Introduction

Over the past several years, developments in public opinion, politics, and policy suggest that attitudes toward the gay and lesbian community have become increasingly positive. For example, in 2011, a majority of Americans supported same-sex marriage for the first time (Newport 2011). Additionally, an increasing number of states have adopted policies to protect lesbian, gay, bisexual, and transgender (LGBT) employee workplace rights (Everly and Schwarz 2015). In the military, the controversial “Don’t Ask, Don’t Tell” policy was repealed so that gay and lesbian military personnel can openly disclose their gay identities. In 2012, Barack Obama announced his support of same-sex marriage marking the first time the President of the United States has publically supported the issue. Three years later, the United States Supreme Court ruled in Obergefell v. Hodges (2015) to legalize same-sex marriages. Finally, organizations have also begun to adopt more progressive policies for LGBT employees, creating workplaces that welcome and embrace diversity (Button 2001; Everly and Schwarz 2015; Johnston and Malina 2008).

As public opinion toward homosexuality becomes increasingly positive and more organizations create environments where gay and lesbian employees feel safe and supported, gay and lesbian employees will feel more comfortable disclosing their identities at work. Research examining gay identity disclosure in the workplace has found that gays and lesbians are less fearful and more likely to disclose their identities in organizations with supportive policies and cultures (Ragins and Cornwell 2001; Ragins et al. 2007). Openly, gay and lesbian employees are also less likely to experience discrimination in organizations with supportive policies (Button 2001). However, despite the shift in attitudes toward homosexuality and the prevalence of gays and lesbians choosing to disclose their identities at work, few studies have examined whether attitudes toward gay men and lesbians have changed and whether these attitudes impact perceptions of gays and lesbians in the workplace.

Although attitudes toward gays and lesbians are changing, most social psychological studies looking at perceptions of gays and lesbians were published many years ago and are not necessarily positive. For example, a meta-analysis of over one hundred studies examining attitudes toward gays and lesbians concluded that attitudes toward gays and lesbians were negative, yet deemed acceptable in American society (Kite and Whitley 1996). These studies included a wide range of samples from college students (Herek 1984, 1986; Kite 1994) to national surveys (Herek 1991; Herek and Glunt 1993) and the results consistently supported the conclusion that anti-gay attitudes were prevalent in America (Kite and Whitley 1996). However, these studies also found that anti-gay attitudes were decreasing among younger cohorts (Kite and Whitley 1996) and women exhibited more positive attitudes toward gays and lesbians than men (Herek 1988; Kite 1994; Kite and Whitley 1996; LaMar and Kite 1998).

So how have perceptions of gays and lesbians changed over time? More recent studies suggest that attitudes toward gays and lesbians have become much more positive, particularly among women. For example, a recent study found that relative to male participants, female participants found gay and lesbian targets to be more likable (Cohen et al. 2009). Importantly, women’s ratings of likability were well above the midpoint of the scale, while men’s ratings were much closer to the midpoint (Cohen et al. 2009). Additionally, in the poll mentioned earlier in which the majority of respondents supported same-sex marriage, the support was stronger among women than men (Newport 2011). Finally, women report greater feelings of honesty and security when interacting with their gay friends (Grigoriou 2004). Taken together, this more recent work supports the argument that attitudes toward gays and lesbians are becoming increasingly positive, but that women more so than men are driving this change. However, some questions that remain unanswered are why women have more positive perceptions of gays and lesbians compared to men and do these perceptions influence judgments of gays and lesbians? Our study attempts to answer these questions by measuring perceptions of gays and lesbians applying for jobs in order to determine how gay and lesbian job applicants are perceived relative to heterosexual job applicants. We propose that stereotypes about gays and lesbians’ warmth and competence may lead men and women to evaluate gays and lesbians differently in the hiring process.

Hiring and Discrimination Against Gays and Lesbians

Previous studies looking at workplace experiences of gays and lesbians show that gay and lesbian employees still experience discrimination in the workplace (Croteau 1996; Myers 2000; Taylor 2002). Specifically, previous studies examining perceptions of gays and lesbians’ hirability in the United States have found that gay and lesbian job applicants are likely to experience discrimination in some form. For example, in one such study, researchers measured formal and interpersonal discrimination against confederates who applied for retail jobs while wearing hats that read “Gay and Proud” or “Texan and Proud” (Hebl et al. 2002). The study found confederates wearing a “Gay and Proud” hat were treated more negatively and with more disinterest from store employees (Hebl et al. 2002). In another study that involved manipulating the gender and sexual orientation of job applicant resumes, researchers discovered that relative to heterosexual men, gay and lesbian job applicants were deemed less hirable (Horvath and Ryan 2003). However, because these studies were conducted several years ago and perceptions of gays and lesbians are becoming more positive in recent times, particularly among women, it is possible that these studies no longer reflect the experiences of gay and lesbian job applicants. In fact, one of the only more recent studies that measured perceptions of gays and lesbians’ hirability found that male college students who were hiring a personal trainer were significantly less likely to hire a gay or lesbian trainer relative to heterosexual trainers, while female participants showed no difference in their perceptions of the trainers’ hirability (Cunningham et al. 2010). This recent study is consistent with the argument that women’s attitudes toward gays and lesbians are becoming increasingly positive and that these attitudes may also influence the perceived hirability of gay and lesbian job applicants.

Because previous work has shown that men are more likely to have negative attitudes toward gays and lesbians and are more likely to discriminate against gays and lesbians, we predict that heterosexual men will be less likely to hire gay and lesbian job applicants relative to heterosexual job applicants. On the other hand, because recent work has shown women are adopting more positive attitudes toward gays and lesbians, we predict that female participants will show the opposite pattern and be more likely to hire gay and lesbian job applicants relative to heterosexual job applicants.

Hypothesis 1

There will be a participant gender by applicant sexual orientation interaction such that women will be more likely to hire gay and lesbian job applicants relative to equally qualified heterosexual job applicants, while men will be less likely to hire gay and lesbian job applicants relative to equally qualified heterosexual job applicants.

Stereotypes of Gay Men and Lesbians

Just as attitudes are becoming more positive toward gays and lesbians, recent work suggests that stereotypes are becoming more positive also, particularly for women. We argue that these positive stereotypes could help explain why men and women may perceive gay male and lesbians job applicants’ hirability differently relative to heterosexual job applicants. For example, there are several positive stereotypes about gay men that women hold more strongly than men. Gay men are generally perceived to be good listeners, open with their feelings, warm in relationships, and tactful (Madon 1997; Taylor 1983). Heterosexual women have also reported that they feel better about themselves when they spend time with gay male friends (Barlett et al. 2009) and they trust the honest advice of their gay friends more than the advice from other women (Russell et al. 2013). Women also tend to score higher on a measure called the homopositivity scale, which measures positive stereotypes of gay men’s warmth including items like, “Gay men are more in touch with their emotions than straight men.” (Morris and Bearden 2007). However, there is much less research specifically examining women’s perceptions of lesbians and therefore little evidence that women perceive lesbians as warmly as gay men. Therefore, we predict women will have positive perceptions of gay men’s warmth relative to lesbians and heterosexual men and women.

Men on the other hand tend to score below the midpoint on the homopositivity measure, suggesting that on average, men disagree with positive stereotypes about gay men’s warmth (Morris and Bearden 2007). Furthermore, previous studies have consistently shown that men are more likely to have negative perceptions of both gay men and lesbians (Kite and Whitley 1996). While women tend to view gays and lesbians as a disadvantaged minority group, men tend to view gays and lesbians as violators of gender roles, which contribute to their negative perceptions (Herek 2000). Therefore, we predict men will have negative perceptions of both gay men and lesbians’ warmth relative to heterosexual men and women.

Hypothesis 2

There will be an interaction between participant gender, applicant gender, and applicant sexual orientation such that women will perceive gay male job applicants as warmer than equally qualified lesbian and heterosexual job applicants, while men will perceive gay male and lesbian job applicants as less warm than equally qualified heterosexual job applicants.

In addition to warmth, there are also stereotypes associated with gay men and lesbians’ competence that are particularly relevant in an organizational context. For example, there are several positive stereotypes about lesbians that might actually help lesbians be perceived as more competent in business contexts (Badgett 1996). In particular, compared to heterosexual women, lesbians are more likely to be perceived as career-oriented rather than family-oriented (Peplau and Fingerhut 2004). Additionally, lesbians are not assumed to be mothers as often as heterosexual women, which leads to the perception that lesbians are more committed to their jobs (Kite and Deaux 1987) and therefore, less likely to experience wage and hiring penalties associated with motherhood (Baumle 2009). Together, these stereotypes of lesbians contribute to the perception that relative to heterosexual women, lesbians are more independent, assertive, competitive, and self-confident, characteristics typically equated with competence and success in business contexts (Peplau and Fingerhut 2004). With respect to gay men, there is evidence that women, but not men, perceive gay men as being particularly competent. The homopositivity scale, which measures positive stereotypes of gay men, also includes items related to competence such as, “gay men are more articulate than straight men,” and “gay men are more intelligent than straight men” (Morris and Bearden 2007). Because women score highly on the homopositivity scale and may endorse positive stereotypes regarding lesbians in the workplace, we predict women will have positive perceptions of both gay men and lesbians’ competence relative to heterosexual men and women. However, because men score lower on the homopositivity scale and may only endorse positive stereotypes regarding lesbians’ competence, we predict men will only have positive perceptions of lesbians’ competence relative to gay men and heterosexual applicants.

Hypothesis 3

There will be an interaction between participant gender, applicant gender, and applicant sexual orientation such that women will perceive gay male and lesbian job applicants as more competent than equally qualified heterosexual job applicants, while men will perceive lesbian job applicants as more competent than equally qualified gay male and heterosexual job applicants.

Finally, because warmth and competence are the two universal dimensions of social perception upon which groups are judged (Fiske et al. 2002) and have been shown to predict a number of important outcomes, including hirability (Correll et al. 2007; Rudman and Glick 2001), we predict that for both male and female participants, perceptions of competence, and warmth toward gays and lesbians as an out-group will mediate the effect on hirability. In other words, to the extent that men and women show differences in their perceptions of gay and lesbians job applicants’ warmth and competence relative to heterosexual job applicants, these differences in perceived warmth and competence should explain men and women’s differential perceptions of job applicants’ hirability.

Hypothesis 4

Perceptions of warmth and competence should mediate the relationship between applicant sexual orientation and hirability.

Study 1

In Study 1, we first wanted to measure perceptions of job applicant hirability to determine whether men and women differentially evaluate gay and lesbian job applicants relative to heterosexual job applicants. To this end, we randomly presented participants with job applicant resumes differing only in the gender and sexual orientation of the applicant. We predicted that female participants would perceive gay and lesbian applicants as more hirable than equally qualified heterosexual applicants. However, for male participants, we predicted that men would perceive gay and lesbian applicants as less hirable than equally qualified heterosexual applicants.

Method

Participants

One hundred and ten participants were recruited from an online subject pool maintained by a large American west coast university. Members of this subject pool are non-student adults who were originally recruited through postings on the Internet. Members of the subject pool are located across the United States and represent a variety of occupations and backgrounds. Participant age ranged from 18 to 62 years (M = 32.93, SD = 10.48; 55 women, 55 men). The self-identified racial breakdown of the sample was as follows: 17 Asian, 4 Black, 76 White, 6 Latino, and 2 multiracial participants. Seven participants self-identified as non-heterosexual. Participants were given $3 for completing the online survey.

Procedure

Participants were told that a large west coast university needed their help evaluating job applicants for a vacant Office Manager position. Participants were then shown a resume that ostensibly belonged to one of the job applicants. In reality, participants were randomly presented with one of four possible resumes that differed only in the gender and sexual orientation of the job applicant. Thus, the experiment was a 2 (applicant gender: male, female) × 2 (applicant sexual orientation: gay, straight) × 2 (participant gender: male, female) between-subjects design.

To manipulate the gender of the applicant, the applicant’s name was listed as either “Greg Johnson” (male) or “Jennifer Lewis” (female). To manipulate the sexual orientation of the applicant, a professional organization was added to each resume that provided information about the applicant’s sexual orientation. For example, the gay male and lesbian resumes indicated the applicant belonged to “Los Angeles Gay Business Professionals.” The straight male and straight female resumes indicated the applicant belonged to “Los Angeles Business Professionals.” See “Appendix” for the resumes used in the gay male and straight female conditions. Similar manipulations for sexual orientation have been used before in previous studies using resumes (Ellis and Vasseur 1993; Horvath and Ryan 2003). Additionally, a pre-test of the sexual orientation manipulation showed that when asked to indicate the job applicants’ sexual orientation on a scale of 1 (straight) to 7 (gay), participants who viewed a gay or lesbian resume were more likely to identify the job applicant as gay (M = 5.57, SD = 1.63) relative to participants who viewed a heterosexual resume (M = 3.19, SD = 1.13), t(54) = −6.23, p < .001, η 2 = .42.

After reading their randomly assigned resume, participants evaluated the applicant’s hirability and answered demographic survey items and a manipulation check item. Finally, participants were shown a debriefing page that thanked them for their participation and explained the true purpose of the study.

Measures

Manipulation Check

To assess whether participants were attuned to the sexual orientation of the job applicant, participants completed one item: “What was the sexual orientation of the job candidate whose resume you evaluated in this study?”

Hirability

To assess the applicants’ hirability, participants were asked to complete three items used by Rudman and Glick (2001). Specifically, participants were asked to rate the probability that they would recommend the applicant to receive a final round interview for the job, they would personally hire the applicant for the job, and the applicant would be hired for the job. Ratings were provided on a 1 (strongly disagree) to 7 (strongly agree) scale (α = .91).

Results

Manipulation Check and Preliminary Analyses

To determine whether participants were attuned to the sexual orientation of the job applicant, we reviewed their answers to the manipulation check item. Participants who viewed a gay or lesbian applicant’s resume were removed if they claimed the job applicant was heterosexual. Similarly, participants who viewed a heterosexual applicant’s resume were removed if they claimed the job applicant was gay or lesbian. Overall, nine participants did not answer this item and were removed from the analysis. An additional eight participants answered incorrectly and were also removed from the analysis. After removing these participants, we were left with a final sample of ninety-three participants. Descriptive statistics of the final sample and correlations among measured variables are reported in Table 1.

Table 1 Study 1 means, standard deviations, and correlations among measured variables
Full size table

Main Analyses

We first conducted a 2 (applicant gender) × 2 (applicant sexual orientation) × 2 (participant gender) ANOVA with hirability as the dependent variable. The results showed no significant three-way interaction (p = .84) or any significant interaction effects of applicant gender (ps > .97). However, consistent with our predictions, a significant Applicant Sexual Orientation × Participant Gender interaction emerged, F(1, 85) = 5.31, p < .05, η 2 = .06. There were no other significant main effects or interactions. To decompose the significant Applicant Sexual Orientation × Participant Gender interaction, we conducted simple effects analyses for both male and female participants. The simple effects analyses revealed that male participants showed no difference between hiring a heterosexual applicant (M = 5.00, SD = .97) and a gay or lesbian applicant (M = 4.51, SD = 1.41), t(46) = 1.39, p = .17, η 2 = .04). However, female participants were significantly more likely to hire a gay or lesbian applicant (M = 5.14, SD = 1.32) over a heterosexual applicant (M = 4.42, SD = 1.00), t(43) = −2.05, p = .05, η 2 = .09) (Fig. 1).

Fig. 1
figure 1

Male and female participants’ ratings of job applicant hirability in Study 1

Full size image

Discussion

The results of Study 1 suggest that men and women evaluate gay and lesbian job applicants’ hirability differently. While men perceive gay and lesbian job applicants as less hirable than equally qualified heterosexual job applicants, this difference did not reach statistical significance. Women, however, perceive gay and lesbian job applicants as significantly more hirable than equally qualified heterosexual job applicants. The findings with respect to male participants are consistent with previous work showing that men are more likely to discriminate against gays and lesbians, although the difference in Study 1 was not significant. It is possible that with a larger sample, a significant difference would emerge consistent with our predictions. The pattern of results for female participants, however, is quite interesting. Although recent work has shown that women are adopting more positive attitudes toward gays and lesbians overall, the results of this study are the first to show that women give gay and lesbian applicants a boost in the hiring process. However, the results of this study are limited by the fact that we were unable to determine whether our participants in Study 1 had experience evaluating resumes. Therefore, we conducted another study with a larger sample of adult (non-student) participants in which we asked about participants’ experience evaluating resumes. We also included measures of warmth and competence to determine whether these perceptions influenced perceptions of hirability.

Study 2

Study 2 was designed to replicate the effects from Study 1 and to examine whether perceived warmth or competence toward gays and lesbians might explain men and women’s differential evaluations of gay and lesbian job applicants’ hirability. Additionally, Study 2 was conducted using a different sample of adult participants in order to strengthen the generalizability of our results. First, we predicted that male participants would perceive the gay and lesbian applicants as less hirable than heterosexual applicants, while female participants would perceive the gay and lesbian applicants as more hirable than the heterosexual applicants (Hypothesis 1). Next, we predicted that women would perceive gay male applicants as warmer than lesbian and heterosexual applicants and that men would perceive gay male and lesbian applicants as less warm than heterosexual applicants (Hypothesis 2). We also predicted that women would perceive gay and lesbian applicants as more competent than equally qualified heterosexual applicants, but that men would perceive lesbian applicants as more competent than gay male and heterosexual applicants (Hypothesis 3). Finally, we predicted that perceptions of job applicants’ competence and warmth would mediate any effects on hirability for male and female participants (Hypothesis 4).

Method

Participants

Two hundred seventy-five Amazon Mechanical Turk (M-Turk) users participated in the study online in exchange for $.50. Participant age ranged from 18 to 82 years (M = 35.03, SD = 12.88; 163 women, 112 men). The self-identified racial breakdown of the sample was as follows: 3 Native American, 25 Asian, 11 Black, 223 White, 8 Latino, and 4 multiracial participants. Thirty-one participants self-identified as non-heterosexual. Although M-Turk comprises participants from around the world, we limited participation in this study to individuals living only in the United States. Several recent studies have verified the advantages and appropriateness of this subject population for survey and experimental research (Buhrmester et al. 2011; Paolacci et al. 2010).

Procedure

Participants followed a procedure almost identical to Study 1. Participants were presented with one of the four possible job applicant resumes that were used in Study 1 (gay male, lesbian, straight male, straight female) and were asked to provide their perceptions of the job applicant. However, in Study 2, participants were told that the job applicant was applying for a position as a Program Manager rather than an Office Manager. Because Office Manager may be perceived as a female-typed position, we wanted to include a job title that was more ambiguous and less likely to be associated with any particular gender. In addition to the hirability measure from Study 1, participants were also asked to rate the job applicants’ competence and warmth. After providing their perceptions, participants completed demographic items and a manipulation check item. Finally, participants were shown a debriefing page that thanked them for their participation and explained the true purpose of the study. After completing the study, participants were able to provide feedback on Mechanical Turk’s website. Although we did not include explicit questions measuring participants’ suspicion throughout the study, no participant reported in their open-ended feedback that they correctly predicted our hypotheses or that they did not believe our cover story.

Measures

Manipulation Check

To assess whether participants were attuned to the sexual orientation of the job applicant, participants completed one item: “What was the sexual orientation of the job candidate whose resume you evaluated in this study?”

Experience Evaluating Resumes

Participants’ prior experience with evaluating resumes was measured using one item. Participants were asked how much they agreed with the statement, “I have a lot of experience evaluating resumes.” Responses were given on a 1 (strongly disagree) to 7 (strongly agree) scale.

Warmth

Participants’ perceptions of job applicant warmth were assessed using a four-item measure with ratings given on a 1 (strongly disagree) to 7 (strongly agree) scale (Rudman and Glick 1999). The four items used were warm, tolerant, sincere, and good-natured (α = .83).

Competence

Participants’ perceptions of job applicant competence were assessed using a five-item measure with ratings given on a 1 (strongly disagree) to 7 (strongly agree) scale (Rudman and Glick 1999). The five items used were confident, competent, intelligent, competitive, and independent (α = .86).

Hirability

Participants completed the same hirability measure as Study 1 (α = .95).

Results

Manipulation Check and Preliminary Analyses

To determine whether participants were attuned to the sexual orientation of the job applicant, we reviewed their answers to the manipulation check item. Participants who viewed a gay or lesbian applicant’s resume were removed if they claimed the job applicant was heterosexual. Similarly, participants who viewed a heterosexual applicant’s resume were removed if they claimed the job applicant was gay or lesbian. Overall, two participants did not answer this item and were removed from the analysis. An additional ten participants answered incorrectly and were also removed from the analysis. After removing these participants, we were left with a final sample of two hundred sixty-three participants. Descriptive statistics of the final sample and correlations between measured variables are reported in Table 2.

Table 2 Study 2 means, standard deviations, and correlations among measured variables
Full size table

Main Analyses

We first conducted a 2 (applicant gender) × 2 (applicant sexual orientation) × 2 (participant gender) ANOVA with hirability as the dependent variable. Similar to Study 1, the results showed no significant three-way interaction (p = .39) or any significant interaction effects of applicant gender (ps > .37). A significant main effect of participant gender emerged such that female participants tended to rate all resumes more favorably (M = 4.76, SD = 1.37) than male participants (M = 4.33, SD = 1.32), F(1, 255) = 14.64, p < .01, η 2 < .01. More importantly, however, a significant Applicant Sexual Orientation × Participant Gender interaction emerged, F(1, 255) = 18.80, p < .001, η 2 = .07. To decompose the significant Applicant Sexual Orientation × Participant Gender interaction, we conducted simple effects analyses for both male and female participants. The simple effects analyses revealed that male participants were significantly more likely to hire a heterosexual applicant (M = 4.66, SD = 1.16) over a gay or lesbian applicant (M = 3.93, SD = 1.39), t(107) = 2.97, p = .004, η 2 = .08). However, female participants were significantly more likely to hire a gay or lesbian applicant (M = 5.11, SD = 1.25) over a heterosexual applicant (M = 4.43, SD = 1.40), t(152) = −3.18, p = .002, η 2 = .06) (Fig. 2).

Fig. 2
figure 2

Male and female participants’ ratings of job applicant hirability in Study 2

Full size image

Next, we conducted a 2 (applicant gender) × 2 (applicant sexual orientation) × 2 (participant gender) ANOVA with warmth as the dependent variable. The results showed no significant three-way interaction (p = .66) or any significant interaction effects of applicant gender (ps > .22). A significant main effect of participant gender emerged such that female participants tended to rate all resumes more favorably (M = 4.69, SD = .79) than male participants (M = 4.50, SD = .73), F(1, 255) = 3.89, p < .01, η 2 = .02. More importantly, however, a marginally significant Applicant Sexual Orientation × Participant Gender interaction emerged, F(1, 255) = 3.16, p = .08, η 2 = .01. To decompose the Applicant Sexual Orientation × Participant Gender interaction, we conducted simple effects analyses for both male and female participants. The simple effects analyses revealed that male participants showed no difference in warmth between the heterosexual applicants (M = 4.50, SD = .56) and the gay and lesbian applicants (M = 4.51, SD = .89), t(107) = -.03, p = .98, η 2 < .01). However, female participants perceived the gay and lesbian applicants as significantly warmer (M = 4.86, SD = .78) than the heterosexual applicants (M = 4.52, SD = .77), t(152) = −2.68, p = .008, η 2 = .05).

To test our specific hypotheses regarding warmth, we then conducted a one-way ANOVA for both male and female participants with experimental condition as the independent variable and warmth as the dependent variable. The analysis for female participants revealed a significant effect between our four experimental conditions F(3, 150) = 3.11, p = .03. A planned-comparisons test revealed that women perceived the gay male applicant as significantly warmer (M = 4.99, SD = .80) than the lesbian applicant (M = 4.73, SD = .76), heterosexual male applicant (M = 4.49, SD = .85), and heterosexual female applicant (M = 4.55, SD = .70), t(3, 150) = 2.71, p = .008. The analysis for male participants revealed no significant effect between our four experimental conditions F(3, 105) = .18, p = .91 (Fig. 3).

Fig. 3
figure 3

Male and female participants’ ratings of job applicant warmth in Study 2

Full size image

Next, we conducted a 2 (applicant gender) × 2 (applicant sexual orientation) × 2 (participant gender) ANOVA with competence as the dependent variable. The results showed no significant three-way interaction (p = .77) or any significant interaction effects of applicant gender (ps > .46). A significant main effect of participant gender emerged such that female participants tended to rate all resumes more favorably (M = 5.00, SD = .95) than male participants (M = 4.76, SD = .88), F(1, 255) = 4.54, p < .05, η 2 = .02. More importantly, however, a significant Applicant Sexual Orientation × Participant Gender interaction emerged, F(1, 255) = 10.70, p = .001, η 2 = .04. To decompose the Applicant Sexual Orientation × Participant Gender interaction, we conducted simple effects analyses for both male and female participants. The simple effects analyses revealed that male participants perceived the heterosexual applicants as significantly more competent (M = 4.93, SD = .82) than the gay and lesbian applicants (M = 4.60, SD = .93), t(107) = 1.97, p = .05, η 2 < .04). However, female participants perceived the gay and lesbian applicants as significantly more competent (M = 5.21, SD = .78) than the heterosexual applicants (M = 4.80, SD = 1.05), t(152) = −2.75, p = .007, η 2 = .05).

To further analyze the competence variable, we then conducted a one-way ANOVA for both male and female participants with experimental condition as the independent variable and competence as the dependent variable. The analysis for female participants revealed a significant effect between our four experimental conditions F(3, 150) = 2.71, p = .05. A planned-comparisons test revealed that women perceived the gay male applicant (M = 5.26, SD = .88) and lesbian applicant (M = 5.17, SD = .66) as significantly more competent than the heterosexual male applicant (M = 4.72, SD = 1.21) and heterosexual female applicant (M = 4.87, SD = .89), t(3, 150) = 2.77, p = .006. The analysis for male participants revealed no significant effect between our four experimental conditions F(3, 105) = 1.30, p = .27 (Fig. 4).

Fig. 4
figure 4

Male and female participants’ ratings of job applicant competence in Study 2

Full size image

Finally, we analyzed the variable measuring participants’ experience evaluating resumes to determine whether participants with this experience responded differently than participants without this experience. Overall, the average participant did not report having significant experience evaluating resumes (M = 3.71, SD = 1.62). However, there was a wide variety of responses to this item. For example, 131 participants (49.8 % of the sample) responded below the midpoint of the scale, meaning that they did not have significant experience evaluating resumes. On the other hand, 104 participants (39.5 % of the sample) responded above the midpoint of the scale, indicating that they did have significant experience evaluating resumes.

To determine whether these participants evaluated job applicants differently, we conducted a regression analysis in which hirability was regressed on participant gender, applicant gender, applicant sexual orientation, experience evaluating resumes, and the various interactions between these variables. The results revealed no significant interactions involving experience evaluating resumes (ps > .23).

To more closely examine this variable, we created a dichotomous variable that captured participants’ experience evaluating resumes. Participants who answered below the midpoint of the scale and lacked experienced were coded as 0, while participants who answered above the scale and had previous experience were coded as 1. We then split the analysis along this variable and conducted a 2 (applicant sexual orientation) × 2 (participant gender) ANOVA. The results of this analysis showed a marginally significant Applicant Sexual Orientation × Participant Gender interaction consistent with the general pattern of results for participants without significant experience evaluating resumes, F(1, 127) = 3.65, p = .06, η 2 = .03. The results also showed a highly significant Applicant Sexual Orientation × Participant Gender interaction consistent with the general pattern of results for participants with significant experience evaluating resumes, F(1, 100) = 12.41, p = .001, η 2 = .11. Overall, this analysis suggests that while both groups of participants evaluated resumes similarly and consistent with our predictions, participants with experience evaluating resumes were even more likely to demonstrate differential perceptions of gay and lesbian applicants.

Mediation Analysis

Given that women perceive gay and lesbian applicants to be more hirable than equally qualified heterosexual job applicants and men perceive gay and lesbian applicants to be less hirable than equally qualified heterosexual applicants, we next examined whether men and women’s perceptions of the job applicants’ competence and warmth mediated the relationship between applicant sexual orientation and hirability. To this end, we conducted moderated mediation analysis with bootstrapping in which we estimated conditional indirect effects using multiple mediators (Preacher and Hayes 2008). For the moderated mediation analysis, hirability was entered as the dependent variable. Applicant sexual orientation was entered as the predictor variable. Competence and warmth were entered as proposed mediators, while participant gender was entered as a moderating variable in the SPSS macro (PROCESS) created by Preacher and Hayes for moderated mediation analysis.

For female participants, the 95 % bias-corrected confidence interval (CI) for the conditional indirect effect (derived from 5000 bootstrap samples) through warmth as a mediator did not contain zero and was therefore significant, 95 % CI = [−.19, −.01]. Similarly, the conditional indirect effect through competence as a mediator did not contain zero and was therefore significant, 95 % CI = [.13, .74]. These results suggest that for female participants, both warmth and competence mediated the relationship between job applicant sexual orientation and hirability.

For male participants, the conditional indirect effect through warmth as a mediator was not significant, 95 % CI = [−.07, .08]. However, the 95 % bias-corrected confidence interval for the conditional indirect effect through competence did not contain zero and was therefore significant, 95 % CI = [−.71, −.01]. These results suggest that for male participants, only competence mediated the relationship between job applicant sexual orientation and hirability.

Discussion

Overall, the results of Study 2 replicate the effects from Study 1, but some of our hypotheses were only partially supported. With respect to hirability, female participants perceived the gay and lesbian applicants as significantly more hirable than equally qualified heterosexual applicants, while male participants perceived the gay and lesbian applicants as significantly less hirable than equally qualified heterosexual applicants, confirming Hypothesis 1. With respect to perceived warmth, women did perceive the gay male applicant as significantly warmer than the other three candidates, partially confirming Hypothesis 2. Women also perceived the gay and lesbian applicants together as more warm than the heterosexual applicants. However, male participants showed no difference in perceived warmth between the four applicants, partially rejecting Hypothesis 2. With respect to perceived competence, women did perceive the gay and lesbian applicants as more competent than heterosexual applicants, partially confirming Hypothesis 3. However, rather than perceiving the lesbian applicant as the most competent, men perceived the lesbian applicant as the least competent of all, partially rejecting Hypothesis 3. Additionally, men perceived the gay and lesbian applicants together as less competent than equally qualified heterosexual applicants. The mediation analyses revealed that perceptions of competence mediated the effect on hiring for both male and female participants, but warmth also mediated the effect on hiring, to a lesser degree, only for female participants. These results partially confirm Hypothesis 4. Interestingly, the results of Study 2 also indicate that participants with experience evaluating resumes show the same pattern of results as participants without experience evaluating resumes. Moreover, participants with significant experience evaluating resumes were more likely to demonstrate bias against gay and lesbian applicants, although this difference was not significant.

Overall, it is unclear why male participants rated all four applicants as equally warm. One explanation is that men found it difficult to assess warmth simply using a brief resume. However, it is also possible that men did not think warmth was a relevant trait for the Program Manager position and therefore did not consider warmth as strongly as they considered competence and hirability. It is also interesting that male participants failed to perceive the lesbian applicant as particularly competent given the positive stereotypes regarding lesbians in the workplace. Because men perceived lesbians and gay men as the least competent, it is possible men’s negative attitudes toward gays and lesbians in general overrode any positive stereotypes related to lesbians in the workplace domain. Future research could examine whether the job attributes related to a position might enhance or suppress stereotypes.

General Discussion

In addition to showing that men perceive gays and lesbians as less hirable than heterosexual job applicants, the results of these studies also show that women find gays and lesbians to be more hirable than heterosexual job applicants. We also find that men and women’s perceptions of gay and lesbian applicants’ hirability are mediated by perceived competence of gays and lesbians. For female participants, perceived warmth also mediated perceptions of gay and lesbian applicants’ hirability, but to a lesser extent. Our results add to the existing literature on attitudes toward gays and lesbians by documenting the positive bias that women seem to have for gays and lesbians. Overall, our results provide evidence that women perceive gays and lesbians to be more competent and warm, and that women’s perceptions of gays and lesbians’ competence actually lead women to find gays and lesbians more hirable.

Theoretically, our results contribute to the body of work examining perceptions of gay men and lesbians and emphasize the important roles of target gender and perceiver gender in these perceptions. For example, the men in our sample perceived gay men and lesbians as less competent than equally qualified heterosexual applicants, but showed no differences in perceived warmth. The women in our sample on the other hand perceived gay men and lesbians as significantly more competent and warm than equally qualified heterosexual applicants. Furthermore, both perceptions of competence and warmth helped explain why women viewed gay male and lesbian applicants as more hirable.

Future research in this area could examine why women perceive gay and lesbian job applicants as more warm and competent. One explanation is that because women experience discrimination and the glass ceiling in organizations, women may believe that gay and lesbian applicants must be more competent than equally qualified heterosexual applicants in order to advance in their careers and overcome interpersonal and institutional discrimination. It is also possible that women view gays and lesbians as more emotionally intelligent due to coping experiences with discrimination and navigating the coming out process. Findings from the stigma compensation literature support this argument. Specifically, stigmatized individuals are better able to decipher subtle cues in interpersonal interactions (Hall 1978; Miller and Myers 1998) and are more attentive to situational cues (Frable et al. 1990). For men, our findings suggest that men do not see gay or lesbian applicants as more competent, despite previous work documenting positive stereotypes associated with lesbian’s competence. Future research could further explore these stereotypes to determine whether they still exist and in what contexts they are likely to be endorsed.

Practically, our results suggest sexual orientation biases are more nuanced than what previous literature has found. Having a better understanding of these biases should help organizations when making important personnel decisions. It is important to note that based on our results, both male and female participants have sexual orientation biases that influence their perceptions of gay and lesbian applicants. However, because women only comprise approximately 40 % of management positions in North American organizations (Davidson and Burke 2011), it is possible that men’s negative bias is actually more common and contributing to widespread discrimination against gay and lesbian applicants. Therefore, it could very well be that placing more women in selection roles could either act as a catalyst for the inclusion of gay and lesbian employees in organizations or at least temper the negative bias that men show against gays and lesbians. It is also possible that hiring decisions made by teams of both men and women could lead to less biased decisions. Future research could further explore this possibility by having male–male, male–female, and female–female dyads follow a similar procedure to these studies to see whether mixed-gender dyads show less bias when evaluating gay and lesbian job applicants. Overall, the results of our study highlight the importance of considering the demographic characteristics of individuals making selection decisions. If these demographic characteristics and resulting biases are not considered, organizations may be unintentionally perpetuating disadvantage against underrepresented groups.

Because previous work has found men and women often have different attitudes toward gay men than to lesbians, it is interesting that in both of our studies, applicant gender did not play a significant role in either male or female participants’ evaluations of the job applicants’ hirability. For male participants, it is possible that men’s more negative views toward homosexuality in general influenced their judgments of both gay men and lesbians in a similar way. For female participants, it is possible that although specific attitudes and stereotypes toward gay men and lesbians differ, both gay men and lesbians have positive stereotypes associated with their competence that lead women to find them more hirable.

It is also interesting that men and women evaluated heterosexual job applicants differently in Study 1, but not Study 2. In Study 1, men evaluated the heterosexual job applicants as more hirable (M = 5.00, SD = .97) than women did (M = 4.42, SD = 1.00), F(1, 44) = 3.92, p = .05, η 2 = .08). However, in Study 2, there were no significant differences between men’s ratings of heterosexual applicants’ hirability (M = 4.66, SD = 1.16) and women’s ratings of heterosexual applicants’ hirability (M = 4.43, SD = 1.40), F(1, 137) = 1.07, p = .30, η 2 = .008). Because our sample in Study 2 was substantially larger and did not replicate the effect, it is possible that this result from Study 1 is not a robust effect.

One limitation of these studies is that although our samples comprise older adults, some of whom have experience making hiring decisions in organizations, we do not directly assess the hiring decisions of managers in real organizations. Conducting a similar study with real organizational decision makers may provide additional evidence that women prefer gay and lesbian job applicants relative to heterosexual job applicants under more realistic conditions. On one hand, we may predict that managers in real organizations would be trained not to make selection decisions based on criteria such as gender and sexual orientation. However, it is also possible that managers who evaluate many resumes very quickly would rely on the use of stereotypes and heuristics even more than the participants in our studies (see Bertrand and Mullainathan 2004). Future work could examine this possibility with a sample of HR managers.

In summary, these results provide evidence that women find gay and lesbian job applicants more hirable than their equally qualified heterosexual counterparts. Although gays and lesbians must regularly contend with discrimination, our results suggest being gay may not always be an obstacle when applying for a job if women are doing the selection.