This paper explains attitudes towards gays and lesbians, and explores the complex relationship of religiosity, youth, masculinity and support for gay rights. Based on a large, reliable and nationally representative study (n=1405) from PEW Center carried out in 2006, we estimate three logistic regression models predicting approval for gay marriage, gay adoption and gays in the military, which helps us to observe the differences. We conclude that while religiosity and fundamentalism negatively affect support for all three issues, the relationship varies by age and gender. Generally, young men do not show differences in their views of gay marriage, but men, especially religious, young men do show more negative attitudes than their female counterparts in support for gay adoption. Finally, men show more positive attitudes towards gays in the military.

Attitudes toward gays and lesbians are an important topic for social scientists, politicians and policy makers. Many recent studies have documented the increasing acceptance of gays and lesbians in the United States (Greenberg and Bystryn, 1982; Loftus, 2001; Werum and Winders, 2001).
     While many organizations have become more accepting and supportive of gay rights, religions and religious institutions have in general been unsupportive. Young people have therefore come of age caught between increasing support for and acceptance of gays from secular authorities and a strong counter-mobilization from the religious right. Young men in particular constitute an interesting research area, as they show less acceptance and tolerance towards gays than their female counterparts, despite the fact that young people on the whole are more tolerant than the rest of the population. Young men, therefore, seem to be situated at the intersection of two forces: one making them more tolerant, the other less. This paper focuses on young men and explores the effects of age, gender and particularly religiosity on their attitudes towards gay men and lesbians. We investigate the complex relationship of religiosity, masculinity and gay rights and unravel the effects of these seemingly contrasting influences.

Prior research

Prior research consistently shows that attitudes towards gay men and women have generally tended to be negative (Louderback and Whitley Jr., 1997). This anti-gay prejudice has been distinct and well documented in research done with convenience samples from college students (Herek, 1984, 1986; Kite, 1994) as well as large scale, representative surveys (Herek 1991; Herek and Capitanio, 1996; Herek and Glunt, 1993). Overall, attitudes towards gay men and lesbians seem to be improving consistently over time as Americans become increasingly liberal in their opinions about civil liberties (Brooks, 2000). They have gained social acceptance from some parts of the U.S. population, but face opposition from others (Loftus, 2001; Werum and Winders 2001).
     Attitudes towards gay men and lesbians have been explained by numerous factors. Individuals holding negative attitudes towards gay men and women tend to be more authoritarian, less educated, more traditional in sex roles and show negative attitudes towards minority groups (Herek, 1984 and 1991).

Men

One of the most central factors in attitudes towards homosexuality is the sex of the respondent. Many studies show that men on average have more negative attitudes towards gays and lesbians than women (Glenn and Weaver, 1979; Lottes and Kuriloff, 1992; Herek and Glunt, 1993; Kirkpatrick, 1993; Louderback and Whitley, 1997; Marsiglio, 1993; Kerns and Fine, 1994; Kite and Whitley, 1996; LaMar and Kite, 1998; Aberson, Swan and Emerson, 1999; Cotten-Huston and Waite, 2000; Wills and Crawford, 2000; Brown and Amoroso, 1975; Kite and Whitley, 1996; Glassner and Owen, 1976; Gurwitz and Marcus, 1978; Hansen, 1982; Kite, 1984; Laner and Laner, 1979; Millham et al., 1976; Minnigerode, 1976; Steffensmeier and Steffensmeier, 1974; Storms, 1978; Weiss and Dein, 1979). However, even though this sex difference is well documented, few attempts have been made to explain it (Herek, 1988).
     Further inquiries also show that attitudes towards gay men and women differ based on the sex of the target in interaction with the respondent’s sex (Kite and Whitney, 1996). Mary Kite and Bernard Whitley (1996) show that men are more negative towards gay men than women are while there are no differences between men and women in their attitudes towards lesbian women.
     Kite and Whitney (1996) explain this difference based on gender belief systems. Gender belief systems define appropriate behaviors for men and women: people use these gender stereotypes to define what is feminine and masculine, and form opinions about others depending on how well they conform to them. They suggest that attitudes towards homosexuals are shaped by these existing gender belief systems. Because society has more strict expectations of masculinity than femininity (Herek, 1986; Hort, Fagot and Leinbach, 1990), men who display feminine traits receive more negative reaction than women who display masculine traits. We would therefore expect that gay men who violate male gender stereotypes to receive more negative reaction than gay women who violate female gender stereotypes. Furthermore, the more one is invested in the gender belief system, the more one is likely to have a negative view of gays and lesbians because they deviate from the gender norm, possibly forcing one to question the system itself. An

additional explanation could be found in men’s traditional definitions of masculinity. To the extent that gay men differ from heterosexual men’s definitions of traditional, normative masculinities, heterosexual men’s masculine identities might be threatened (Epstein, 1995, 1998; Herek and Capitano, 1999), leading to more negative feelings towards gay men.
     Support for this view of masculinity threat leading to more negative views of gay men can be found in the differential views of men and women towards both gay men and women. In addition to the main effect of sex on attitudes towards homosexuality, where men have more negative attitudes on average towards gays than women do, there is also an indirect effect of sex on attitudes towards homosexuality. While women show no difference in their attitudes towards gay men and women, men have more negative attitudes towards gay men that towards gay women. While gay men seem to threaten heterosexual male’s gender belief systems, lesbianism is seen as erotic and therefore unthreatening (Reiss, 1986; Louderback et al., 1997).

Religion

Religion is an important factor in the understanding of discrimination against gay men and lesbians (McFarland, 1989). First, religious orientation is identified as a factor leading to discrimination, not just against gay men and lesbians, but racial discrimination as well (Allport and Ross, 1967; Batson, 1971). In the now-classic Allport and Ross model (1967), religious orientation is classified as extrinsic, where the individual uses religion to gain “security, comfort, status or social support” (p. 441) or intrinsic, where the individual uses religion only for personal and individual reasons. Similarly, C. D. Batson’s (1971) three factor model classifies religious orientation as Religion as Means, where religion is a means to reach an end, Religion as End, where religion is as an end in itself and finally Religion as Quest, where religion is a way to reach truth. Extensive research shows extrinsic religion is positively related, intrinsic religion unrelated and quest religion negatively related with discrimination (McFarland, 1989). Most of this research focuses on racial discrimination, however. Some prior studies point to the effects of religious orientation on gender discrimination and discrimination towards gays and lesbians; unfortunately, there is little research focusing exclusively on attitudes towards gays and lesbians. Extant research points to higher discrimination against women and gays and lesbians in intrinsic religion (McClain, 1979; McFarland, 1989), suggesting that the relationship between religious categories and attitudes towards gays is similar to that of religious categories and attitudes towards blacks.
     Recent studies that focus exclusively on gay and lesbian discrimination show that some religions are more conservative and less accepting of gays and lesbians than others. Jews, those with no religious affiliation, and inactive Christians have higher rates of gay and lesbian support than Catholics, who are, in turn, more tolerant than Protestants, who show the lowest levels of tolerance towards gays and lesbians (Irwin and Thompson, 1977; Glenn and Weaver, 1979; Henley and Pincus 1978; Lottes and Karloff, 1992; Wills and Crawford, 2000).

Religiosity

In addition to the overall effect of religious orientation, fundamentalism and being a born-again Christian are specifically identified as factors associated with negative attitudes towards gays and lesbians (Herek, 1987). Within the large category of Protestants, fundamentalism – a belief in the literal truth of the Bible – is a particular factor related to attitudes towards gays and lesbians. Herek’s (1987) findings show that fundamentalism increased prejudice towards gays and lesbians. Other studies confirm the effects of fundamentalism on negative attitudes towards gays and lesbians (Wagenaar and Barton, 1977; Herek and Glunt, 1993; Kirkpatrick, 1993; Marsiglio, 1993; Cutton, Hudson and Waite, 2000). Unfortunately, since fundamentalism is not included as a control in other studies (i.e. McClain, 1979; Griffin et al., 1987), it is hard to predict the direct and indirect effects of religion and fundamentalism.
     Finally, in addition to religion and fundamentalism, religiosity or attending services is identified as a distinct factor associated with negative attitudes toward gay men and lesbians. Many studies have found that the more individuals attend services, outside of weddings and funerals, the less tolerant they are of gays and lesbians (Beatty and Walker, 1984; Cochran and Beeghley, 1991, Herek and Glunt, 1993). Randy Fischer et al. (1994), however point out that, for individuals who belong to more progressive and accepting religions, attendance at services has no effect on attitudes towards homosexuality. It remains unclear if the effects are due to the reinforcement of an anti-gay message at less progressive churches, or due to a selection effect, with less tolerant individuals attending church more often.

Political ideology

Religiosity is also closely related to political ideology as both deal with the idea of morality. According to Paul Brewer (2003), public opinion on gays and lesbians is partially explained as an issue of equality (McClosky and Zaller, 1984; Wilcox and Wolpert, 1996 and 2000), partially linked to political party affiliation and political ideology (Haeberle, 1999; Lewis and Rogers, 1999; Wilcox and Norrander, 2002) and partially as a moral issue (Lewis and Rogers, 1999). Therefore, the political affiliation and views of individuals are an important factor in understanding their views on gay issues.

Contact

Furthermore, knowing someone who is gay tends to lead to more positive attitudes towards gays and lesbians (Gentry, 1987; Herek, 1988; Schneider and Lewis, 1984). Further and more recent studies confirm these findings (Ellis and Vasseur, 1993; Herek and Capitanio, 1996; LaMar and Kite, 1998; Cotten-Huston and Waite, 2000; Wills and Crawford, 2000). However, this interpretation has been criticized methodologically because just as having openly gay relatives and friends could make one more likely to have positive attitudes, the direction of causality could be in the other direction, so that it could be argued that people who support gay rights tend to associate with people who are openly gay.
     Further research also points to other correlates such as income, education, geographic region and race and ethnicity (Herek, 1984; Schneider and Lewis, 1984), which are important control factors.

Attitudes and policy areas

In the literature, partly due to data restrictions, attitudes toward gay men and lesbians are generally measured as a unified category, mostly through a feeling thermometer as to how positive or negative one feels towards gay men and lesbians on a scale from 0-100. However, attitude towards gay men and lesbians is not a single unified category. While feeling thermometers are useful, it is important to see the inner differences within the overall attitude. Unpacking policy views in this way will give us crucial insight into the differences in opinion regarding different issues. Different factors may be associated with different concrete aspects of gay related issues: for instance, having gays in the military may be more of a threat to heterosexual men’s masculinity than allowing gays to adopt children. Therefore, we shall look at concrete, policy related issues such as gay marriage, gay adoption and gays in the military.
     Furthermore, each of these issues is separate and divides survey respondents in distinct ways. Rather than grouping them all together, it is important to model approvals and factors explaining opinion in each category separately. This will allow for a more nuanced understanding of attitudes towards gays and lesbians.
     By exploring attitudes towards these issues, this paper focuses on the complex and interrelated relationship between gender, age and religion. The intersection of these areas creates a unique place for young men. While men are traditionally less supportive of gay rights, young people are more supportive. Young men, therefore, are at the center of two opposing social forces. How does belonging to two categories of contrasting views predominate young men’s attitudes towards gays and lesbians? Furthermore, religion and religiosity are central factors in explaining support for gay rights: we will pay special attention to the attitudes of young religious men and unravel the intertwined relationship of gender, age and religiosity in explaining attitudes towards gay men and lesbians.

Methods

Our data come from the Pew Center, which conducts regular national surveys that measure social and political attitudes, values and public attentiveness. Our data come from the recent March 2006 survey on attitudes towards homosexuality. This dataset provides extensive information on factors predicting attitudes towards gays and lesbians, ranging from demographic factors to attitudinal factors, providing the opportunity to estimate a comprehensive model in understanding opinions. It offers a very large, nationally representative sample (n=1405).
     Our aim is to understand attitudes towards gay men and lesbians. As noted before, however, though these attitudes are multi-faceted, most datasets on the topic do not typically include many different variables. The dataset allows us to see differential opinions based on the issue, measured by three dependent variables: approval of gay marriage, approval of gay adoption and approval of openly gay people serving in the military. These three variables were recoded as dichotomous variables, coded 1 if the respondent approved and 0 otherwise.

     Three separate models were estimated, predicting the above dependent variables. Since the dependent variables are dichotomous, logistic regression models were estimated. The independent variables included in the models were uniform to enable comparison. The first set of independent variables included in the model is demographic variables. Sex of the respondent was recoded as a dummy variable where 1=Male and 0=Female. Age of the respondent was asked in years as a continuous variable. However, in addition to age as a continuous variable, a dummy variable for being 18-24 year-olds of age was included (labeled “youth” in the tables presenting the regression results). While the continuous age variable captures the gradual effect of age, the dummy variable should capture any threshold effect. Income, measured in dollars was included, as was race, recoded into a dummy variable as white=1 and non-white=0; Hispanic was coded as 1 if Hispanic and 0 otherwise. Finally, marital status and parenthood are important demographic factors, which could potentially affect attitudes towards gay men and lesbians, especially given their established relationship with authoritarianism (Altemeyer, 1996). Therefore, both these variables were recoded to test for the effects of being married, coded 1 if married and 0 otherwise and being a parent, coded 1 if parent and 0 otherwise. Finally, political affiliation was included as two separate dummy variables: Republican (coded 1 if Republican and 0 otherwise) and Democrat (coded 1 if Democrat and 0 otherwise), leaving political independents as the excluded category.
     In addition to demographic factors which affect attitudes towards homosexuality, we have included a series of factors on religion, religiosity and being a born-again Christian. First, we have coded the religious affiliations of the respondents as dummy variables. In addition to the effects of religious affiliation, religiosity was measured through attending services aside from weddings and funerals, measured in number of times the respondent attends religious services on a weekly basis.
     Finally, being a born-again Christian was included as a separate category, where the respondents who identified themselves as born-again Christians were coded as 1 and 0 otherwise. While being born-again is not exactly the same as fundamentalism, it is a closely related concept, and should be indicative of many of the same attitude structures.
     We have also included attitudinal variables in predicting attitudes towards gay men and lesbians, such as attitudes towards abortion, coded as 1 if approve and 0 otherwise. This inclusion is not intended to imply that views on abortion lead to views on policies relating to homosexuals, but rather to control for general attitudes towards culture war issues (Lindaman and Haider-Markel, 2002). Controlling for these attitudes in such a way allows us to isolate the effects of the other variables specifically on gay rights issues, rather than on the broader category of cultural policy questions. Furthermore, we have included a measure of media exposure, predicting attitudes towards homosexuality such that respondents, who have access to media and are exposed to homosexuality would have more positive views, through access to the Internet coded 1 for access and 0 otherwise.
     In addition to the direct effects of these variables, in explaining attitudes towards gay marriage, gay adoption and gays in the military, we have included a series of interaction effects to capture the interactive effects of these variables

through gender. We hypothesize that being a male, particularly a young male, would affect how some of the above variables would affect attitudes towards homosexuality. For this purpose, we have included interaction effects of Male by white, parent, married, born-again, Republican, Democrat, Internet Access, Abortion Attitudes, Religiosity and Age.
     To explain attitudes towards gay men and lesbians, three logistic regression models were estimated, predicting attitudes towards gay marriage, gay adoption and gays in the military, all using the same independent variables to allow for easy comparison. To isolate the effects of being male and young instead of dividing the dataset and losing sample size, dummy variables were employed instead as well as interactions to capture both direct and indirect effects.

Initial results

First, we looked at the descriptive statistics on attitudes towards gay marriage, gay adoption and gays in the military. While these three aspects all constitute attitudes towards gay men and lesbians in our society, each issue differs in terms of approval rates. Table 1 shows attitudes toward each issue in percentages in the overall population and amongst 18-24 year-old men.

     Among these three, allowing gays to serve openly in the military has the highest approval rate, at 64.4 percent of the overall population, with 20.7 percent strongly favoring. This is followed by gay adoption, with 46.8 percent (13.3 percent

strongly favoring) of the overall population supporting policies that would allow gays to adopt children. This is followed closely by gay marriage, with 39.1 percent of the overall population supporting gay marriage, though only about 10 percent strongly support it.
     When we look specifically at young men between the ages 18-24 – the same group represented by the dummy variable in the logistic regression models – we see that young men have higher levels of support on all three issues. However, their ranking of support follows the same pattern as the overall population, even though their approval in every category is higher. The highest approval rate is for gays in the military, where 67.8 percent of young men support openly gay people serving in the military, almost half of those strongly supporting it. This is followed by gay adoption, where 51.7 percent of young men support gay adoption (19.6 percent of them strongly) and finally 43.5 percent of young men support gay marriage, while an almost equal number, 39.1 percent, strongly oppose it.
     Compared to the overall population, young men between the ages 18-24 show higher support for three issues, but they follow general society’s rankings of these issues. As Figure 1 shows, young men also seem to cluster around more extreme categories: strongly agree, strongly disagree, rather than in the middle categories.

     On all three issues, young men are more likely to “strongly support” gay-friendly policies. Also, the proportion of young men strongly opposing gay adoption and gays in the military are much lower than those in the overall population. The

exception is in gay marriage, where the proportion strongly opposing is rather higher at 39.1 percent, opposed to 32.7 in the overall population.

Logistic regression results

While these descriptive statistics show us initial differences between young men and the overall population, a more accurate picture requires that we isolate the effects of sex and age. The first logistic regression model explains attitudes towards gay marriage.

     Interestingly, there is no independent effect of sex on attitudes towards gay marriage, meaning that, on the whole, men are not less likely than women to approve of gay marriage. Also, there is no threshold effect of age, meaning that the effect of age on approval is relatively linear, and that 18 to 24 year-olds are not

different as a group than members of other age groups. The linear effect of age can be seen in the age variable, which shows that the older people are, the less likely they are to approve of gay marriage. So, while age is a marginally significant predictor, young men do not show differential approval rates.
     Overall, approval of gay marriage is predicted by a combination of demographic, attitudinal and religious variables. Higher income, being a Democrat, being Roman Catholic and favoring abortion increases the likelihood of approving of gay marriage. Being religious, identifying oneself as a born again Christian, being a Republican and being a parent decreases the likelihood of approving gay marriage.
     While men and women show no differences in their attitudes towards gay marriage, the one marginally significant interaction effect is between being male and being a parent. While being a parent makes one less likely to approve of gay marriage, this is not the case for men who are parents: they are more likely to approve of gay marriage. So, overall, when we look specifically at gay marriage and model its approval, we see that young people are more supportive, yet being a male has no significant direct effect.
     Our second model looks at approval of gay adoption. While age was a significant factor in explaining gay marriage – younger people were more likely to approve – age does not seem to be a factor in approval for gay adoption, either as a continuous or threshold effect. Sex also has no direct effect: men and women on average do not seem significantly different in their attitudes towards gay adoption. Therefore, solely being a young male does not seem to lead to differences in approval of gay adoption.
     The direct significant effects seem comparable to attitudes towards gay marriage. The significant predictors of approval of gay marriage are religiosity, where the more often the respondent attends services, the less likely he or she is to approve of gay marriage, being married, where being married makes the respondent less likely to approve of gay adoption, being a Republican, which makes one less likely to approve of gay adoption, being a born-again Christian, which makes one less likely to approve of gay marriage, being Roman Catholic, which makes one more likely to approve and approving of abortion, which makes one more likely to approve of gay adoption. This final result indicates that as with approval of gay marriage, being more liberal on other culture war issues makes it more likely that the respondent will be more liberal with regard to gay adoption. A new significant predictor is Internet access, which makes people more likely to approve of gay adoption. This might be because the Internet provides more exposure and provides more information, leading to more positive attitudes toward gay adoption. However, it could also be interpreted as a self-selection bias, where younger and more liberal people, in ways not captured by other variables, tend to have Internet access.
     While being male has no direct effects on approval of gay adoption, there are many significant and interesting interaction effects. Overall, the less frequently a respondent attends religious services, the more likely he or she is to approve of gay adoption. However, men who attend services are less likely than women who attend religious services to approve of gay adoption.

Table 3: Logistic regression estimates predicting attitudes towards gay adoption. Shaded coefficients significant at α=0.05

     Also, among married people – even though being married makes one less likely to approve of gay adoption – married men, compared to married women, are more likely to approve. Finally, white men are less likely than white women to approve of gay adoption.
     Our final model deals with attitudes towards gays in the military. Here, we see a completely different picture. While being male had no significant effects on the gay marriage and gay adoption, it is a significant predictor of approval of gays in the military, where men are more likely to approve of allowing gays to serve openly in the military. Age, however has no significant effect: there is no difference between older and younger people in their attitudes towards gays in the military.

     Religiosity is a significant predictor of approval of gays in the military. Interestingly, we also detect a significant interaction effect of religiosity and being male. Being a religious man, compared to a religious woman, makes one less likely to approve of gays in the military. Similarly, male parents are less approving of gays in the military than their female counterparts.

Discussion

This analysis yields a complex mapping of relationship between young men’s attitudes towards different gay rights issues and their religiosity. In all three issues, we see religiosity and being born-again as leading to disapproval. Parallel with the

argument made by Glock and Stark (1966), our results show that more religious and fundamentalist people display a more “closed-minded, ethnocentric mindset, which is shown here as a general tendency to discriminate” (McFarland, 1989, p. 333).
     However, men do not have less support for all three issues. When we separate the issues, for gay marriage, gender has no effect: therefore men and women are not different in their attitudes on at least one of the issues. This might also be because of the lack of a linear relationship: it is possible that some men approve and some disapprove, resulting in a curvilinear relationship between gender and support for gay marriage, which requires further research. Age, however, is not as important as we predicted. While it has a marginally positive effect on approval of gay marriage, it does not affect men differently than women.
     When it comes to gay adoption, there is no direct effect of being a man, but there are interaction effects, where religious men are less supportive than religious women and white men are less supportive than white females of gay adoption. However, married men are more supportive than married women. Therefore, in attitudes towards gay adoption, gender is a complex issue. In both issues, however, age does not seem to be important: younger and older men do not seem significantly different.
     Gays in the military is a very different issue than the previous ones, for men. This is the only issue in which being a man leads to more support of gays rights. While men who are religious are less supportive of gays in the military than religious females, the direct effects of gender are in the opposite direction. Such a finding shows that for men, attitudes towards homosexuality are not a unified category, but one that is separated by issues.

Conclusion and future direction limitations

This paper has looked at the attitudes towards homosexuality in the United States and shows they do not consist of one issue, but rather have inner differences based on the issue, which result in different attitudes. Traditional studies have agreed upon the effects of gender, age and religiosity on attitudes towards homosexuality. But rather than simply looking at “attitudes towards homosexuality” as a large, reified category, this paper looks at three main issues: attitudes towards gay marriage, adoption, and gays in the military. Furthermore, rather than making assumptions about young men, this paper looks at young men’s attitudes in-depth and provides a more nuanced understanding. Attitudes towards gay marriage and adoption are not different for men, though men are more likely to support allowing gays to serve openly in the military.
     What is also interesting is to see the interaction effects of men and religiosity: where religious men, rather than religious women, are less likely to support gay adoption and allowing gays in the military. Overall, rather than looking at simplistic relationships between men and attitudes towards homosexuality, we have tried to demonstrate the complexity of the relationship, and the next step would be to unravel why men see these issues so differently through in-depth interviews.

References

Yasemin Besen, Assistant Professor
Gilbert Zicklin, Associate Professor
Department of Sociology, Montclair State University
Montclair, NJ 07043/USA
e: besenymail [dot] montclair [dot] edu

Download this paper (348 KB PDF)
Download Vol. 1 No. 3 (1 MB PDF)

 Previous: The Homosexual Subject: Coming-Out as a Political Act  Top: Volume 1, Number 3: Table of Contents  Next: Review of Victor J. Seidler, Young Men and Masculinities: Global Cultures and Intimate Lives