An empirical examination of factors affecting the incomes of gay men.INTRODUCTION
Numerous studies have demonstrated that gay men earn less than their heterosexual counterparts (e.g. Allegretto & Arthur, 2001; Badgett 1995; Berg & Lien, 2002; Carpenter, 2004, 2005, 2008a, 2008b; Heller Clain & Leppel, 2001). These studies report gay men earn 2.4 to 22 percent less than similarly qualified heterosexual men and lesbians earn 3 to 30 percent more than similarly qualified heterosexual women. The mixed findings, in particular the results for lesbians, would suggest that there are a number of factors affecting income in addition to, or in conjunction with, discrimination on the basis of sexual orientation. These studies typically control for age, ethnicity, educational attainment, experience, organizational size, and job and organizational tenure. Researchers have offered a number of possible contributing factors to these income disparity findings. These factors include sexual orientation discrimination (Badgett, 1995), the greater likelihood that gay men work in traditionally female occupations (Blandford, 2003), or other as-yet unobserved labor-market traits (Zavodny, 2008). These studies compare gay men to their heterosexual counterparts rather than exploring factors affecting the range of incomes of gay men. Although determining the factors that affect lesbian's incomes is important, this study focuses on the factors that affect gay men's incomes.
Blandford (2003) suggests that some gay men may choose to work in traditionally female dominated occupations, or in organizations with a greater proportion of women. The incomes of both men and women has been shown to be inversely related to the proportion of women in that job (Elvira & Graham, 2002; England, Farkas, Kilbourne & Dou, 1988; Sorensen, 1990). Gay men may choose these occupations or organizations because women, compared to men, tend to hold more positive attitudes toward gay men (Herek, 1988; Herek & Capitano, 1996). Gay men may also choose to work in occupations or organizations that employ greater numbers of other gay men and lesbians. Organizations with larger numbers of women and gay men and lesbians have been described as "safe havens" for gay men (Ragins, 2004). Within these "safe havens", gay men report do less discrimination. There is some evidence that gay men constrain their career choices in an attempt to avoid discrimination and other manifestations of heterosexism (Chung, 2001; Morrow, Gore & Campbell, 1996). This study assesses whether working in organizations with larger proportions of women and other gay men and lesbians is associated with lower incomes for gay men. This study will also assess whether gay men's incomes are associated with other manifestations of heterosexism.
Heterosexism has been defined as valuing heterosexuality as superior to and/or more natural or normal than gay and lesbian sexual orientations (Morin, 1977). Heterosexism includes a much broader range of discrimination than homophobia The distinction is critical because it focuses on heterosexual privilege and draws attention to the constancy of the experience and not just episodic harassment or violence (Herek & Berrill, 1992). An analogy exists with racist behaviors which may be a result of fear or aversion or may be based on self-interests, beliefs, group norms or social institutions (Allport, 1954). As with racism or sexism, heterosexism may refer to the behaviors of individual or to institutions. Examples of institutionalized organizational heterosexism include the lack of policies that prohibit discrimination on the basis of sexual orientation and failure to provide gay and lesbian employees benefits equal to those provided to heterosexual employees. Examples of institutionalized societal heterosexism include lack of legislative protection against discrimination in the workplace. Organizations are embedded within society and individuals' heterosexist behaviors occur within and are affected by both the organization and the society. For the purposes of this research, institutional heterosexism will be referred to as heterosexism and individual heterosexist behaviors in the workplace will be referred to as perceived workplace discrimination. Perceived discrimination on the basis of sexual orientation within the workplace has been shown to be positively associated with the level of organizational heterosexism which, in turn, is negatively related to protective legislation (Ragins & Cornwell, 2001). Causal directions are ambiguous, and possibly reciprocal, but one can conclude that workplace discrimination is less likely to be perceived in less heterosexist organizations in less heterosexist societies.
Heterosexist organizations are characterized by an absence of protective legislation, absence of supportive organizational policies and practices, and workgroup composition of majority heterosexuals, and each of these organizational heterosexism variables was significantly related to perceived workplace discrimination (Ragins & Cornwell, 2001). Organizational heterosexism has been shown to be significantly related to lower disclosure of sexual orientation in the workplace, lower job satisfaction, lower organizational commitment, and higher turnover intentions (Button, 2001; Ragins & Cornwell, 2001, Waldo, 1999) for gay employees.
The presence of legislative protection against discrimination on the basis of sexual orientation has been shown to associated with lower levels of observed discrimination on the basis of sexual orientation in workplaces (Ragins & Cornwell, 2001) and with greater likelihood of the adoption of same-sex partner health benefits (Chuang, Church & Ophir, 2011). The presence of such a state law may indicate more positive attitudes towards gay men (Wald, Button & Rienzo, 1996), a greater degree of acceptance of equal treatment (Button, Rienzo & Wald, 2000), and a stronger commitment to equitable compensation (Chuang et al., 2011). The incomes of gay men should be less negatively affected in environments with less observed discrimination in the workplace, greater general positive attitudes towards gay men, and a greater degree of acceptance of equal treatment. Based on this research, I hypothesize:
[H.sub.1] Income reported by gay men will be inversely related to the proportion of women in their organizations.
[H.sub.2] Income reported by gay men will be inversely related to the proportion of gay men and lesbians in their organizations.
[H.sub.3] Income reported by gay men will be positively related to the presence of state laws forbidding discrimination on the basis of sexual orientation.
[H.sub.4] Income reported by gay men will be positively related to the number of supportive organizational policies and practices.
[H.sub.5] Income reported by gay men will be will be inversely related to reported discrimination based on sexual orientation.
Studying gay men and lesbians, as a population, presents numerous challenges. Other demographic information is widely shared and tracked by organizations. Many gay men and lesbians have considerable interest in not being found or at the least, not having their sexual orientations known by their organizations. Probability sampling requires that all cases in the population are randomly selected and have a known probability of being included in the sample (Singleton & Straits, 1999). This is not possible with this population. Nonprobability sampling was therefore used. Nonprobability sampling introduces several problems: it does not control for investigator bias in the selection of units and the pattern of variability cannot be predicted from probability sampling theory which makes calculation of sampling error or estimation of sample precision impossible.
A survey conducted on-line using a form of convenience sampling to develop the sampling frame. Consistent with Ragins & Cornwell (2001) and Day & Schoenrade (1997), the sampling procedure involved the members of gay and lesbian rights organizations and on-line communities. Using only the members of gay rights organizations introduces several important limitations and may limit the generalizability of the results. Members of "mainstream" gay rights organizations are disproportionately white and well educated (Croteau, 1996), may be more likely than nonmembers to be out at work, may be more likely to be sensitive to gay and lesbian issues in the workplace and may be more likely to seek employment in organizations that are supportive of gay and lesbian employees (Day & Schoenrade, 1997).
The sampling frame was also increased by snowball sampling. Snowball sampling involves identifying members of the target population and requesting that they assist in the identification of other members of the target population. Survey respondents were encouraged to send the survey's link to other gay men and lesbians. Snowball sampling rests on the assumption that members of the target population know each other. The sample included all those who self-identified as gay or lesbian regardless of the degree of disclosure of their sexual orientations in the workplace or within their personal lives, actual sexual behaviors, or current relationship status.
The survey instrument developed was pre-tested on a group of 15 persons including gay men, a number of persons for whom English is not a first language, and persons whose computer skills could be described as limited. These pre-tests helped ensure clarity of the instrument and the instructions. The first web page explained the purpose of the research, sought informed consent, provided instructions for completion, and provided assurances of anonymity. All analysis was conducted using SPSS 18.0 statistical package.
Demographic/Individual Characteristics were measured using single item measures. The characteristics included sex, age, ethnicity, sexual orientation, educational levels, employment status, income, state in which they worked, organization size, organization tenure, and job tenure. Income categories provided were as follows: 1= under $10,000, 2=$10,000-$19,999, 3=$20,000-$29,999, 4=$30,000-$39,999, 5=$40,000-$49,999, 6=$50,000-$59,999, 7=$60,000-$69,999, 8=$70,000-$79,999, 9=$80,000-$89,999, 10=$90,000-$99,999, and 11= Over $100,000. Organization size was measured using a 5-point scale with 1= "less than 10" and 5="over 10,000". Respondents were asked to indicate their organizational and job tenure for their current job in years and months. These were converted to months for analysis. Respondents were asked the degree of disclosure of their sexual orientation in the workplace by asking "at work, have you disclosed your sexual orientation to: 1) no one, 2) some people, 3) most people and 4) Everyone.
To measure the sexual orientation composition of the work group, respondents were asked about the sexual orientations of their co-workers. They were given a 4-point scale with 025% gay or lesbian and 75 - 100% gay and lesbian as anchors and a Don't Know option which was coded as missing data. To measure the gender composition of the work group respondents were given a 4-point scale with 0-25% women and 75 - 100% women as anchors and a Don't Know option which was coded as missing data.
The presence of protective legislation was determined by the state of residence reported by the respondents and the statewide anti-discrimination laws and policies reported by the civil rights organization, Human Rights Campaign. At the time of the data collection, 14 states provided protective legislations against discrimination on the basis of sexual orientation.
A 6-item scale developed by Ragins & Cornwell (2001) was used to assess the supportiveness of organizational policies and practices for gay and lesbian employees. The scale includes the following items: Does your organization: 1) have a non-discrimination policy that includes sexual orientation? 2) include sexual orientation in the definition of diversity? 3) include awareness of gay/lesbian/bisexual issues in diversity training? 4) offer same-sex domestic partners benefits? 5) offer gay/lesbian/bisexual resource/support groups? and 6) welcome same-sex partners at company social events? Responses indicating the presence of the supportive policy were coded as 1; the absence of the policy as 0, and Don't Know was coded as missing. The items were then summed to create an overall scale of organizational policies and practices with values ranging from 0 to 6 with 0 representing the absence of all listed policies and practices and 6 the presence of all. For this study, principal component analysis yielded a single factor for organizational policies and practices with an eigenvalue of 2.90 accounting for 48.3% of the variance (coefficient alpha of .78).
Perceived workplace discrimination was measured using another scale developed and tested by Ragins & Cornwell (2001) which was based upon the 15-item Workplace Prejudice/Discrimination Inventory (James, Lovato, & Cropanzano, 1994). The original inventory measured racial discrimination in the workplace and was found to be single factor, reliable and valid (James et al., 1994). Ragins & Cornwell (2001) substituted references to race with references to sexual orientation. The Workplace Prejudice/Discrimination Inventory modified for sexual orientation includes items such as: "Prejudice against gays and lesbians exists where I work" and "At work I am treated poorly because of my sexual orientation". A 5-point Likert scale with Completely Disagree and Completely Agree as anchors was used. For this study, principal component analysis yielded a single factor with an eigenvalue of 6.88 accounting for 57.3% of the variance (coefficient alpha of .93).
The sample consisted of 264 individuals who self-identified as gay men, who live in the United States, and who reported being employed fulltime. Respondents reporting that they were unemployed, retired, or self-employed were eliminated from this analysis. The ethnicity of the respondents was 81.8% Caucasian (n=216), 6.1% African-American (n=16), 3.8% Asian (n=10), 3.0% Aboriginal (n=8), 8.0% Latino or Hispanic (n=21), and 1.1% Middle Eastern or North African (n=3) (percentages exceed 100 as some respondents reported multiple ethnicities). Because the proportions of most ethnicities were too low for analysis, the categories were combined creating the variable called Majority. Respondents indicating their ethnicity to be Caucasian only were coded 1 (n=209), all others (n=55) were coded 0.
When asked about the percent of gay or lesbians coworkers, 71.2% indicated 0-25%, 4.2% indicated 26-50%, 1.5% indicated 51-75%, 2.3% indicated 76-100% and 20.8% indicated they didn't know or didn't respond to that question. When asked about the percent of women coworkers, 14.7% indicated 0-25%, 13.6% indicated 25-50%, 30.3% indicated 50-75% and 15.5% indicated 75-100%.
Means, standard deviations, correlation coefficients, and reliability coefficients are displayed in Table 1.
The majority of respondents had completed some form of post secondary education. 0.8% of respondents had completed some high school, 11.7% had completed high school, 6.8% had completed technical or vocational training, 10.6% an Associate's degree, 41.3% a Bachelor's degree, 4.5% a professional degree, 19.3% a Master's degree and 4.9% a doctoral degree. Regarding disclosure of sexual orientation, 18.3% reported having disclosed their sexual orientation to no one at work, 28.5% to some people, 22.4% to most people and 30.8% to everyone.
Quick inspection of the correlation tables shows that incomes is significantly correlated with age, education, organizational size, organizational tenure, job tenure, protective legislation, the proportion of gay and lesbian coworkers, and perceived discrimination. Ragins and Cornwell (2001) found that perceived discrimination was negatively correlated with protective legislation and supportive organizational policies and practices. This finding was confirmed with these data.
Multiple regression was used to test the hypotheses. The results are presented in Table 2.
On the first step, the demographic characteristics (age, ethnicity, and education,) were entered to isolate the effects that they may have on income. On the second step, the general employment-related variables (organizational size, organizational tenure, and job tenure) were entered to isolate their effects on income. On the third step, the heterosexism variables (the presence of legislative protection against discrimination on the basis of sexual orientation at state-level, disclosure of sexual orientation at work, percentage of women co-workers, percent gay or lesbian co-workers, supportive policies & procedures, and perceived workplace discrimination) to determine their effects on income. Multicollinearity is not a problem as the highest correlation was .53 and the largest variance inflation factor was only 1.91.
Two of the control variables, age and education, were strong predictors of income for both gay men. For gay men, income was significantly related to age ([beta]= .19, p<.05) and education ([beta]= .26, p<.001). None of the other demographic or employment variables were significantly related to income. Of the heterosexism variables only the proportion of gay and lesbian coworkers was significant, therefore only hypothesis 2 was supported. There was a negative, statistically significant relationship between gay men's income and the proportion of gay and lesbian coworkers ([beta]= -.18, p<.01).
Past studies have clearly shown that gay men's incomes are lower than their heterosexual counterparts. This study tested the effects of several heterosexism-related variables on the incomes of gay men. The effects of all but one of these heterosexism-related variables on income disappeared when controlling for the demographic and employment variables. This study does show that gay men's incomes are negatively related to the proportion of gay men and lesbian coworkers.
Past studies have suggested that gay men may choose to work in "safe haven" organizations with a high proportion of women or gay men and lesbians and that leads to lower incomes. This study did not confirm that gay men's incomes are related to the proportion of women coworkers but were related to the proportion of gay and lesbian coworkers providing some support to the "safe haven' hypotheses suggested by previous research. Whether or not gay men consciously choose organizations that are "safe havens" and accept lower incomes as part of the tradeoff or whether there are barriers to the hiring of gay men in organizations with lower proportions of other gay men and lesbians is beyond the scope of this study. Further research is required to determine the degree to which this is a conscious choice and the degree to which this choice is driven by the desire to avoid heterosexism.
Allegretto, S.A. & Arthur, M.M. (2001). An empirical analysis of homosexual/heterosexual male earnings differentials: Unmarried and unequal", Industrial and Labor Relations Review, 54, 631-646.
Allport, G.W. (1954). The nature of prejudice. Cambridge: Addison-Wesley.
Badgett, M.V.L. (1995). The wage effects of sexual orientation discrimination. Industrial and Labor Relations Review, 48, 726-739.
Berg, N. & Lien, D. (2002). Measuring the effect of sexual orientation on income: Evidence of discrimination?, Contemporary Economic Policy, 20, 394-414.
Blandford, J. M. (2003). The nexus of sexual orientation and gender in the determination of earnings. Industrial and Labor Relations Review, 56, 622-642.
Button, S. B. (2001). Organizational efforts to affirm sexual diversity: A cross-level examination. Journal of Applied Psychology, 86, 17-28.
Button, J.M., Rienzo, B.A., & Wald K.D. (2000). The Politics of Gay Rights at the Local and State Level. In C.A. Rimmerman, K.D. Wald, & C. Wilcox (Eds.), Politics of Gay Rights (pp. 269-289). Chicago: University of Chicago Press.
Carpenter, C. (2004). New evidence on gay and lesbian household incomes. Contemporary Economic Policy, 22, 78-94.
Carpenter, C. (2005). Revisiting the income penalty for behaviorally gay men: Evidence from NHANES III. Labour Economics, 14, 25-34.
Carpenter, C. (2008a) Sexual orientation, income, and non-pecuniary economic outcomes: New evidence from young lesbians in Australia. Review of Economics of the Household, 6, 391-408.
Carpenter, C. (2008b) Sexual orientation, work and income in Canada. The Canadian Journal of Economics, 41, 1239-1261.
Chuang, Y.T., Church, R., & Ophir, R. (2011). Taking sides: The Interactive influences of institutional mechanisms on the adoption of same-sex partner health benefits by Fortune 500 corporations, 1990-2003. Organization Science, 22, 190-209.
Chung, Y. B. (2001). Work discrimination and coping strategies: Conceptual frameworks for counseling lesbian, gay, and bisexual clients. The Career Development Quarterly, 50, 33-44.
Croteau, J. M. (1996). Research on the work experiences of lesbian, gay, and bisexual people: An integrative review of methodology and findings. Journal of Vocational Behavior, 48, 195-209.
Day, N.E. & Schoenrade, P. (1997). Staying in the closet versus coming out: Relationships between communication about sexual orientation and work attitudes. Personnel Psychology, 50, 147-163.
Elvira, M.M. & Graham, M.E. (2002). Not just a formality: Pay system formalization and sex-related earnings effects. Organization Science, 13, 601-617.
England, P., Farkas, G., Kilbourne, B.S. & Dou, T. (1988). Explaining occupational sex segregation and wages: Findings from a model with fixed effects. American Sociological Review, 53, 544-558.
Herek, G.M. (1988). Heterosexuals' attitudes toward lesbians and gay men: Correlates and gender differences. The Journal of Sex Research, 25, 451-477.
Herek, G.M. & Capitano, J.P. (1996). "Some of my best friends": Intergroup contact, concealable stigma, and heterosexuals' attitudes toward gay men and lesbians. Personality and Social Psychology Bulletin, 22, 412424.
Huffman, M.L. & Velasco, S.C. (1997). When more is less. Sex composition, organizations, and earnings in U.S. firms. Work and Occupations, 24, 214-244.
Heller Clain, S. & Leppel, K. (2001). An investigation into sexual orientation discrimination as an explanation for wage differences. Applied Economics, 33, 37-47.
Herek, G. M. & Berrill, K. (1992). Hate crimes: Confronting violence against lesbians and gay men. Thousand Oaks, CA: Sage.
James, K., Lovato, C. & Cropanzano, R. (1994). Correlational and known group comparison validation of a workplace prejudice/discrimination inventory. Journal of Applied Social Psychology, 24, 1573-1592.
Morin, S. (1977). Heterosexual bias in psychological research on lesbianism and male homosexuality. American Psychologist, 32, 629-637
Morrow, S. L., Gore, P. A. Jr., & Campbell, B. W. (1996). The application of a sociocognitive framework to the career development of lesbian women and gay men. Journal of Vocational Behavior, 48, 136-148.
Plug, E. & Berkhout, P. (2004). Effects of sexual preferences on earnings in the Netherlands. Journal of Popular Economics, 17, 117-131.
Podskoff, P. M., MacKenzie, S. B., Lee, J-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903.
Ragins, B. R. (2004) Sexual orientation in the workplace: The unique work and career experiences of gay, lesbian and bisexual workers. Research in Personnel and Human Resource Management, 23, 37-122.
Ragins, B.R. & Cornwell, J.M. (2001). Pink triangles: Antecedents and consequences of perceived workplace discrimination against gay and lesbian employees. Journal of Applied Psychology, 86, 1244-1261.
Singleton, R. A. & Straits, B. C. (1999). Approaches to social research. NY: Oxford University Press.
Sorensen, E. (1990). The crowding hypothesis and comparable worth. Journal of Human Resources, 25, 55-89.
Wald, K.D., Button, J.W., & Rienzo, B.A. (1997). All politics is local: Analyzing local gay rights legislation. Washington, DC: National Gay and Lesbian Task Force Policy Institute.
Waldo, C.R. (1999). Working in a majority context: A structural model of heterosexism as minority stress in the workplace. Journal of Counseling Psychology, 46, 218-232.
Zavodny, M. (2008). Is there a 'marriage premium' for gay men. Review of Economics of the Household, 6, 369-389.
Robin Church, Ryerson University
Table 1 Means, Standard Deviations, Correlation Coefficients, and Reliability Coefficients M SD 1 2 3 1 Age 38.7 10.23 -- 2 Majority .79 .41 .11 -- 3 Education 5.95 1.68 19 ** .00 -- 4 Org Size 3.34 1.19 .02 -.07 .02 5 O Tenure 79.6 87.8 .53 ** .04 .08 6 J Tenure 60.7 69.7 .50 ** .07 .00 7 Protect .44 .50 .13 * -.09 .05 8 Out 2.66 1.10 -.03 .11 -.08 9 Women 3.55 .95 -.03 .00 .01 10 LGB 2.21 .64 -.10 -.12 -.10 11 Policies 2.99 1.94 -.03 -.07 17 ** 12 Discrim 2.45 .83 .10 -.01 .08 13 Income 6.03 2.57 32 ** .01 39 ** 4 5 6 7 8 1 Age 2 Majority 3 Education 4 Org Size -- 5 O Tenure 15 ** -- 6 J Tenure .00 .61 ** -- 7 Protect .05 .04 .02 -- 8 Out -.08 -.02 .05 .03 -- 9 Women .04 -.07 -.03 .03 .13 * 10 LGB -.22 ** -.16 * -.10 .01 .24 ** 11 Policies 19 ** -.06 -.07 .21 ** .30 ** 12 Discrim .26 ** .18 * .13 * -14 ** -.35 ** 13 Income .13 * .26 ** .21 ** 14 ** .05 9 10 11 12 1 Age 2 Majority 3 Education 4 Org Size 5 O Tenure 6 J Tenure 7 Protect 8 Out 9 Women -- 10 LGB -.05 -- 11 Policies .12 * .21 ** (.78) 12 Discrim -.06 -.24 ** - 32 ** (.93) 13 Income -.06 -.22 ** .20 ** -.08 ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). Table 2: Regression Predicting Income [beta] Step 1 Step 2 Step 3 Step 1: Demographic Variables Age .31 *** .20 * .19 * Majority -.04 -.05 -.06 Education .30 *** .29 *** .26 *** Step 2: Employment Variables Organization Size .13 .08 Organization Tenure .15 .17 Job Tenure .01 -.08 Step 3: Heterosexism Variables Legal Protection .03 Disclosure of Sexual Orientation at Work -.02 Women Co-workers -.09 Gay & Lesbian Co-workers -.18 ** Supportive Policies and Practices .15 Perceived Discrimination -.11 [R.sup.2] .22 *** .26 * .32 * [DELTA][R.sup.2] .22 *** .04 * .06 * Adjusted [R.sup.2] .21 *** .23 * .26 * F 15.982 9.575 6.150 * p < .05, ** p < .01, *** p < .001