: First, national random samples often fail to ask questions about sexual orientation. Second, online surveys make it easier to implement quasi-experimental research designs. Or, in more technical terms, the internet allows researchers to implement a “case–control” design that starts with non-equivalent comparisons on the dependent variable and collects retrospective data on the independent variable (Babbie 2008). By using political and apolitical listervs to find respondents, we were able to create comparison groups of gays and lesbians who were involved in activist and non-activist social circles. This use of nonequivalent comparison groups has never been done before in studies on gay and lesbian political participation. Previous studies of gay and lesbian activism collected their data through snowball samples of activists (Jones 2002; Tester 2004; Waldner 2001), mailing surveys to members of gay and lesbian organizations (Rollins and Hirsch 2003; Simon et al. 1998; Sturmer and Simon 2004) or distributing surveys at political events (Elbaz 1996; Lombardi 1999; Waldner 2001).

Respondents were selected through a purposive stratified sample of several email listservs. The first stratum included two listservs of members in gay and lesbian rights organizations. These political listservs were run by the umbrella group “Fairness Alliances.” These coalitions of political and human service organizations seek equality for sexual minorities by encouraging leadership development, public education, and participation in the democratic process. Memberships in these email groups were free and most of their participants resided in Midwestern and Mid Atlantic states, with the largest contingencies from Kentucky, Ohio, Indiana, Tennessee, and West Virginia.

The second stratum served as a comparison group of gays and lesbians involved in less politically-engaged networks. The researchers selected Yahoo groups that met three criteria. First, the group had to exist for explicitly social purposes (i.e., they did not mention anything political in the description of their listserv). Some of these groups concentrated on hobbies (e.g., “Dykes on Bykes,” “Gay Square Dancers,” or “GLBT Horselovers”) while others displayed support group qualities (e.g., “Lavender Mothers” or “Kentucky Pride”). Second, we excluded groups that served as romantic or sexual match-making sites in order to avoid biases toward single participants. Finally, we looked for groups that mentioned the regions that were most common in the political listservs (e.g., Queer Kentucky, Rural Pride of Tennessee, or Gay in Ohio).

The cover letters sent via the listservs asked potential respondents to click over to a surveymonkey web site. The letter solicited the involvement of adults who considered themselves gays or lesbians. In addition to the standard discussion of anonymity, voluntarism and informed consent, we described the educational and professional backgrounds of the primary investigators. The response rate to this letter was impossible to calculate because we cannot estimate the number of people who belonged to each listserv, but 285 of the 296 who clicked on the survey actually completed the survey.

The sample of 285 participants included more men (58 % gay men) and mostly White participants (79 % Euro American, 7 % Native American, 2 % African American, 1 % Asian American, 1 % Latino(a)s, and 10 % “refuse to answer”). When addressing the intersectionality of gender and race statuses, 74 % of the women were Euro American and 83 % of the men were Euro American. Ages in the sample spanned a wide range, from age 18–75, with 24 % under age 30, 54 % ages 30–50, and 22 % ages 51–75. The sample included a diverse array of incomes, including 10 % below $20,000 per year, 27 % $20,000–50,000 per year, 25 % 50,000–80,000 per year, and 31 % over $80,000 per year, with 8 % being missing data. Similar to most samples of “out” gay and lesbian participants, our sample was highly educated, with 3 % having earned a high school degree, 58 % having some college or a bachelor’s degree, and 32 % having a graduate degree. Participants tended to be distributed in many types of urban and rural spaces, with 26 % residing in a large urban center, 18 % residing in a suburb of a large city, 18 % residing in a mid-sized city, and 32 % residing in smaller towns or rural areas. Because the majority of recruitment took place in the geographic South of the United States, 65.2 % of participants lived in the South, 15.7 % lived in the Midwest, 6.2 % lived in the West, and 4 % lived in the East, with Kentucky, Ohio, Indiana, Tennessee, and West Virginia representing the most respondents. Unfortunately, due to conditions of anonymity set by the IRB, it is impossible to estimate possible selection biases found in the different listservs.

Measures

The anonymous online survey contained 88 close-ended items. The items on self-perceptions or perceptions of the gay and lesbian community were mostly measured via a 5-point scale (strongly disagree to strongly agree). Demographics and experiences of discrimination were handled through more idiosyncratic scales. For example, respondents were offered the responses of never, once, or twice or more to the item: “Because of your sexual orientation, have you ever had verbal insults directed your way since you were 16-years-old?”

Political Activism

In democratic societies, citizens can choose between a range of conventional and unconventional political behaviors. One channel is when citizens express their concerns to governmental leaders through the orthodox or “insider” avenues of electoral politics (e.g., testifying at legislative hearings, writing letters to Congress, or making political donations to favorite politicians). Compared to other tactics, these approaches take less time and effort and require little risk. Citizens can also draw the attention of authority figures through the “outsider” and confrontational means (e.g., engaging in protests, strikes, boycotts or civil disobedience). As a whole, outsider tactics require greater efforts to join and can be more risky than electoral activism (e.g., outsider tactics can lead to arrest, chastisement, ridicule).

To match these insider and outsider tactics to the gay and lesbian rights movement, we provided four measures of gay and lesbian rights activism. After creating four single-item dependent variables, participants were asked if they had ever voted, written a letter, attended a demonstration, or engaged in civil disobedience on the behalf of gay and lesbian rights. Activism items were answered in a binary yes or no format (yes = 1, no = 0). In total, 83 % of participants had voted for gay rights, 76 % had signed a petition, 41 % had been to a demonstration, and 6 % had engaged in civil disobedience for the cause of expanding the rights of sexual minorities.

Gender (Men)

Answers to “What is your gender” were coded in a binary fashion, from 0 (lesbian women) to 1 (gay man).

Race (Euro American)

The responses of “Please identify your race” were coded into a single dummy variable. Euro American/White received a one, while African American, Asian American, Native American, and Latino(a) all received a zero. The separate racial minorities were collapsed into non-White category due to the small number of people in several of the racial classifications. All of the “refused to answer” responses were not assigned any race.

Economic Bias Experiences

Sexual prejudices are expressed and reinforced through the practices of mainstream institutions. While every institution privileges heterosexuality, the consequences of employment and housing discrimination are especially important (i.e., access to key resources). Accordingly, we used Herek’s (2009) study to pose two questions about how often participants experienced discrimination at the workplace or in a residential setting in their lifetime. As an additive scale where participants responded as “never,” “once,” “twice or more,” respondent answers ranged between 2 and 6 (Cronbach ά = .78).

Hate Crime Experiences

Herek’s victimization scales (2009) established the extent of hate crime violence in participants’ lifetimes (similar to Waldner 2001). Respondents were asked three questions about the frequency with which they have been physically or verbally attacked because of their perceived sexual identity. One question read: “How often have people directed verbal insults at you because of your sexual identity?” Another one asked how often the respondent was “punched, hit, kicked or beaten because of your sexual identity?” Individual items had scores ranging from 1 (never) to 3 (two times or more), while the composite scale went from 3 to 9, with higher scores indicating a greater degree of victimization (Cronbach ά = .66).

Disclosure of Sexual Identities

To see how people managed the public disclosure of gay and lesbian identities, we chose an item from the public identification subscale of the Lesbian Internalized Homophobia Scale (Szymanski and Chung 2001). Our item focused on issues of disclosure or the willingness to publicly express a gay or lesbian identity: “I try not to give any signs that I am gay or lesbian.” Scores were coded in the direction of more disclosure as responses ranged from 1 (strongly agree) to 5 (strongly disagree). Note that the cover letter asked for volunteers who were self-identified gays and lesbians, so the question itself was not a form of disclosure to the researchers.

Perceived Collective Efficacy

Collective efficacy judgments are future-oriented expectancies about the likelihood of a group achieving its intended goals. When focusing on group potency, our measure addressed the perceived collective capacities of the gay and lesbian community: “When gays and lesbians work together, they can solve the problems facing them” (see Yeich and Levine 1994). Responses ranged from 1 (strongly disagree) to 5 (strongly agree).

Activist Identity

We focused on personal self-designations to determine if people had a solid and long-term commitment to social justice (Kelly and Breilinger 1995; Oliver 1984; Opp 1990). By modifying an “activist commitment” measure from the Feminist Identity Development scale (Bargad and Hyde 1991), our item asserted: “It is important for me to fight for gay and lesbian rights.” Responses ranged from 1 (strongly disagree) to 5 (strongly agree).

Results

Group Differences and Correlational Data

To start our analysis, we conducted a number of tests of group differences for our gender and race variables. Table 1 displays the means, standard deviations, and frequencies for every variable in the study (gender by race figures presented). Chi-squares (χ²) were performed to examine race and gender differences among the dichotomous political action variables (voting, petition signing, protesting, and civil disobedience). When exploring the within-gender differences by race, only two Chi-squares were significant. For lesbians, race showed significant differences for signing a petition χ² (1, 98) = 4.24, p < .05, and going to a protest χ² (1, 98) = 8.27, p < .01. For gays, race showed offered no significant associations with any of the political activities, thus suggesting that the effect of race on petition signing and protesting seems contingent upon a sexual minority’s gender.

Table 1 Descriptive statistics for gender by race subsamples (n = 285)
Full size table

To address the gender and race differences among the continuous variables, a multivariate analysis of variance (MANOVA) was conducted with economic discrimination, hate crimes, concealed sexuality, collective efficacy, and activist identity as the dependent variables. To study this gender by race intersections in greater detail, we conducted separate MANOVA for each gender group on the effect of race for economic discrimination, hate crimes, concealed sexuality, collective efficacy, and activity identity. The overall multivariate effect of race was never significant for either gays or lesbians and none of the within-gender univariate effects indicated significance for race.

To extend our intersectional analysis, bivariate relationships between the independent and dependent variables were ascertained through a series of Pearson product–moment correlations for gay and lesbian subsamples (see Table 2). When addressing significant associations with the dependent variables, race was significant for the petition signing and protesting actions of lesbians but not gays; that is, lesbians of color were significantly less likely to join gay and lesbian rights protests than White lesbians.

Table 2 Pearson correlation matrix
Full size table

To address other predictors, exposure to heterosexist discrimination in the workplace significantly increased the tendency for gays to sign petitions and protest for LGB rights, but economic discrimination had no significant relationships for lesbians. Elevated petition signing, and protest behaviors was also significantly linked to hate crimes for gays and lesbians, but hate crime experiences was only linked to more voting for lesbians but not more civil disobedience for either gender. Every political action except civil disobedience was significantly more common among gays who routinely divulged their sexuality, while concealment only influenced the petition signing and protest actions of lesbians. Activism was consistently elevated when gays and lesbians expressed high levels of perceived gay and lesbian power. Finally, the internalization of activist identities was important for each of the political action of gays but none of the lesbians. Finally, the shifting significance of these framing factors supports the intersectional assumption that the relative importance of different collective action frames often changed between the genders.

Binary Logistic Regressions

The simultaneous analysis of our demographic and framing variables was derived from several binary logistic regression analyses. Logistic regressions were well suited at analyzing dichotomous dependent variables as they calculated a likelihood estimation of a certain event occurring (Hosmer and Lemeshow 2000). Logistic regressions are also well suited for our data because their use is not confined to many of the strict requirements other sorts of regressions (i.e., a normal distribution in the dependent variable or no problems of homoscedasticity). As expected, the data met all the conditions for a logistic regression, in that the outcome variable was coded in categorical binary fashion, linear relationships were assumed between the independent variables and the logit of the dependent variables and there was an absence of outliers or high multicolinearity. Multicolinearity diagnostics suggested low variance inflation factors (VIF) in each regression because the VIF was below 1.53 for every independent variable. Due to a pairwise treatment of missing data, any person who failed to answer an item was automatically removed from the regressions in this study.

The following tables contain the intersectional analyses for the four dependent variables of voting, petition signing, protesting, and engaging in civil disobedience. To accomplish the intersectional analysis, we split our sample into two gender subgroups (women and men). For each regression, the variables were simultaneously entered in the formulas and participants who had missing data on the gender and political action variables were excluded from the regression.

Table 3 presents the findings for electoral activism (voting or signing petitions for gay and lesbian rights). With four regressions in the table, the statistical columns display the results of gender specific sub-samples. We initially explored the significance of the entire model and later discussed the significance of specific individual variables. This table is designed to test the hypotheses that Euro Americans and gay men do more electoral activism than lesbians and sexual minorities of color (even after controlling for framing factors). It also explores the associations between electoral activism and racial background, experiences of heterosexist discrimination, sexuality disclosure patterns, perceptions of LGB efficacy and an activist self-concept among gays and lesbians.

Table 3 Logistic regression estimates of voting and petition signing for gay and lesbian rights (n = 263)
Full size table

When addressing the combined strength of the demographic and framing factors in Table 3, the full model’s goodness-of-fit χ² was significant for every regression except for gay men’s voting behaviors and lesbian women’s petition singing. The full model was able to account for between 32 % and 9 % of the variance in petition actions with model being the weakest for voting actions of gay men. Although the model could explain 32.4 % of the variance for voting among lesbians, the model presented the meager R² of .09 for gay men.

Specific race and framing variables displayed some interesting findings. Hypothesis 1 predicted that Euro American gays and lesbians would do more electoral activism that gays and lesbians of other races. Being Euro American significantly augmented the voting tendencies of lesbians (B = .67, p < .05) but was not as crucial for gay men. Hypothesis 3 predicted greater activism when gays and lesbians experienced more discrimination, publicly shared their sexual identity, internalized activist identities, and imagined greater power among the LGB community. When looking at the framing factors, the salience of specific factors often fluctuated for the voting behaviors of gays and lesbians. Only twice did the framing factors significantly predict the voting patterns of lesbians while these factors never significant for gays. Collective efficacy was especially important for voting among lesbians (B = 1.33, p < .01) but enduring hate crimes also cultivated more voting actions among lesbians (B = .39, p < .05). Every other framing factor was unable to reject the null for either gays or lesbians.

In columns 4 and 5 the analysis of petition activities was completed for gender subsamples. Being White had a much larger impact on the petition tendencies of lesbians than gays (B = 1.45, p < .05). Like in voting, the relative potency of the framing factors for petition signing was partially contingent upon the gender of survey participants. Disclosure patterns had bigger repercussions for gay men (B = .74, p < .01) than for lesbians. Similarly, activist identities (B = .63, p < .05) and hate crime victimization (B = .31, p < .05) mattered for gay men but were less crucial for lesbians. Gender also intervened in the importance of power interpretations as perceptions of efficacy were only significant for lesbians (B = .79, p < .05) but not for gay men. The last framing factor of economic discrimination was never significant for either gender.

Table 4 offers the logistic analyses for the “non-electoral” behaviors of protesting and engaging in civil disobedience. In two of the four computations, the combined effects of the variables were significantly related to those who did and did not engage in these political actions. The model was more adept at explaining protesting actions because the R² for attending demonstrations ranged between .48 and .33 (all χ² were significant at p < .05). Conversely, the same factors mustered the smaller R² of .27 to .21 for involvement in civil disobedience for LGB rights.

Table 4 Logistic regression estimates of protesting and civil disobedience for gay and lesbian rights (n = 265)
Full size table

When addressing specific variables race was a crucial antecedent to protesting. Net of framing influences, the significance of racial status remained intact but the importance of racial ancestry was more important for gays than lesbians (B = 1.77, p < .01; .74 p < .05). Similar to the bivariate tables, the direction of the association ran against the argument that African Americans and Latino/as are more inclined to protest than Euro Americans (rejecting hypothesis 2).

In corroborating the assumptions of intersectional theories, the relative importance of framing factors changed somewhat in the gender subgroups. For lesbians, surviving hate crimes and collective efficacy interpretations were the best predictors of joining LGB demonstrations (B = .44 and .69, p < .05). However, lesbian protest inclinations were not connected to personal experiences of economic discrimination, disclosure practices, or the internalization of activist self-concepts. In contrast to lesbians, protest involvement for gays was not related to perceptions of gay power and surviving hate crimes. Instead, the likelihood of gay men joining a LGB protested was enhanced by greater disclosures of sexual identities (B = 1.05, p < .001), embracing an activist self-definition (B = 1.24 p < .001), and enduring of economic disadvantages (B = .54, p < .05).

Our model was at its weakest when predicting engagement in civil disobedience. The blended effects of demographic and framing factors could only account for about one-fourth to one-fifth of the variance and none of the models significantly accounted civil disobedience activities. For the specific variables in the general sample, race was never significant (rejecting the hypothesis that predicted greater protesting among African Americans). When addressing framing factors, only some parts of hypothesis 3 were substantiated. When exploring the intersectional aspects of collective actions frames, the importance of efficacy appraisals (B = .57, p < .05) was much more pronounced among lesbians, while activist identities were especially crucial to gay men who engaged in civil disobedience (B = 2.16, p < .001). Notably, gender barely modified the impact of economic discrimination, hate crime encounters, and disclosure practices but issues of power interpretations displayed opposite directions for gays and lesbians.

Discussion

This study analyzed the ways gender and racial statuses, as well as five framing process, were connected to four forms of gay and lesbian rights activism among a national sample of self-identified gay and lesbian individuals. To do so, this study used logistic regressions to locate important factors within the total sample and also utilized an intersectional analysis to determine if demographic and framing antecedents worked the same way for gays and lesbians. The intersectional analysis tested the assumption that this crossing or sharing of social statuses established a distinct set of political behaviors for the sexual minorities from each combination of interlocking social positions.

This study demonstrates the value of combining demographic and framing factors into an integrative theoretical model. Four of the 8 regressions found significant χ² for the model and the combination of variables routinely explained between 9–48 % of the variance in political behavior outcomes. The model had its greatest potency when predicting protest activities and it was weakest at predicting civil disobedience behaviors. Although the same size was not equivalent for the gender subsamples, the goodness-of-fit calculations were crucial for the voting and protest inclinations of gays and lesbians. This study also validated that an intersectional framework was beneficial to nuance the findings, as the power of many of the predicator variables seemed contingent upon the genders and races of the sexual minorities in this study.

The data offered many insights into the ways that gender and racial identities influenced the enactment of gay and lesbian rights advocacy. The division the sample into gender subsamples revealed that LGB activism was sometimes governed by different factors for gay men and lesbian women. Hence, the connections between political habits and gender, race, and framing factors were far from universal because the combinations of different race, gender, and interpretative configurations produced different activist inclinations within segments of the LGB community.

Patterns of how race and gender influenced gay and lesbian rights activism were somewhat complicated and idiosyncratic. In general, Whites were disproportionably overrepresented in the gay and lesbian rights movement. However, the impact of Whiteness was mostly connected to how race affected lesbian involvement in voting, petition signing, and protest. That is, White gay men rarely differentiated themselves from gay men of other races, but lesbian women of color were much less inclined to do three of the four political actions. While this intersection of race and gender was noteworthy, these results demand further analysis. Ensuing studies should construct stratified or purposive samples that insure greater racial diversity within the sample. With such an approach, it would be easier to ascertain the possibly changing relationships between activism, gender, and framing factors for sexual minorities with African American, Asian American, Latino(a), and Native American ancestries.

While it is possible that our findings may overstate the Whiteness of LGB movements, we do not think it is simply a consequence of methodological decisions. Several qualitative studies suggest that LGB movement often prioritizes the demands of gay men over lesbian women and White sexual minorities over sexual minorities of color (Alimahomed 2010; Levitsky 2007; Ward 2008). Future research should more clearly determine the mechanisms behind these different participation rates among racial groups. Perhaps gay and lesbian identities were less relevant to the lives of Latino(a), Asian, Native American or Black sexual minorities, as racial identities had more salience (Fingerhut et al. 2005). As such, perhaps sexual minorities of color spend most of their political energies fighting against racism, or perhaps a combination of homophobia within racial minority communities and the racism within the gay and lesbian rights movement present a doubly marginalized reality that inhibits activism. Differences in political participation could also be related to more macro factors. The entrenched nature of racial segregation probably reaches into the living and work arrangements of sexual minorities. In turn, distinct gay and lesbian communities may be somewhat broken into isolated racial enclaves that block important discussions about common problems and the ways to coordinate multiracial political mobilizations.

In general, this study also confirmed most of our hypotheses about specific collective action frames. In the intersectional analysis, types of discrimination were highly dependent on the gender of the target. Exposure to economic biases was associated with activism for both genders but the link between protesting behaviors and economic hardship was much higher for gay men than for lesbians. Perhaps gay men are more outraged by economic discrimination because they are accustomed to the hidden benefits of workplace sexism (e.g., greater pay, better chances of advancement, not being channeled into pink-collar jobs). It is also possible that lesbian women express their disdain for workplace discrimination more through the women’s movement. Hate crimes also presented some varying effects, in that enduring homophobic attacks was an especially strong predictor of voting and protest actions of lesbians but only the petition signing of gay men. This may speak to the gendered expectations of how men and women are expected to respond to stranger violence. When experiencing brutal attacks, men may have become more accustomed to handling violence through individual means (e.g., fighting back or being too embarrassed to admit that they were injured by physical violence). Conversely, women are typically told to be silent about their abuse, but lesbians who challenge public acts of violence may be more comfortable with collective responses than gay men.

The data suggests that there are occasionally political consequences of being a “closeted” or “out” sexual minority. This analysis suggests that the effect of disclosure patterns was contingent upon a person’s gender. Although “out” lesbians were drawn more to activism than “closeted” lesbians, this factor only displayed significant results for the protesting and petitioning actions of gay men. Again, future research should try to explain why disclosure patterns may have different effects on gay men and lesbian women.

The last two framing factors highlight the value of doing intersectional analysis. However, the significance of each factor was intricately patterned along clear gender divisions. Regardless of the political behavior, believing in the collective efficacy of sexual minorities always led to greater amounts of activism among lesbians. For gay men, however, levels of efficacy never significantly linked to any political action. Likewise, the salience of activist identities was not equally applicable to both genders. Visualizing oneself as a champion of gay and lesbian rights had a positive relationship with petition signing, protesting, and civil disobedience for gay men. In sharp contrast, activist identities were never significant for lesbians, and higher levels of this factor had an inverse relationship with protesting among lesbians.

Strengths and Limitations

This study offers some theoretical and methodological rigor. By exploring theoretical breadth, this is the first study to quantitatively use an intersectional analysis when predicting gay and lesbian rights activism. This was important because the relationships between variables were often contingent among the different combinations of gender and race variables. Our choice of four dependent variables added greater specification, as some variables were better suited for either electoral or protesting actions. Our creation of a national sample was unique because all of the previous studies on gay and lesbian rights activism have been confined to a single region of the United States. Finally, our sampling of political and apolitical listservs allowed for a clear analysis between activists and non-activists. Most studies on this topic restrict their analysis to only people who have been involved in a local mobilization for LGB rights.

When moving to limitations, the findings of this study should not be considered universal because race dynamics, gender norms, and the acceptance of sexual minorities often vary by country (Adamczyk and Pitt 2009). Scholars may also ask why social class measures were not included in the intersectional part of the study (Taylor et al. 2009). The omission was due to a problem of insufficient cells sizes for a regression and the recognition that studies often find no associations between income and political activism among sexual minorities (Elbaz 1996; Swank and Fahs 2011). For American populations, quasi-experimental designs of this sort have non equivalent comparison groups and are sometimes prone to errors of selection bias. For example, we have an unequal proportion of activists and non activists and the study included a higher percentage of men than women (unfortunately do Human subject constraints we were unable to determine in any of the listserves added any large selection biases) This study also did not include self-identified bisexuals or transsexuals in the sample as there is evidence that bisexuals are less inclined to protest for LGB rights (Herek et al. 2010; Lewis et al. 2011). Moreover it is possible that this study could have different findings if we focused on “same-sex sexual contact” rather than self-identities in defining the population. Some studies claim that African American and Latino “men who have sex with men” are more reticent when embracing a gay identity than White men who have the same sexual practices (Barrett and Pollack 2005).

Measurement errors are also possible. Problems of overdemanding recall could lead to inaccurate results in that people may have difficulty remembering whether they experienced discrimination or engaged in any political activities. Questions of concealed sexual identities may elude generalities because gays and lesbians are rarely totally “out” or “closeted” to everybody they meet. Our close-ended responses for gender and racial statuses might have forced people into exclusive categories that were not totally accurate (e.g., transgendered people or the variety of different groupings within “people of color”). Questions of social desirability may have been especially relevant to our activist identity measures, as participants may have wanted to affirm themselves by overstating the amount that they fight for social justice. Finally, there could be problems of temporal ordering for all of the framing variables, as people may embrace an activist identity or display themselves as “more out to others” after joining the gay and lesbian rights movement.

In total, this study offered unique insights into the virtues of intersectional analyses of race, gender, framing, and sexual variables. For example, the intersectional analysis discovered that White lesbians were more likely to vote, sign petitions, and protest than women of color. Conversely, race only mattered for the protesting actions of men. Framing factors also revealed the benefits of intersectional studies. Although activist identities, economic discrimination, and perceptions of efficacy were often crucial in the larger sample, the relative strength of these factors fluctuated between the genders. The strongest predictors of activism for gay men included economic discrimination, disclosure patterns, and activist identities, while experiencing hate crimes and perceptions of collective efficacy were more vital to lesbian activism.

Accordingly, this study offers insights about gay and lesbian activism to researchers and the general population at large. For researchers, this study suggests that political participation among sexual minorities is complex, and intersectional approaches are helpful in sorting through such complexity. For activists and concerned citizens, gay and lesbian activism is clearly driven by different factors that may require different strategies to recruit and nurture a vibrant LGB activist community. This study should serve as a warning that sexual minorities of color are less affiliated with the gay and lesbian rights movement and that sexual minorities and their allies must make greater efforts to diminish divisive racial divisions. In particular, men, and Whites must better recognize and challenge their institutionalized privileges in order to make a stronger and more inclusive LGB movement. In doing so, they can guard against the cruel irony of progressive social movements replicating many of the oppressive practices they seek to challenge and subvert.