Skip to main content
NCBI home page
Psychological Science logoLink to Psychological Science
. 2021 Sep 20;32(10):1684–1696. doi: 10.1177/09567976211018624

A Contextual Approach to the Psychological Study of Identity Concealment: Examining Direct, Interactive, and Indirect Effects of Structural Stigma on Concealment Motivation Across Proximal and Distal Geographic Levels

Micah R Lattanner 1,✉, Jessie Ford 2, Na Bo 3, Wanzhu Tu 4, John E Pachankis 5, Brian Dodge 6, Mark L Hatzenbuehler 7
PMCID: PMC8907493  PMID: 34543132

Abstract

Psychological theories of identity concealment locate the ultimate source of concealment decisions within the social environment, yet most studies have not explicitly assessed stigmatizing environments beyond the immediate situation. We advanced the identity-concealment literature by objectively measuring structural forms of stigma related to sexual orientation (e.g., social policies) at proximal and distal geographic levels. We linked these measures to a new, population-based data set of 502 gay and bisexual men (residing in 44 states and Washington, DC; 269 counties; and 354 cities) who completed survey items about stigma, including identity-concealment motivation. Among gay men, the association between structural stigma and concealment motivation was (a) observed across three levels (city, county, and state), (b) conditional on one’s exposure at another geographic level (participants reported the least motivations to conceal their identity if they resided in both cities and states that were lowest in structural stigma), and (c) mediated by subjective perceptions of greater structural stigma.

Keywords: structural stigma, concealment, sexual orientation, open materials

Psychological research has documented the unique predicament of identity concealment among the stigmatized (Quinn & Chaudoir, 2009). Whereas concealing one’s identity confers protection against discrimination (Pachankis & Bränström, 2018), it can also increase risk for numerous adverse psychosocial and health outcomes (Pachankis, 2007). Psychological theories of identity concealment locate the ultimate source of concealment decisions within the social environment (Pachankis, 2007), yet these environments have been measured almost exclusively at the interpersonal level, typically in the form of stigma-related cues in a given situation (Newheiser & Barreto, 2014). This research has been important, but these studies offer an incomplete test of psychological theories of identity concealment because they did not explicitly measure stigmatizing environments beyond the immediate situation.

In the past decade, researchers have begun to measure these stigmatizing social environments in the form of structural stigma, which refers to “societal-level conditions, cultural norms, and institutional policies that constrain the opportunities, resources, and wellbeing of the stigmatized” (Hatzenbuehler & Link, 2014, p. 2). Accumulating evidence indicates that structural stigma is associated with a range of adverse outcomes, including psychological distress, psychiatric disorders, self-stigma, and social isolation (Hatzenbuehler, 2014, 2016, 2017). Recently, studies have begun to document associations between structural stigma and concealment outcomes in particular. Individuals with HIV/AIDS (Miller et al., 2011) and sexual minorities (Pachankis & Bränström, 2018; Pachankis et al., 2015; Riggle et al., 2010) report greater disclosure concerns and more identity concealment in communities with higher levels of structural stigma (e.g., communities in which individuals report more HIV-related prejudice, countries that lack legal protections based on sexual orientation). However, critical questions remain regarding the multilevel relationship between structural stigma and identity concealment. Our study advances the previous literature in three ways.

First, prior studies have examined objectively measured structural stigma at a single level, such as European countries (e.g., Pachankis et al., 2015). However, aggregation to only one level ignores heterogeneity at each geographic level. The inclusion of more than one level is particularly important in the United States, given that many cities have policies and attitudinal contexts that differ markedly from those at the state level. We overcame this limitation by assessing the relationship between objectively measured structural stigma and identity-concealment motivation across three geographic units: states, counties, and cities. This approach also affords the opportunity to answer new questions about how different levels of structural stigma interact. For instance, does living in a structurally supportive city attenuate identity concealment, even when one lives in a structurally stigmatizing state?

Second, if objectively measured structural stigma increases motivations to conceal one’s identity, it raises the question of what psychological pathways underlie this association. One possibility is through an appraisal pathway, as anticipated by minority-stress theory (Meyer, 2003) and the cognitive-affective-behavioral model of concealment (Pachankis, 2007). According to these theories, individuals with stigmatized identities are likely aware that the citizens of their states and cities hold negative attitudes about their group and that they have implemented policies that target them for social exclusion. Although subjectively perceived structural stigma is a plausible mechanism underlying identity concealment, we are unaware of any study that has tested this hypothesis by simultaneously measuring structural stigma both objectively and subjectively. We therefore created a new measure of perceived structural stigma—defined as the perception that one’s rights, opportunities, and inclusionary status are restricted because of the laws, policies, and normative climate of one’s social environment—and examined it as a mechanism linking objectively measured structural stigma to concealment.

Statement of Relevance.

Concealing a stigmatized identity provides protection against discrimination but also increases risk for adverse health outcomes. Most psychological research has examined whether cues in the immediate situation drive concealment decisions. Here, we consider whether decisions to conceal one’s identity are also influenced by stigmatizing features of the broader social context. We objectively measured structural forms of stigma related to sexual orientation (e.g., social policies) across three geographic levels—states, counties, and cities—and examined their association with identity concealment among gay men. We found that the association between structural stigma and identity concealment was apparent at each level; further, this association was conditional on exposure to structural stigma at another level: Concealment was lowest when participants resided in cities and states with the least structural stigma. Finally, we identified a psychological mechanism explaining why structural stigma increases concealment. These results advance understanding of the psychological consequences of living in high-stigma contexts.

Third, the few studies that have evaluated stigma across individual and structural levels (e.g., Miller et al., 2011; Pachankis & Bränström, 2018) have all used nonprobability samples, which can introduce selection biases. For example, studies that recruit sexual minorities may result in samples that are more open about their sexual orientation (Kuyper et al., 2016), which could lead to a restricted range of identity-concealment behaviors. We addressed this limitation by recruiting a new population-based sample of gay and bisexual men to enhance generalizability.

Drawing on research on structural stigma (Hatzenbuehler, 2016) and identity concealment (Pachankis, 2007), we examined the relationship between structural stigma and concealment motivation (i.e., the motivation to prevent one’s identity as a sexual minority from being known by others). Whereas prior work has documented associations between structural stigma, identity concealment and disclosure, and outness (e.g., Miller et al., 2011; Pachankis & Bränström, 2018), the present research focused instead on concealment motivation, which captures the cognitive, affective, and motivational antecedents to subsequent concealment behaviors and health outcomes (Mohr & Kendra, 2011). We tested the following hypotheses at each geographic level of analysis. Objectively measured structural stigma will be associated with greater concealment motivation (Hypothesis 1), environments objectively higher (vs. lower) in structural stigma will be subjectively appraised as more stigmatizing (Hypothesis 2), subjective appraisals of structural stigma will be associated with greater concealment motivation (Hypothesis 3), and subjectively appraised structural stigma will mediate the association between objectively measured structural stigma and concealment motivation (Hypothesis 4). Additionally, we tested interactions between state- and city-level structural stigma in predicting concealment motivation. Although we expected that participants would report the least motivation to conceal if they resided in both cities and states that were lowest in structural stigma, we were less certain about how other combinations of structural stigma would relate to concealment motivation. For instance, it is possible that signals of structural support at distal (i.e., state) levels could reduce motivations to conceal one’s identity even in proximal (i.e., city) environments with higher levels of structural stigma; conversely, if proximal environments matter most, concealment motivation would be similar across states with high and low structural stigma. Because this is the first article, to our knowledge, to examine these questions, we did not make a priori hypotheses regarding these interactions.

Method

Data collection

Our data come from the Ipsos (formerly GfK) KnowledgePanel, an online panel that is representative of the adult U.S. population. KnowledgePanel employs an address-based sampling methodology, which allows probability-based sampling of addresses from the Delivery Sequence File of the U.S. Postal Service. Address-based sampling is a statistically valid sampling method with a published sample frame of residential addresses that covers approximately 97% of U.S. households. Samples from KnowledgePanel cover all households regardless of their phone status.

Participants

KnowledgePanel consists of approximately 50,000 adult members (age 18 years and older), of which approximately 900 identify as gay and bisexual men. We partnered with Ipsos to recruit a sample of gay and bisexual adult men (age 18 years and older, English- and/or Spanish-language speakers) to become part of the National Study of Stigma and Sexual Health. Participants were paid $25 (U.S.) in Ipsos incentives as compensation for their time. The final sample consisted of 502 sexual-minority men (gay: n = 371; bisexual: n = 131). On average, participants identifying as gay (M = 53.48, SD = 14.36, range = 19–91) and bisexual (M = 53.05, SD = 16.02, range = 18–84) were similar in age. Gay men were predominantly non-Hispanic White (73.85%), followed by Hispanic (15.09%), non-Hispanic Black (6.74%), non-Hispanic other (2.16%), and non-Hispanic multiracial (2.16%). Bisexual men were predominantly non-Hispanic White (75.57%), followed by Hispanic (15.27%), non-Hispanic Black (4.58%), non-Hispanic other (2.29%), and non-Hispanic multiracial (2.29%). Additional sociodemographic information on the subsample of gay men is provided in Table 1, and sociodemographic information on the subsample of bisexual men is provided in SOM-R Table 2.1 in the supplementary online material (SOM), available on OSF at https://osf.io/ytcsu/.

Table 1.

Descriptive Statistics for the Analytic Subsample of Gay Men (n = 371)

Variable Value Observed range
Age (years) M = 53.48 (SD = 14.36) 19–91
Relationship status
 Single n = 219 (59.03%)
 In a relationship n = 152 (41.97%)
Race/ethnicity
 White, non-Hispanic n = 274 (73.85%)
 Black, non-Hispanic n = 25 (6.74%)
 Other, non-Hispanic n = 8 (2.16%)
 Hispanic n = 56 (15.09%)
 Two or more races, non-Hispanic n = 8 (2.16%)
Urbanicity
 Metro n = 345 (92.99%)
 Nonmetro n = 26 (7.01%)
Education
 Less than high school n = 7 (1.89%)
 High school n = 33 (8.89%)
 Some college n = 117 (31.54%)
 Bachelor’s degree or higher n = 214 (57.68%)
Perceived structural stigma: proximal environment M = 2.11 (SD = 0.67) 1.00–4.00
Perceived structural stigma: distal environment M = 2.19 (SD = 0.67) 1.00–4.00
Concealment motivation M = 2.31 (SD = 0.95) 1.00–4.00
Objective city-level structural stigma (k = 489) M = 0 (SD = 0.91) −1.60 to 1.41
Objective county-level structural stigma (k = 2,990) M = 0 (SD = 0.82) −3.77 to 5.38
Objective state-level structural stigma (k = 51) M = 0 (SD = 0.94) −2.74 to 1.44
Open in a new tab

Note: See SOM-R Table 2.1 in the supplementary online material (SOM), available on OSF at https://osf.io/ytcsu/, for descriptive statistics for the subsample of bisexual men (n = 131).

Because the lesbian, gay, and bisexual (LGB) population has not been appropriately enumerated in the United States (given that the U.S. Census does not currently assess sexual orientation), no study, including our own, can state with certainty that it has a nationally representative sample of LGB individuals. Nevertheless, the strength of our approach is that we used a randomly selected sample drawn from the general population. This approach minimizes selection biases that can occur when nonprobability samples of LGB individuals are used. This sampling approach was particularly important for our research question, given that nonprobability samples may overrepresent LGB individuals who are more open about their sexuality, which could lead to a restricted range of identity-concealment behaviors. For this reason, the use of probability-based samples is the gold standard of sampling for survey research with LGB populations (Meyer & Wilson, 2009, p. 24).

Measures

Dependent variable

Participants completed the three-item Concealment Motivation subscale of the Lesbian, Gay, and Bisexual Identity Scale to measure motivation to prevent one’s identity as a sexual minority from being known by others (e.g., “I keep careful control over who knows about my same-sex romantic relationships”; Mohr & Kendra, 2011). These items are rated on a 4-point Likert-type scale ranging from 1 (do not agree) to 4 (agree strongly). We computed an average concealment-motivation composite score (M = 2.31, SD = 0.95); higher scores represent a higher motivation to conceal one’s sexual orientation (Cronbach’s α = .87).

Mediating variable

Participants completed a new eight-item scale designed to measure perceived structural stigma. Because no measure of this construct exists, these items were developed for the current project. Items were created on the basis of interviews conducted with lesbian, gay, bisexual, and transgender (LGBT) individuals across the United States (Bruni, 2017) that addressed broad dimensions of structural stigma (i.e., societal-level conditions, cultural norms, and institutional policies). Because subjective appraisals of structural stigma may differ with respect to one’s proximal environment (e.g., cities and counties) compared with one’s distal environment (e.g., state), we measured perceived structural stigma for both proximal and distal environments. This approach also allowed us to assess the relationship between objectively measured structural stigma at a particular geographic level (i.e., state, county, and city levels) and subjectively appraised structural stigma at a corresponding geographic distance (i.e., distal or proximal social environment). Two items on the scale were tailored to the distal social environment (i.e., state), and two items were tailored to the proximal social environment (i.e., town or city; e.g., “There are LGBT laws in my state [town/city] that I would like to see changed”). The other four items were tailored to the social climate in general. For the full perceived-structural-stigma scale, see Appendix 1 in the SOM. Items were rated on a 4-point Likert-type scale ranging from 1 (do not agree) to 4 (agree strongly); higher scores represent a stronger perception of structural stigma. For the current study, we removed one item, “I can be 100 percent open about being gay in my community,” because of conceptual overlap with our dependent variable of interest, concealment motivation. We computed an average score from the two state-level items and the three general-social-climate items to form a five-item composite score indexing distal-environment perceived structural stigma (M = 2.11, SD = 0.67), which was used for state-level analyses (Cronbach’s α = .67). We also computed an average score from the two city-level items and the three general-social-climate items to form a five-item composite score indexing proximal-environment perceived structural stigma (M = 2.19, SD = 0.67), which was used for city- and county-level analyses (Cronbach’s α = .68).

Independent variables

Following prior research (Hatzenbuehler & McLaughlin, 2014; Pachankis & Bränström, 2018) and the conceptualization of structural stigma (Hatzenbuehler & Link, 2014), we quantified structural stigma by statistically combining indicators of institutional policies (e.g., state laws), cultural norms (e.g., attitudes toward sexual minorities), and societal-level conditions (e.g., density of same-sex couples), as described below. To statistically combine the indicators, we first performed three separate factor analyses on standardized structural-stigma indicators at the state, county, and city levels, respectively. Second, we quantified structural stigma at the city, county, and state levels using the standardized factor-scoring coefficients generated from the factor analyses to compute objectively measured structural-stigma factor scores. We chose a factor-analytic approach because it (a) recognizes that different dimensions of structural stigma (e.g., norms, policies) are highly correlated, (b) provides evidence of construct validity (i.e., shows that multiple items related to structural stigma load onto a single factor), and (c) reduces measurement error (i.e., tapping into shared variance, thereby minimizing unique variance). For a description of, data source for, and computation of each structural indicator, as well as the results of factor analyses conducted at each geographic level, see Appendices 3 to 10 in the SOM.

State-level structural stigma

The state-level structural-stigma factor score consisted of eight indicators: (a) presence of 33 protective and discriminatory state laws and policies related to sexual orientation, (b) explicit attitudes toward acceptance of homosexuality and legality of same-sex marriage, (c) explicit policy-specific attitudes toward rights for LGB people and same-sex couples, (d) implicit attitudes toward gay men and lesbian women, (e) proportion of openly LGBT elected government officials, (f) proportion of schools with gay–straight alliances (also known as gender–sexuality alliances), (g) density of LGBT adults, and (h) density of same-sex couples (Cronbach’s α = .95). The eight indicators loaded onto a single factor (factor loadings ranged from 0.65 to 0.92). Figure 1 shows the level of objectively measured structural stigma across U.S. states according to our indicators.

Fig. 1.

Fig. 1.

Open in a new tab

Distribution of objectively measured structural stigma related to sexual orientation across U.S. states. Darker colors indicate greater stigma.

County-level structural stigma

The county-level structural-stigma factor score consisted of four indicators: (a) explicit attitudes toward gay men, (b) explicit attitudes toward lesbian women, (c) implicit attitudes toward sexual minorities, and (d) density of same-sex couples (Cronbach’s α = .60). The four indicators loaded onto a single factor (factor loadings ranged from 0.20 to 0.74).

City-level structural stigma

The city-level structural-stigma factor score consisted of five indicators: (a) presence of nondiscrimination laws and policies related to sexual orientation, (b) presence of equality in benefits and protections provided to LGBT persons employed by the city and their partners, (c) availability of city services and programs for LGBT residents, (d) quality of relationship between law enforcement and the LGBT community, and (e) explicit commitment by city leadership to inclusive practices and equality (Cronbach’s α = .81). The five indicators loaded onto a single factor (factor loadings ranged from 0.62 to 0.80).

Covariates

We controlled for five individual-level sociodemographic covariates: relationship status, age, level of education, urbanicity, and racial/ethnic-minority status. These covariates were selected on the basis of past research (Pachankis & Bränström, 2018), bivariate correlations with perceived structural stigma and/or concealment motivation, and preliminary analyses showing that they did not moderate either the direct or indirect effects.

Analytic strategy

Prior studies have indicated that prejudice related to bisexuality involves different stereotypic content and sources from prejudice related to homosexuality (Bostwick & Dodge, 2019; Dodge et al., 2016; Worthen, 2013), that measures of structural stigma that are not specific to bisexuality do not predict psychosocial and health outcomes among bisexuals (e.g., Hatzenbuehler et al., 2018), and that identity-concealment concerns differ for gay and bisexual men (Pachankis et al., 2020). As a result, initial analyses were examined separately by sexual-orientation identity. These results indicated that only two associations among the bisexual sample reached statistical significance: Objectively coded structural stigma was associated with greater perceived structural stigma at the state and county levels. Consequently, the analyses below are presented only for the gay male subsample. The full set of analyses for the bisexual subsample can be found in Appendix 2 in the SOM and are summarized in the Discussion section.

Given the nested structure of the data, we used nested mixed-effects models to test the associations between objectively measured structural stigma, perceived structural stigma, and concealment motivation at the city, county, and state levels. We then estimated the indirect effects between objectively measured structural stigma and concealment motivation via perceived structural stigma at each geographic level. We quantified an indirect effect as the product of the effect of the independent variable on the mediator variable and the effect of the mediator on the dependent variable; 95% confidence intervals for the indirect effect were calculated via a normal approximation (MacKinnon et al., 2004).

Finally, we explored the interacting influences of objectively measured structural stigma measured at the state and city levels by using generalized additive mixed models (GAMMs; Wood & Scheipl, 2014). We focused on interactive effects of state- and city-level structural stigma on concealment motivation for conceptual and statistical reasons. Conceptually, the state and city levels represent the most meaningful differentiation of social environments with respect to the measurement of objective structural stigma because laws and policies are legislated at the state and city levels. Statistically, the Pearson correlation between objectively measured state- and city-level structural stigma was relatively modest (r = .18, p = .012) compared with the correlation between structural stigma at state and county levels (r = .57, p < .001); therefore, examining the interactive effects of state- and city-level structural stigma minimized concerns of multicollinearity in statistical modeling.

Unlike the traditional parametric linear regression models, semiparametric regression models such as GAMMs are often used to explore nonlinear associations between the independent and dependent variables (Wood, 2017). In this research, we used two different types of semiparametric regression techniques to assess the simultaneous influences of city- and state-level structural-stigma scores on concealment motivation: (a) to express associations between state-level structural stigma and concealment motivation as a smooth function of the city-level structural-stigma score (i.e., city-level structural stigma as the moderator), while first controlling for individual-level covariates, and (b) to express the estimated level of concealment motivation as a bivariate smooth function of the city- and state-level structural-stigma scores (Li et al., 2017). To visualize the interaction effects, we first graphed the estimated function curve for the state-level structural-stigma effect; then we graphed the estimated bivariate function as a colored contour plot.

The sample size was chosen on the basis of several factors, including GfK’s recommendation of the maximum sample size possible (given their substantial experience with recruitment) and power considerations (see Appendix 11 in the SOM). Distal environments were operationalized as state of residence, and proximal environments as county and city of residence. Analyses performed at the state level were based on the subsample of 371 gay men and 131 bisexual men (residing in 44 and 38 states, respectively), and analyses at the county level were based on 370 gay and 131 bisexual men (residing in 212 and 103 counties, respectively). The city-level nested mixed-effects models and the interaction-effects analyses (i.e., state- and city-level structural stigma on concealment motivation) were completed among a subsample of gay men (n = 206) and bisexual men (n = 51; residing in 106 and 39 cities, respectively). The smaller sample size for city-level analyses reflected the fact that gay men in the National Study of Stigma and Sexual Health sample resided in only 106 of the 489 cities for which we had data on objectively measured structural stigma. All analyses were conducted in SAS (Version 9.4) and RStudio (Version 1.2.5033-1; RStudio Team, 2019), and p values less than .05 were considered statistically significant.

Results

See Table 2 for results from unadjusted models and models adjusted for covariates among gay men.

Table 2.

Mixed-Effects Model Estimates: Associations Between Objectively Measured Structural Stigma at State, County, and City Levels; Concealment Motivation; and Perceived Structural Stigma Among Gay Men

Association and stigma level Model 1 Model 2
b 95% CI p b 95% CI p
Objectively measured structural stigma and concealment motivation
 State-level structural stigma 0.18 [0.04, 0.32] .014 0.17 [0.04, 0.31] .01
 County-level structural stigma 0.76 [0.50, 1.02] < .0001 0.73 [0.45, 1.00] < .0001
 City-level structural stigma 0.16 [−0.02, 0.33] .078 0.18 [0.01, 0.36] .046
Objectively measured structural stigma and perceived structural stigma
 State-level structural stigma 0.36 [0.26, 0.46] < .0001 0.34 [0.24, 0.44] < .0001
 County-level structural stigma 0.62 [0.43, 0.81] < .0001 0.52 [0.32, 0.73] < .0001
 City-level structural stigma 0.15 [0.03, 0.28] .019 0.14 [0.01, 0.27] .04
Perceived structural stigma and concealment motivation
 State-level perceived structural stigma 0.27 [0.12, 0.42] < .001 0.22 [0.07, 0.36] .003
 County-level perceived structural stigma 0.35 [0.21, 0.50] < .0001 0.28 [0.13, 0.43] < .001
 City-level perceived structural stigma 0.33 [0.15, 0.52] < .001 0.34 [0.15, 0.53] < .001
Open in a new tab

Note: Model 1 was unadjusted. Model 2 adjusted for relationship status (single, in a relationship), race/ethnicity (non-Hispanic White, other), age (continuous), education (less than high school, high school, some college, bachelor’s degree or higher), and urbanicity (metro, nonmetro). State of residence was available for 371 gay men. County of residence was available for 370 gay men. City-level data were collected through the Human Rights Campaign. The Municipality Equality Index was available for 106 cities, in which 206 of the 371 gay men resided. CI = confidence interval.

Direct association between objectively measured structural stigma and concealment motivation at state, county, and city levels

In support of our first hypothesis, results showed that objectively measured structural stigma was associated with greater concealment motivation among gay men, and this association was observed at the state, county, and city levels.

Association between objectively measured structural stigma and perceived structural stigma

As hypothesized, objectively measured structural stigma was associated with greater perceptions of structural stigma among gay men at state, county, and city levels.

Association between perceived structural stigma and concealment motivation

Consistent with our third hypothesis, results showed that perceived structural stigma was associated with concealment motivation among gay men at state, county, and city levels.

Multilevel mediation models

Our fourth hypothesis was that perceptions of structural stigma would mediate the association between objectively measured structural stigma and concealment motivation. We tested the indirect effect of perceived structural stigma on the association between objectively measured structural stigma and concealment motivation among gay men, controlling for individual-level covariates. The results of the mediation models are shown in Figure 2 (state level), Figure 3 (county level), and Figure 4 (city level). There was an indirect effect of objectively measured structural stigma at the state and county levels on concealment motivation through perceived structural stigma among gay men. Although there was an attenuation in the association between city-level structural stigma and concealment motivation when perceived structural stigma was included in the model, the indirect effect did not reach statistical significance at this level.

Fig. 2.

Fig. 2.

Open in a new tab

Multilevel mediation model showing the (a) total effect and (b) indirect effect of state-level structural stigma on concealment motivation among gay men through perceptions of structural stigma. The model adjusted for relationship status, race/ethnicity, age, education, and urbanicity. Values in brackets are 95% confidence intervals.

Fig. 3.

Fig. 3.

Open in a new tab

Multilevel mediation model showing the (a) total effect and (b) indirect effect of county-level structural stigma on concealment motivation among gay men through perceptions of structural stigma. The model adjusted for relationship status, race/ethnicity, age, education, and urbanicity. Values in brackets are 95% confidence intervals.

Fig. 4.

Fig. 4.

Open in a new tab

Multilevel mediation model showing the (a) total effect and (b) indirect effect of city-level structural stigma on concealment motivation among gay men through perceptions of structural stigma. The model adjusted for relationship status, race/ethnicity, age, education, and urbanicity. Values in brackets are 95% confidence intervals.

Interactions between state- and city-level structural stigma on concealment motivation

The GAMM results and accompanying estimated functional curve shown in Figure 5 indicate the presence of a state- by city-level interaction (p = .01) among gay men. State-level structural stigma had a positive association with concealment motivation when city-level structural stigma was low; in contrast, this association decreased and approached zero when city-level structural stigma increased.

Fig. 5.

Fig. 5.

Open in a new tab

Multilevel association between state-level structural stigma and concealment motivation (y-axis) as a function of city-level structural stigma (x-axis). The error band shows the pointwise 95% confidence interval.

To further unpack the pattern of this interaction, we generated a colored contour plot (Fig. 6), which shows the estimated means of concealment motivation at all combinations of objectively measured city- and state-level structural stigma. In the contour plot, colder colors (blues and greens) indicate lower concealment motivation, whereas warmer colors (yellows and oranges) indicate higher concealment motivation. There are two characteristics to focus on when interpreting the pattern of the interaction: (a) consistency versus gradation in color concentration and (b) the numbers on the contour lines, which indicate the estimated means of concealment motivation.

Fig. 6.

Fig. 6.

Open in a new tab

Interaction effects of state- and city-level structural stigma on concealment motivation among gay men. The numbers on the contour lines represent concealment values. Colder colors (blues, greens) indicate lower concealment motivation, whereas warmer colors (oranges, yellows) indicate higher concealment motivation. The white space indicates where there were no participants represented in our data set.

Both the gradation in color concentration and values on the contour lines in Figure 6 support the interpretation that city-level structural stigma has a stronger association with concealment motivation than does state-level structural stigma. For example, the lowest levels of concealment motivation are in the lower left quadrant (e.g., blue area), which is where city- and state-level structural stigma were both at their lowest. Further, as depicted by the concentration of green, when city-level structural stigma was at low levels (e.g., at −1.5), concealment motivation remained generally low (between 2.2 and 2.3), even at relatively high levels of state-level structural stigma. In contrast, as depicted by the concentration of yellow and orange, when city-level structural stigma was at average to relatively high levels (e.g., greater than 0), concealment motivation remained virtually unchanged (between 2.5 and 2.6) as state-level structural stigma increased. Thus, state-level structural stigma did not increase motivations to conceal one’s identity when structural stigma at the city level approached higher levels, indicating the lack of an additive effect.

Discussion

Although research has produced a wealth of understanding about the determinants and consequences of identity concealment, studies have focused primarily on interpersonal factors, such as situational features that make one’s stigma particularly salient, particularly likely to be discovered, or particularly consequential if discovered (Pachankis, 2007). Our study extends a growing body of evidence (e.g., Miller et al., 2011; Pachankis & Bränström, 2018) suggesting that concealment decisions among the stigmatized may also be driven, in part, by stigmatizing features of the broader social context beyond the immediate situation.

Whereas previous studies have measured structural stigma at only a single level (e.g., Pachankis & Bränström, 2018), we objectively measured structural stigma at three geographic levels simultaneously. Gay men reported significantly greater motivations to conceal their identities in states, counties, and cities with higher levels of structural stigma, after analyses controlled for individual-level covariates. Observing consistent associations at three diverse geographic levels helps to minimize concerns of confounding because variables that potentially confound the association between structural stigma and concealment motivation likely differ across these levels. Our measurement approach also enabled us to examine, for the first time, cross-level interactions between structural stigma at proximal (i.e., city) and distal (i.e., state) levels. As hypothesized, gay men reported the least motivations to conceal their identity if they resided in both cities and states that were lowest in structural stigma. Moreover, living in a city with relatively low structural stigma attenuated the motivation of gay men to conceal their sexual orientation, even if they lived in a state with high structural stigma. These findings underscore the utility of measuring structural stigma across multiple levels, which has the potential to reveal novel insights regarding the psychological consequences of structural stigma. Additionally, we advanced prior work on associations between structural stigma and identity concealment (e.g., Pachankis et al., 2015) by identifying a new psychological mechanism underlying this relationship. We found support for an appraisal pathway, whereby gay men report greater perceived structural stigma in environments that were objectively assessed as higher in structural stigma, and these perceptions in turn were associated with greater motivations to conceal their identity. Further, there was an indirect effect of objectively measured structural stigma on concealment motivation among gay men via perceived structural stigma, although we found weaker evidence for mediation at the city level.

In contrast to the results for gay men, we did not observe statistically significant associations between objectively measured structural stigma and concealment motivation among bisexual men at any of the geographic levels. However, the estimated effects at the county and city levels were similar in magnitude to the effect sizes among gay men (see SOM-R Table 2.2 in the SOM), suggesting that the lack of statistical significance may be due to reduced power. Further, objectively measured structural stigma was significantly associated with elevated perceptions of structural stigma at state and county levels among bisexual men. Thus, there is suggestive evidence that the first two mediation pathways operate similarly between gay and bisexual men. The primary group difference emerged in the third pathway: Perceived structural stigma was not associated with greater concealment motivation among bisexual men at any geographic level. This finding highlights the need to explore alternative mechanisms linking structural stigma to concealment motivation in this group and to identify factors (e.g., lower identity centrality, opposite-sex-partnership status) that reduce the likelihood that perceptions of structural stigma increase bisexual men’s concealment motivations.

We note several limitations of this study. Our data are cross-sectional. Because it is unlikely that greater concealment would cause more structural stigma, reverse causality is not a plausible alternative explanation for these findings. However, because mediation processes, by definition, unfold over time (MacKinnon, 2012), our mediation hypotheses should be examined prospectively in future research. At the city and county levels, we were unable to incorporate the same comprehensive breadth of structural-stigma indicators available at the state level, both because some indicators are not operable at certain levels (e.g., laws that affect sexual minorities are not legislated at the county level) and because some indicators are not routinely collected at certain levels (e.g., city-level data on attitudes toward sexual minorities). Thus, although our measures of structural stigma include indicators that are appropriate to their respective levels, we could not directly compare the magnitude of estimated effects across levels.

We also note several factors that may affect the generalizability of our results. First, participants resided in counties and cities with lower average levels of structural stigma compared with all U.S. counties and cities from which the objective structural-stigma factor scores were computed. Because this restricted range would reduce statistical power, our estimates are likely conservative, but the restricted range limits generalizability to the contexts included in our sample. Second, the proximal environment was operationalized as residential city/county, which may not capture all relevant spatial contexts in which sexual-minority men socialize, such as neighborhoods (Duncan et al., 2014). Thus, future research should determine whether our results are generalizable across different measures of proximal environments (e.g., residential and social). Third, the relatively small sample of racial/ethnic minorities (n = 97) likely reduced statistical power to detect moderation and prevented us from examining the association between structural stigma and concealment motivation within racial/ethnic-minority groups. Motivations to conceal one’s identity, level of concealment, and associated outcomes differ by race/ethnicity (e.g., Moradi et al., 2010; Pachankis et al., 2020; Villicana et al., 2016), perhaps because of different attitudes toward sexuality held by racial/ethnic groups (Cerezo et al., 2020; Greene, 1994). Thus, future research would benefit from incorporating an intersectional approach to the study of structural stigma and identity concealment.

Conclusion

Using data from a new population-based survey of sexual-minority men, we addressed heretofore unanswered questions about structural stigma and identity concealment, including documenting that, among gay men, this association is (a) apparent at both proximal and distal levels (main effect), (b) conditional on one’s exposure to structural stigma at another level (interaction effect), and (c) mediated by subjective appraisals (indirect effect). Determining whether these findings generalize to members of other groups who also conceal aspects of their stigmatized identities is an important area for future inquiry. Because individuals may have multiple stigmatized identities that they conceal at any given moment, this research would benefit in particular from intersectional approaches that capture how structural forms of stigma across multiple axes of social stratification shape concealment behaviors among stigmatized individuals.

Footnotes

ORCID iD: Mark L. Hatzenbuehler Inline graphic https://orcid.org/0000-0002-7430-0853

Transparency

Action Editor: Steven W. Gangestad

Editor: Patricia J. Bauer

Author Contributions

M. R. Lattanner and M. L. Hatzenbuehler developed the study concept. M. L. Hatzenbuehler, B. Dodge, J. Ford, J. E. Pachankis, and M. R. Lattanner designed the study. N. Bo, W. Tu, and M. R. Lattanner analyzed the data. M. L. Hatzenbuehler and M. R. Lattanner drafted the manuscript. All the authors provided revisions and approved the final manuscript for submission.

Declaration of Conflicting Interests: The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

Funding: This work was supported by National Institute of Mental Health Grants T32MH013043 and R01MH112384.

Open Practices: All materials for this study have been made publicly available in the supplementary online material (SOM), available on OSF at https://osf.io/ytcsu/. Data used to create the objectively measured structural-stigma factor scores were obtained from publicly available sources (see Appendices 3, 6, and 9 in the SOM). Individual-level data were collected through the National Study of Stigma and Sexual Health (see Appendix 11 in the SOM) and are available on request from M. L. Hatzenbuehler at markhatzenbuehler@fas.harvard.edu. For participant location, only state- and county-level data can be provided because of data-use agreements with Ipsos. The analytic code for all analyses is provided in Appendix 11. The design and analysis plan for the study were not preregistered. This article has received the badge for Open Materials. More information about the Open Practices badges can be found at http://www.psychologicalscience.org/publications/badges.

graphic file with name 10.1177_09567976211018624-img2.jpg

References

  1. Bostwick W. B., Dodge B. (2019). Introduction to the special section on bisexual health: Can you see us now? Archives of Sexual Behavior, 48, 79–87. 10.1007/s10508-018-1370-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bruni F. (2017, August 25). The worst (and best) places to be gay in America. The New York Times. https://www.nytimes.com/interactive/2017/08/25/opinion/sunday/worst-and-best-places-to-be-gay.html
  3. Cerezo A., Cummings M., Holmes M., Williams C. (2020). Identity as resistance: Identity formation at the intersection of race, gender identity, and sexual orientation. Psychology of Women Quarterly, 44(1), 67–83. 10.1177/0361684319875977 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Dodge B., Hebenick D., Friedman M. R., Schick V., Fu T. C., Bostwik W., Bartelt E., Munoz-Laboy M., Pletta D., Reece M., Sandfort T. G. (2016). Attitudes toward bisexual men and women among a nationally representative probability sample of adults in the United States. PLOS ONE, 11, Article e0164430. 10.1371/journal.pone.0164430 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Duncan D. T., Kapadia F., Halkitis P. N. (2014). Examination of spatial polygamy among young gay, bisexual, and other men who have sex with men in New York City: The P18 cohort study. International Journal of Environmental Research and Public Health, 11, 8962–8983. 10.3390/ijerph110908962 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Greene B. (1994). Ethnic-minority lesbians and gay men: Mental health and treatment issues. Journal of Consulting and Clinical Psychology, 62(2), 243–251. [DOI] [PubMed] [Google Scholar]
  7. Hatzenbuehler M. L. (2014). Structural stigma and the health of lesbian, gay, and bisexual populations. Current Directions in Psychological Science, 23(2), 127–132. 10.1177/0963721414523775 [DOI] [Google Scholar]
  8. Hatzenbuehler M. L. (2016). Structural stigma: Research evidence and implications for psychological science. American Psychologist, 71, 742–751. 10.1037/amp0000068 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Hatzenbuehler M. L. (2017). Structural stigma and health. In Major B., Dovidio J. F., Link B. G. (Eds.), The Oxford handbook of stigma, discrimination, and health (pp. 105–121). Oxford University Press. [Google Scholar]
  10. Hatzenbuehler M. L., Bränström R., Pachankis J. E. (2018). Societal-level explanations for reductions in sexual orientation mental health disparities: Results from a ten-year, population-based study in Sweden. Stigma and Health, 3, 16–26. 10.1037/sah0000066 [DOI] [Google Scholar]
  11. Hatzenbuehler M. L., Link B. G. (2014). Introduction to the special issue on structural stigma and health. Social Science & Medicine, 103, 1–6. 10.1016/j.socscimed.2013.12.017 [DOI] [PubMed] [Google Scholar]
  12. Hatzenbuehler M. L., McLaughlin K. A. (2014). Structural stigma and hypothalamic–pituitary–adrenocortical axis reactivity in lesbian, gay, and bisexual young adults. Annals of Behavioral Medicine, 47, 39–47. 10.1007/s12160-013-9556-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kuyper L., Fernee H., Keuzenkamp S. (2016). A comparative analysis of a community and general sample of lesbian, gay, and bisexual individuals. Archives of Sexual Behavior, 45, 683–693. 10.1007/s10508-014-0457-1 [DOI] [PubMed] [Google Scholar]
  14. Li Z., Liu H., Tu W. (2017). A generalized semiparametric mixed model for analysis of multivariate health care utilization data. Statistical Methods in Medical Research, 26, 2909–2918. 10.1177/0962280215615159 [DOI] [PubMed] [Google Scholar]
  15. MacKinnon D. P. (2012). Introduction to statistical mediation analysis. Routledge. [Google Scholar]
  16. MacKinnon D. P., Lockwood C. M., Williams J. (2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 99–128. 10.1207/s15327906mbr3901_4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Meyer I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129, 674–697. 10.1037/0033-2909.129.5.674 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Meyer I. H., Wilson P. A. (2009). Sampling lesbian, gay, and bisexual populations. Journal of Counseling Psychology, 56, 23–31. 10.1037/a0014587 [DOI] [Google Scholar]
  19. Miller C. T., Grover K. W., Bunn J. Y., Solomon S. E. (2011). Community norms about suppression of AIDS-related prejudice and perceptions of stigma by people with HIV or AIDS. Psychological Science, 22(5), 579–583. 10.1177/0956797611404898 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mohr J. J., Kendra M. S. (2011). Revision and extension of a multidimensional measure of sexual minority identity: The Lesbian, Gay, and Bisexual Identity Scale. Journal of Counseling Psychology, 58, 234–245. 10.1037/a0022858 [DOI] [PubMed] [Google Scholar]
  21. Moradi B., Wiseman M. C., DeBlaere C., Goodman M. B., Sarkees A., Brewster M. E., Huang Y. P. (2010). LGB of color and White individuals’ perceptions of heterosexist stigma, internalized homophobia, and outness: Comparisons of levels and links. The Counseling Psychologist, 38(3), 397–424. 10.1177/0011000009335263 [DOI] [Google Scholar]
  22. Newheiser A. K., Barreto M. (2014). Hidden costs of hiding stigma: Ironic interpersonal consequences of concealing a stigmatized identity in social interactions. Journal of Experimental Social Psychology, 52, 58–70. 10.1016/j.jesp.2014.01.002 [DOI] [Google Scholar]
  23. Pachankis J. E. (2007). The psychological implications of concealing a stigma: A cognitive-affective-behavioral model. Psychological Bulletin, 133, 328–345. 10.1037/0033-2909.133.2.328 [DOI] [PubMed] [Google Scholar]
  24. Pachankis J. E., Bränström R. (2018). Hidden from happiness: Structural stigma, sexual orientation concealment, and life satisfaction across 28 countries. Journal of Consulting and Clinical Psychology, 86, 403–415. 10.1037/ccp0000299 [DOI] [PubMed] [Google Scholar]
  25. Pachankis J. E., Hatzenbuehler M. L., Hickson F., Weatherburn P., Berg R. C., Marcus U., Schmidt A. J. (2015). Hidden from health: Structural stigma, sexual orientation concealment, and HIV across 38 countries in the European MSM Internet Survey (EMIS). AIDS, 29, 1239–1246. 10.1097/QAD.0000000000000724 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Pachankis J. E., Mahon C. P., Jackson S. D., Fetzner B. K., Bränström R. (2020). Sexual orientation concealment and mental health: A conceptual and meta-analytic review. Psychological Bulletin, 146(10), 831–871. 10.1037/bul0000271 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Quinn D. M., Chaudoir S. R. (2009). Living with a concealable stigmatized identity: The impact of anticipated stigma, centrality, salience, and cultural stigma on psychological distress and health. Journal of Personality and Social Psychology, 97, 634–651. 10.1037/a0015815 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Riggle E. D., Rostosky S. S., Horne S. (2010). Does it matter where you live? Nondiscrimination laws and the experiences of LGB residents. Sexuality Research and Social Policy, 7, 168–175. 10.1007/s13178-010-0016-z [DOI] [Google Scholar]
  29. RStudio Team. (2019). RStudio: Integrated development for R (Version 1.2.5033-1) [Computer software]. http://www.rstudio.com/
  30. Villicana A. J., Delucio K., Biernat M. (2016). “Coming out” among gay Latino and gay White men: Implications of verbal disclosure for well-being. Self and Identity, 15(4), 468–487. 10.1080/15298868.2016.1156568 [DOI] [Google Scholar]
  31. Wood S. (2017). Generalized additive models: An introduction with R. CRC Press. [Google Scholar]
  32. Wood S., Scheipl F. (2014). gamm4: Generalized additive mixed models using lme4 and mgcv (R package Version 0.2-3). https://rdrr.io/cran/gamm4/man/gamm4.html
  33. Worthen M. G. F. (2013). An argument for separate analyses of attitudes toward lesbian, gay, bisexual men, bisexual women, MtF and FtM transgender individuals. Sex Roles, 68, 703–723. 10.1007/s11199-012-0155-1 [DOI] [Google Scholar]

Articles from Psychological Science are provided here courtesy of SAGE Publications

RESOURCES