1 Introduction

Gay neighborhoods have been a familiar part of America’s urban landscape since at least World War II. Famed examples during this time include New York’s West Village and the Castro in San Francisco, though distinct gay districts also surfaced in places like Buffalo, New York; Worcester, Massachusetts; and Columbia, South Carolina (Ghaziani 2014). Even today, despite evidence showing that many historic gay neighborhoods are in decline, there are still numerous residential areas across the country where gay men, lesbians, and other sexual minorities are disproportionately concentrated, as indicated by the percentage of households headed by same-sex couples, a measure based on U.S. Census data (Gates and Ost 2004). These include areas that continue to have businesses and other institutions that specifically cater to sexual minorities, such as bars, bookstores, sex shops, churches, nonprofits, and community centers, as well as areas that may be better characterized as “queer-friendly” (Gorman-Murray and Waitt 2009), in that heterosexuals tend to dominate the residential and commercial spaces, but sexual minority residents, businesses, and organizations are generally welcomed in the neighborhood.

Like ethnic neighborhoods, residential areas with relatively large concentrations of sexual minorities can be understood as places where marginalized people find resources, including protection from discrimination, a sense of community, shared values, and opportunities for social support (Carpiano et al. 2011; LeVay and Nonas 1995; Green 2003). In ethnic neighborhoods, such resources often have important health benefits, especially for mental health (Yuan 2008). For example, studies have found that for some ethnic minority groups, living in areas with high concentrations of people from similar ethnic backgrounds is associated with better mental health and, on some measures, better physical health than living in areas with low concentrations, even when other neighborhood factors are taken into account (Halpern 1993; Pickett and Wilkinson 2008; Stafford et al. 2010). In this chapter we consider whether gay and queer-friendly neighborhoods may have similar health effects for sexual minorities.

In general, sexual minorities are more likely to self-report problems with mental health than other men and women (Meyer 2003). This includes internalizing symptoms, such as depression and low self-esteem (Marshal et al. 2011; Ueno 2010a). The pattern holds not only for persons who identify as sexual minorities, such as gay men and lesbians, but also for those who report same-sex attractions and/or behaviors (Ueno 2010a, b). Although multiple factors appear to contribute to sexual minorities’ poorer mental health (Ueno 2010b), most researchers believe that the stress caused by sexual stigma and prejudice is the biggest factor (Meyers 2003). To the extent that “minority stress” is the main culprit, it seems plausible that living in areas where sexual minority people form a sizable portion of the population will mitigate some of the effects, not to mention the degree of stress exposure. Although the health consequences of these environments may not close the sizable mental health gap that exists between sexual minorities and members of the sexual majority, there may in fact be health benefits to living in areas where sexual minorities are especially numerous.

This chapter explores this issue with a focus on young adults, comparing the mental health of those who live in neighborhoods with relatively high concentrations of sexual minorities to that of those who live in neighborhoods with relatively low concentrations. Further, because many individuals residing in neighborhoods with relatively high concentrations of sexual minorities are not themselves sexual minorities, we also compare the mental health of sexual minority young adults to that of their nonminority, heterosexual peers. In each comparison, we use the proportion of same-sex couple households as a proxy measure of sexual minority neighborhood concentration. Our data come from a nationally representative sample of young men and women, and our analysis controls for both neighborhood- and individual-level factors, including those that may influence neighborhood selection.

2 Background

Gay Neighborhoods and Minority Coping. Meyer’s (1995, 2003) minority stress perspective, which is an elaboration of social stress theory, provides a useful theoretical starting point for thinking about how neighborhoods with higher densities of sexual minorities might promote or protect the mental health of sexual minority young adults. According to this perspective, individuals who belong to stigmatized minority groups have unique, chronic stressors in their lives as a result of their disadvantaged social status. Meyer (1995) refers to these unique psychosocial stressors as “minority stressors” because they are activated when individuals encounter experiences that reinforce their minority status, including prejudice events, such as discrimination and violence, stigma, including expectations of rejection, and the internalization of negative societal attitudes. The contention is that these stressors, which are experienced over and above the routine stressors that all people encounter, increase the likelihood that minority group members will experience mental health disparities. Although Meyer (1995) had self-identified sexual minorities in mind when he first proposed this perspective, similar arguments have been used to explain the poorer psychological well-being observed among persons with same-sex attractions and/or behaviors (Ueno 2010a, b), as well as that of other minority groups, including women, racial-ethnic minorities, and poor people (Amato and Zuo 1992; Kessler and McLeod 1984; Turner and Avison 2003).

This perspective also recognizes that minority individuals have a range of unique resources available to them that may help to alleviate the impact of minority stressors. Meyer (1995, 2003) uses the term “minority coping” to describe any group-level resources that are related to a stigmatized group’s ability to establish self-enhancing structures and values in the face of stigma. As part of the larger social structure, these group-level resources are potentially available to all minority group members, and thus differ from individual-level resources, which vary from person to person (Meyer 2003). From this perspective, the residential clustering found among minority group members can be conceptualized as a collective coping mechanism with possible health promoting and protective effects. This may explain why for certain ethnic minority groups, areas of high ethnic density are associated with lower rates of mental health disparities (Halpern 1993; Stafford et al. 2010). It may be that those who reside in such areas are better shielded from exposure to minority stressors and have more resources to cope with stressors.

Much of the scholarly literature describing neighborhoods with relatively large sexual minority populations mirrors Meyer’s conceptual framework, with its emphasis on minority coping. From early ethnographic accounts of gay and lesbian enclaves to more recent work using U.S. Census data on same-sex partner households, the literature almost uniformly describes these types of neighborhood environments as “safe spaces” for sexual minorities—meaning, places where they can openly express their sexuality, find refuge from sexual prejudice, and meet and form relationships with others without fear (Castells 1983; Frye et al. 2008; Ghaziani 2014; Hayslett and Kane 2011; Weston 1995). In theory, such spaces allow individuals who feel constrained by heteronormativity, whether because of their sexual behavior, attractions, or identity, to experience social environments that challenge the heterosexual status quo. This, in turn, should lessen their likelihood of encountering stigma and other minority stressors. Feeling included and welcome in such spaces also may offset the sense of isolation and difference that many same-sex attracted people face in everyday heteronormative spaces, such as the workplace or school, and may well improve their self-esteem (Finkelstein and Netherland 2005).

Neighborhoods with sizable sexual minority populations also offer greater opportunities for members to develop social networks with one another than what might be possible elsewhere (Finkelstein and Netherland 2005). In turn, these social networks may provide the kind of social support and solidarity that they need to adequately cope with exposure to minority stressors (Ueno 2010b). These social connections may be particularly important for sexual minorities who have little or no family support, an experience which is not uncommon among young adults with same-sex desires and/or behaviors (Needman and Austin 2010). For example, in a qualitative study of young people living in sexual minority enclaves, many participants described their social networks as substitutes for family relationships (Valentine and Skelton 2003). Also, given the relatively large pool of potential same-sex interested partners in such neighborhoods, young people may have more opportunities than elsewhere to realize sexual desires and to date and connect romantically, conditions which may provide additional coping resources (Finkelstein and Netherland 2005).

Finally, many neighborhoods with significant gay and lesbian residential concentration have amenities and services that either cater to or are tolerant of sexual minorities, as well as social events that celebrate sexual diversity (Levay and Nonas 1995). Having these kinds of institutional resources readily available in the neighborhood may reinforce a sense of pride and affirm people’s non-normative expressions of sexuality. Some areas also may have more tangible health resources, including LGBTQ health programs, queer-friendly counseling and support services, and educational workshops on issues related to gay life, including HIV-prevention (Carpiano et al. 2011). In short, because of the greater availability of these types of resources, it is plausible that, when all other relevant factors are controlled, young sexual minorities living in their areas will report better mental health outcomes than their peers who live elsewhere.

This is not to say that neighborhoods with substantial sexual minority populations pose no health risks to sexual minority young adults. For example, many neighborhoods with high densities of sexual minorities also have high levels of sexual orientation-based hate crimes, a finding which raises questions about the level of safety in these locations (Stotzer 2010). In fact, the very visibility of sexual minorities in neighborhoods may aid perpetrators in identifying victims. Further, residence in these areas may include exposure to subcultural groups that engage in risky behaviors, including substance use, heavy drinking, and risky sexual behaviors (Buttram and Kurtz 2012; Carpiano et al. 2011; Green 2003; Kelly et al. 2012). In this respect, sexual minority young adults who live in such neighborhoods may be at greater risk of immersing themselves in a subcultural context that promotes risk taking. Finally, as prior studies show, these neighborhoods are places where young sexual minorities may encounter various forms of social exclusion, including by race, class, and gender (Valentine and Skelton 2003). For example, Green’s (2008) research of a gay enclave in New York City suggests that there may be a collective status order in many gay neighborhoods that strongly favors white men, in addition to those who are young, masculine, and middle-class, As a result, women, nonwhites, and other lower status residents may be more vulnerable to poor mental health outcomes (Green 2008).

Gay Neighborhoods and Heterosexual Residents. Many individuals residing in neighborhoods with high concentrations of sexual minorities are not themselves sexual minorities. In fact, in most instances, nonminority heterosexuals constitute the majority of the neighborhood population (Carpiano et al. 2011). Further, many historically gay neighborhoods, such as the Castro in San Francisco and West Hollywood, California, are undergoing demographic change. Soaring property taxes and rents have driven many sexual minorities out of these areas, while many straight professionals and their families have moved in and replaced them (Ghaziani 2014). It is unclear how these types of environments, if at all, influence the lives of nonminority heterosexuals.

There has been some research on the experiences of straight women in gay-identified venues, such as bars and clubs (Casey 2004; Skeggs 1999). According to this work, these spaces may benefit women by providing them a measure of protection from “the constant male gaze present in heterosexual space,” which can objectify them in potentially threatening ways (Skeggs 1999: 225). On the other hand, heterosexuals may encounter risks, as they are not always welcomed by gay and lesbian patrons; nor do they have the same protections and privileges they experience in other contexts, where heterosexuality is generally assumed and institutionally enforced (Casey 2004). It should also be noted that in historically homophobic societies like the U.S., at least some straight individuals may feel uncomfortable being in areas where sexual minorities are relatively numerous. In short, it may be expected that for heterosexual young adults, living in neighborhoods with relatively high concentrations of sexual minorities will have weak or no effects on their mental health, or possibly even a negative association.

3 Method

Data. The current study uses The National Longitudinal Study of Adolescent Health (Add Health), conducted by the University of North Carolina’s Population Center. The Add Health data rely on a national longitudinal stratified random sample of adolescents enrolled in school beginning when respondents were in grades 7–12, with the first wave collected in 1995 (Tourangeau and Shin 1999). Subsequent waves were collected in 1996, 2002, 2008, and 2016–2018. The third wave, collected in 2002 when respondents were young adults, provides the dependent measures for our analyses. Advantages of this data set include a very large nationally representative sample, which is particularly necessary for this project as it allows for the identification of a sufficient number of sexual minorities. The data also provide information on neighborhood (i.e., Census tracts) characteristics by linking respondents’ addresses to Census data.

The final sample (N = 13,888) used in these analyses was restricted to respondents who had valid weights, valid data in the neighborhood characteristics used, and valid data on all of the dependent measures. In the final sample, 13% (f = 1875) of respondents are classified as sexual minority young adults, 47% of respondents are male, 18% are Latino/Hispanic, 21% are non-Hispanic Black, 1% non-Hispanic Native American, 7% non-Hispanic Asian American, and about 52% non-Hispanic White. The average age is 22 years old and ranges from 18 to 28.

Dependent Variables. All dependent measures are from the Wave III data. The first mental health outcome examined is a measure of “self-esteem.” This 4-item scale is comprised of items asking respondents how much they agree with the following statements as representative of the past 7 days: I have many good qualities, I have a lot to be proud of, I like myself just the way I am, I have been doing things right. Responses on individual items, ranging from 1 = strongly agree to 5 = strongly disagree, were reverse coded so that higher scores indicate higher levels of self-esteem and then averaged (range: 1–5; M = 4.2, SD = 0.6, α = 0.79). The second indicator of mental health is an index of “depressive symptoms,” comprised of 9 items asking respondents how often certain things were true in the past week (e.g., couldn’t shake off the blues, felt too tired to do things, felt sad, were bothered by things that usually don’t bother you, etc.). Responses on individual items, ranging from 0 = never or rarely to 3 = most or all of the time, were averaged (reverse coded when necessary) so that higher scores indicate higher levels of depressive symptoms (range: 0–3; M = 0.5, SD = 0.4, α = 0.81).

Wave I versions of the dependent measures were included in regression models to control for initial levels. Including the lagged measures of the dependent variables means that regression coefficients should be interpreted as effects on (or multivariate associations with) change in self-esteem and depression symptoms.

Focal Independent Variables. One of the two main predictors of interest in this study is whether or not respondents are “sexual minority young adults” (SMYA). In the most general sense, a sexual minority is an individual who has experience with same-sex sexuality, whether at the level of attraction, behavior, or identity. The term reflects the fact that regardless of how one self-identifies, any experience with same-sex sexuality violates societal norms prescribing exclusive heterosexuality, thereby making that person a sexual minority (Diamond 2008). For this study, we use 3 measures of sexual minority status. (1) Respondents can indicate that they self-identify as 100% heterosexual, mostly heterosexual, bisexual, mostly homosexual, or 100% homosexual. Respondents answering something other than 100% heterosexual were classified as SMYA. (2) If respondents were currently involved in or had been involved in a sexual or romantic relationship since Wave II, they were asked the sex of their partner. If the partner was of the same sex, they were classified as SMYA. (3) Respondents were also asked if they had ever been romantically attracted to a male, and separately, to a female. If they answered “yes” in regard to the same sex, they were classified as SMYA.

The other focal predictor for this study is the extent to which respondents live in neighborhoods with a relatively high concentration of sexual minorities. To ascertain this information, we use a U.S. Census measure of the proportion of same-sex couple households in a neighborhood (i.e., Census tract). Respondents live in neighborhoods that range from 0 to 0.19 (19%) same-sex partner headed households, with the average respondent living in a neighborhood that is 0.0058 (SD = 0.008) same-sex partner headed households (0.58%). In our sample, 53.1% live in neighborhoods with no same-sex couple households, 46.9% live in neighborhoods with at least 1% same-sex couple households (with 38.9% living in tracts with 1%), 8% in neighborhoods with at least 2% same-sex couple households (with 6.3% living in tracts with 2%), and 1.7% live in neighborhoods with 3% or more. If these percentages seem low, it is because we are using a proxy measure of sexual minority neighborhood concentration. The Census does not ask about sexual orientation directly, so it leaves out sexual minorities without partners, those who do not live with their partners, and those unwilling to report living with a same-sex partner. This results in a likely underestimation of the proportion of sexual minorities in neighborhoods. On the other hand, according to the 2010 Census, same-sex partner households account for just over half of one percent of all households in the U.S. (Kolko 2012). Thus, even a neighborhood with just 3% same-sex couple households is nearly 6 times the national average. In fact, a neighborhood with a concentration of sexual minorities of that size would be on par with other, more researched types of neighborhood concentration, including ethnic concentration (Spring 2013).

Other Neighborhood-Level Predictors. The proportion of same-sex couple households in a neighborhood may be associated with other neighborhood characteristics that influence mental health. Thus, we used Census measures to control for other significant neighborhood characteristics that may be associated with the proportion of same-sex households. Following precedent (Carpiano et al. 2011), we measured the relative concentration of neighborhood economic disadvantage with a “concentrated disadvantage” index by using a weighted factor score variable based on the proportion of the population over age 16 who are unemployed, the proportion receiving public assistance, the proportion over age 25 without a high school diploma, and the proportion living below the poverty level. Further, to assess the relative rate of residential turnover, we created the variable “residential instability,” which indicates the percentage of the population that has moved in the last 5 years. To ease interpretation of regression coefficients, we collapsed the percentage of those who had moved into deciles (1 = 0 to 9.9%, 2 = 10 to 19.9%, etc.). We also included a measure of the percentage of residents who reside in an urban area (in each tract), collapsed into deciles.

Concurrent Individual-Level Predictors. We also included several Wave III individual-level measures. One variable was the respondents’ relationship status—that is, whether or not they are “currently in a relationship.” Other variables were created in lieu of more conventional socioeconomic status variables. As others have noted (Booth et al. 2012), there are unique challenges involved in measuring the socioeconomic status of young adults. To crudely capture their financial status, we included a dummy coded variable indicating whether or not the respondent receives any public assistance. To measure economic potential, we included a dummy coded variable indicating that the respondent attained at least a junior college degree. We also include a variable coded 1 if the respondent lives with a parent and 0 if no parent figure is recorded in the household roster, because while many were independent in their twenties some were still living with a parent, which could impact mental health. Finally, because we are essentially examining change in self-esteem and depression symptoms between adolescence and young adulthood, we include a dummy coded control for whether or not the respondent moved (0) or still lives in the same house as indicated in Wave I.

Other individual-level factors were included simply as demographic controls. These controls include age, sex (dummy coded into Male = 1 if male; Male = 0 if female); and a dummy set measuring racial/ethnic identity. Hispanic or Latino ancestry was coded 1 if respondents indicated such and 0 if not; Latino ethnicity took coding priority as respondents could identify with any racial group. Most respondents identified with one racial group (non-Hispanic Black, non-Hispanic Native American, non-Hispanic Asian, and non-Hispanic White). Those who identified as multi-racial were subsequently asked which single category best defined them.

Selection Factors. We also included several controls for possible selection into neighborhoods with varying concentrations of same-sex couple households. Specifically, we considered factors that may be important in determining whether or not sexual minority young adults choose to reside in neighborhoods with higher densities of other sexual minorities. In our study, young adults aged 18 to 28 may have transitioned between Wave I and Wave III from a home with a parental figure to one without. Using data from the parent interview at Wave I, we measured whether the responding parent was employed or not (1 = employed) and whether or not the responding parent received any public assistance (1 = received assistance). A measure of the frequency of parental alcohol use is included and ranges from 1 “no alcohol use” to 6 “nearly every day.” To tap into parental concerns about their residence, we include the parent’s perception of how crime ridden their neighborhood was on a scale of 1 “no problem” to 3 “big problem.” We also include a proxy for family of origin’s economic potential that is the average educational attainment of the responding parent and their spouse (if no spouse, we used the parent’s education alone). Parental educational attainment was coded 1 “no formal schooling or 8th grade and less” to 6 “professional training beyond 4 year university.” Finally, using the youth’s Wave I report, we include a measure of positive parental relationship. This index is the average of three items regarding the how frequently the parent-youth relationship is warm and loving, how satisfied the youth is with the way they communicate, and their overall satisfaction with the parent-youth relationship. Reverse coded items were scored so that higher scores indicate greater agreement (warmer relationship) where 1 was “strongly disagree” and 5 was “strongly agree.” For most respondents this index taps their relationship with mothers. If data on the mother was missing, data on the father-respondent relationship were utilized.

We also included a Wave III measure of the incidence of violent victimizations in the last year, as it may influence residence choice. It is also associated with our dependent measures (see for example, Meyer 1995; Ueno 2010b). Violent victimization is a summed index of 6 items tapping the number of different incidents experienced, including whether someone pulled a gun on the respondent, someone pulled a knife on them, someone shot them, someone stabbed them, someone beat them up without robbing them, and someone beat them up and robbed them. Respondents who said “don’t know” to individual items were assigned the mode of no for those items which may result in an undercount of victimizations while respondents with missing data were omitted. The final index ranges from 0 to 6; on average respondents experienced 0.13 different incidents (SD = 0.54).

Analytical Strategy. Descriptive, bivariate, and regression analyses using OLS (ordinary least squares) are presented. Given the complexity of the data’s sampling method, all appropriate sample and individual-level weights are used in all analyses (weights for strata, cluster, and individuals). Stata 13 was used to run the regression analyses. We first estimated equations for the total sample and include a product term for concentration of same-sex headed households centered at its mean (which is essentially 0) and sexual minority young adult status and all other variables. Subsequently, we estimated separate equations for sexual minority and sexual majority young adults. To determine if coefficients are significantly different across equations by SMYA status, we re-estimated equations for the full sample and included product terms for SMYA status and all other variables in the model. Significant product terms (where t values are significantly larger than by chance) indicate that the coefficients displayed in the separate equations are significantly different from each other; these differences are noted in Table 8.3 with bold (p < 0.05) and bolded italicized font (p < 0.01). This method produces results similar to z-tests for the equality of regression coefficients across equations (Paternoster et al. 1998). Lagged levels of the outcome measures, initial levels during adolescence, are included in the model allowing us to focus upon change in depression, self-esteem, logged drug use, and logged excessive drinking.

4 Results

Table 8.1 presents the sample characteristics, as well as compares sexual minority and sexual majority young adults in terms of those characteristics. We can see that sexual minority young adults differ significantly from their sexual majority peers in several important ways. Our data shows that SMYA have significantly higher levels of depressive symptoms and significantly lower levels of self-esteem (see Table 8.1). The patterns we see when the respondents were in their 20s also appear to have existed in adolescence. SMYA reported significantly higher depression and significantly lower self-esteem at Wave 1.

Table 8.1 Sample description

In our sample, sexual minority and sexual majority young adults are equally likely to be in a current relationship (59%), have at least a two-year college degree (19%), and be on public assistance (15% and 13%, respectively). SMYA are less likely than their sexual majority peers to still live with a parent and are more likely to have moved since Wave I. They are also significantly more likely to live in urban neighborhoods and neighborhoods where there is more population mobility but where there is slightly less concentrated disadvantage.

In our sample, sexual minority young adults are significantly and substantially more likely to be female (70%) than sexual majority young adults (50%). This is consistent with other research which shows that young women are more likely than young men to report same-sex attractions and same-sex behaviors (Diamond 2008; Ueno 2010a, b). SMYA report a slightly lower average for positive maternal relationship. Their parents self-reported slightly higher levels of alcohol use at Wave I and levels of educational attainment. Parents of sexual minority and sexual majority young adults reported equal perceptions of how crime ridden their Wave I neighborhoods were, and were equally likely to be employed and on public assistance.

Table 8.2 presents the weighted least squares regressions of 2 dependent variables on a set of contemporaneous correlates including neighborhood characteristics and demographics, Wave I controls including parental support and social characteristics, and a lagged version of the dependent measure to examine how proportion of same-sex households, SMYA status, and their interaction affect change in mental health. As seen in Table 8.2, key bivariate differences by sexual minority status remain significant in the multivariate context. Sexual minorities report a greater increase in depression symptoms and lower self-esteem, net all other variables. The product term for sexual minority young adult status and proportion same-sex headed households suggests a differential effect of the latter for sexual minority and sexual majority young adults in the case of depression symptoms and self-esteem. We explore this difference and others in separate equations for sexual minority and sexual majority young adults in Table 8.3.

Table 8.2 Weighted OLS regressions of 2 health indicators on sexual minority young adult status, neighborhood concentration of same-sex headed households and other factors
Table 8.3 Separate weighted OLS regressions of 2 health indicators for sexual minority and majority young adults

Our primary interest was in the association between the proportion of same-sex headed households and our 2 outcome measures. As seen in Table 8.3, neighborhood concentration of same-sex headed households significantly decreases depressive symptoms for sexual minority young adults and significantly increases self-esteem for sexual minority young adults, net adolescent levels of depression and self-esteem and the effects of all other variables. A one unit increase in the proportion of households headed by same-sex couples decreases depressive symptoms by 3.2 for sexual minority young adults and has no effect for sexual majority young adults. The difference in these two coefficients is statistically significant (t = 2.22, p < 0.05). A one unit increase in the proportion of households headed by same-sex couples increases self-esteem for sexual minority young adults by 6.1 and is not significantly related to self-esteem for sexual majority young adults. Again this difference across equations is statistically significant according to the full sample model with all SMYA interactions included.

5 Discussion

The results from this study suggest that gay and queer-friendly neighborhoods are important contexts for understanding the mental health of sexual minority young adults. While prior studies have shown that sexual minority young adults experience poorer mental health than their sexual majority peers, including higher levels of internalizing symptoms, such as depression and low self-esteem (Marshal et al. 2011; Ueno 2010a), our study finds that this relationship may depend on the characteristics of the neighborhood environments in which they live. Specifically, the proportion of same-sex partner households in a neighborhood appears to influence the degree to which sexual minorities, but not heterosexuals, report poorer mental health, over and above the influence other neighborhood- and individual-level factors.

Our analysis shows that for sexual minority young adults, living in neighborhoods with higher concentrations of same-sex couples is associated with significantly better mental health outcomes than living in neighborhoods with lower concentrations of same-sex couples. Specifically, we found that those who live in areas with higher densities of sexual minorities have lower rates of depression symptoms and higher levels of self-esteem. Conversely, for heterosexual young adults, we found no association between the proportion of same-sex couples in a neighborhood and mental health outcomes. Thus, it appears that only sexual minorities are advantaged by living in neighborhoods where same-sex couples are more densely concentrated.

Neighborhood selection factors, such as the respondents’ past experiences with victimization, their level of parental support, parental perceptions of neighborhood crime, and their parents’ financial status, cannot account for the lower rates of depression symptoms and higher rates of self-esteem found among sexual minority young adults who live in neighborhoods with heavier concentrations of same-sex couples. Nor can other individual-level factors, such as race-ethnicity, education, or the respondents’ relationship status. This suggest that there is something about the neighborhood environment, rather than characteristics of the individuals, that explains why sexual minorities living in neighborhoods with higher levels of same-sex couples report better mental health. Given that we controlled for other neighborhood-level factors, including the relative concentration of economic disadvantage, it seems likely that the presence of other sexual minorities in and of itself is the driving protective factor.

Why might living in neighborhoods with higher concentrations of same-sex partner households lead to better mental health outcomes among sexual minority young adults? Although we cannot definitely answer this question based on the data used in this study, our results are consistent with the assumptions of the minority stress perspective, with its emphasis on minority coping (Meyer 1995, 2003). This perspective asserts that while sexual minorities have unique stressors in their lives as a result of their disadvantaged social status, they also have unique resources to cope with stressors, including potentially protective and health promoting neighborhood environments. As is the case in some ethnic minority neighborhoods (Stafford et al. 2010; Yuan 2008), neighborhoods with relatively large concentrations of sexual minorities may function protectively to generate resiliency in the face of minority stress, with potentially positive consequences for mental health. Sexual minority young adults who live in such neighborhoods may be better shielded from exposure to minority stressors, such as discrimination and violence, and have more resources to deal with stressors, such as social and institutional support. The implication is that sexual minorities who choose to live in neighborhoods with large numbers of other sexual minorities are not just acting out their personal preferences; they also may be seeking the health benefits that these neighborhoods confer (Valentine and Skelton 2003). Then again, even if some sexual minorities choose to live in such neighborhoods for other reasons, such as economic or cultural, they may still reap health benefits by virtue of their proximity.

This is not to say that neighborhoods with substantial sexual minority populations pose no health risks to sexual minority young adults. As prior studies have shown, gay and queer-friendly neighborhoods also may be places where young sexual minorities may encounter antigay violence, subcultural norms that promote risky behaviors, including substance use, heavy drinking, and risky sexual behaviors, as well as various forms of social exclusion, including by race, class, and gender (Buttram and Kurtz 2012; Carpiano et al. 2011; Green 2008; Kelly et al. 2012; Stotzer 2010; Valentine and Skelton 2003).Yet, whatever risks young people may face in these kinds of neighborhoods, they do not appear to have a negative effect on their level of depression symptoms or self-esteem. In this regard, the risks of living in areas with higher than average sexual minority populations do not outweigh the benefits, at least not for sexual minority young adults.

6 Limitations and Conclusion

This study involved several limitations. First, for practical reasons related to data availability, we determined the proportion of the neighborhood population that is composed of sexual minorities using a census-driven measure of the percentage of same-sex partner households. This is a crude proxy that measures sexual minority population concentration indirectly, as it leaves out sexual minorities without partners, those who do not live with their partners, and, as in the case of all surveys, those unwilling to identify themselves. Unfortunately, the Census only collects data on the residential patterns of same-sex partner households, resulting in a significant underestimation of the presence of sexual minorities in any given neighborhood. On the other hand, while the Census may not be ideal, the fact that it significantly underestimates the extent of sexual minority neighborhood concentration gives us greater confidence that the neighborhood effects we did find in this study are robust.

Second, while our sample included a sizeable number of sexual minority young adults (n = 1875), the subsample living in neighborhoods with relatively high concentrations of same-sex partner households was quite small. Small sample size reduces statistical power and makes it difficult to detect group differences in the population. Further, sample size limitations precluded us from exploring other potential variations among sexual minorities, including by gender, racial-ethnicity, and class. When possible, future work should consider how sexuality intersects with other social statuses. Studies also need to consider if variations exist among sexual minorities depending on their sexual identity. Not all young people with same-sex attractions or behaviors adopt a sexual minority identity, such as “gay” or “bisexual,” and there may be differences in neighborhood effects between those who do and do not.

Third, the age range of our sample was restricted to 18- to 28-year-olds. Thus, while the findings presented here may generalize to this particular age group, it remains to be seen whether the same effects will be found in older populations. We also dealt with a single age cohort. Neighborhoods with higher concentrations of sexual minorities may have different effects on different cohorts, including future cohorts. For example, there is some evidence that historic gay neighborhoods are on the decline (Ghaziani 2014). If so, this may alter their impact on the mental health of sexual minorities in the future.

Finally, despite our attempts to control for potential selection effects, we cannot be certain of the causal direction between neighborhood residence and mental health. Although we believe that it is more plausible that neighborhood contexts influence mental health, it is also possible that these associations reflect the selection of persons into neighborhoods based on other characteristics related to mental health. For example, with respect to sexual minority young adults, it is possible that healthier sexual minorities are “selected into” neighborhoods with higher densities of same-sex headed households, while their less healthy peers are “selected out.” On the other hand, the reverse could also be operating, which would have significant implications for thinking about the health effects of high-density sexual minority neighborhoods. While we were able to control for whether the respondents moved between Waves I and III, we could not track when Wave III depressive symptoms and self-esteem scores started vis-à-vis the move to the Wave III residence.

Despite these limitations, our study provides suggestive evidence that gay and queer friendly neighborhoods—i.e., residential tracts with relatively large concentrations of sexual minority residents—have a positive impact on the mental health of sexual minority young adults, above and beyond the influence of their individual characteristics. Our study thus underscores the importance of striving for contextual understandings at the neighborhood level of sexual minorities and their mental health problems. Future work should consider exploring the mechanisms underlying the protective association between neighborhoods with relatively high concentrations of same-sex couples and mental health, and if the mechanisms at work in these areas are similar to those found in ethnic minority neighborhoods.