Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Sep 26.
Published before final editing as: Psychol Sex. 2026 Sep 16:10.1080/19419899.2026.2734898. doi: 10.1080/19419899.2026.2734898

Confirmatory factor analysis and measurement invariance of the sexual orientation concealment scale among nonbinary LGBTQ+ adults

Peter D Goldie a, Michael T Bixter a, Suzanne McLaren b
PMCID: PMC13614439  NIHMSID: NIHMS2211249  PMID: 42799242

Abstract

Sexual orientation concealment is a prevalent minority stressor among LGBTQ+ people (including nonbinary people) and it is associated with a series of adverse mental health outcomes. The Sexual Orientation Concealment Scale (SOCS) is a widely used brief measure assessing sexual orientation concealment. Although the SOCS has been used with diverse cisgender populations, its validity for nonbinary people, including those of varied racial/ethnic backgrounds and sexual orientations, remains unknown. Based on data from an international sample of 340 nonbinary LGBTQ+ individuals, the present study reports the findings of a confirmatory factor analysis to assess the one-factor structure and validity of the SOCS, as well as measurement invariance by racial/ethnic identity (i.e. white and Black, Indigenous, and People of Colour [BIPOC]) and sexual orientation (i.e. monosexual, bi+, and queer). The model fit the data well for the one-factor solution, and we observed partial scalar invariance by racial/ethnic identity and sexual orientation. These findings indicate that the SOCS is valid for use with nonbinary people and provide preliminary support for its use across racial/ethnic and sexual orientation groups. We outline implications for future investigation to increase the cohesion of this body of work, and for clinical practice with nonbinary people.

Keywords: Nonbinary, LGBTQ+, sexual orientation concealment, invariance, confirmatory factor analysis

Lesbian, gay, bisexual, transgender (trans), queer, and other (LGBTQ+) adults experience disproportionately high levels of mental health challenges and conditions (e.g. suicidality, depression, substance dependence; Blondeel et al., 2016; Reisner et al., 2016; Wittgens et al., 2022; Zeeman et al., 2019). Minority stress theory, which proposes that sexual and gender minority people encounter additional stressors as they navigate heterosexist environments and interactions (Brooks, 1981; Meyer, 1995, 2003), helps to explain these mental health disparities. Although minority stress theory was originally developed to conceptualise the experiences of cisgender (cis) sexual minority people, scholars (e.g. Hendricks & Testa, 2012; Matsuno et al., 2026) have used and/or modified this framework to examine the experiences of trans people (i.e. individuals who identify with a gender that is different from their assigned sex) and nonbinary people (i.e. individuals who do not exclusively identify as men or women and often also identify as trans; Matsuno & Budge, 2017). An updated, more gender-inclusive minority stress framework proposes that four factors characterise minority stress: prejudice events, expectations of rejection and discrimination, identity concealment, and internalised heterosexism/homophobia/transphobia (Hendricks & Testa, 2012; Meyer, 1995, 2003).

Gender groups within the LGBTQ+ community have distinct experiences. Much is known about the experiences of LGBTQ+ people with some genders (e.g. cis men), yet far fewer studies have investigated how nonbinary LGBTQ+ people’s experiences may diverge. Existing studies have, however, revealed meaningful differences in the experiences of binary and nonbinary trans people (e.g. discrepant levels of comfort with gender labelling processes; Fiani & Han, 2019), and nonbinary and cis LGBTQ+ people (e.g. nonbinary people report lower levels of perceived gender determinism; Bradford & Catalpa, 2019). Moreover, Matsuno et al. (2026) found that, although nonbinary and binary trans adults encounter similar stressors (e.g. discrimination, misgendering, internalised stigma, gender identity concealment), nonbinary people experience them in unique ways. Matsuno et al. (2026) also noted that there are stressors unique to nonbinary people. For example, nonbinary people report frequent experiences of binary gender normativity (e.g. having no choice but to use gendered bathrooms), psychological burden resulting from educating others about nonbinary gender identities, and invalidation of their gender identities. It is integral to continue leveraging rigorous scientific methods to understand nonbinary people’s experiences of minority stress. Doing so is particularly important given that the nonbinary population is rapidly growing, especially among younger people (A. Brown, 2022; Ipsos, 2021). Indeed, a recent study (Hammack et al., 2022) showed that approximately one quarter of LGBTQ+ adolescents are nonbinary.

Sexual orientation concealment and mental health outcomes

Sexual orientation concealment (SOC) refers to a minority stress process by which individuals withhold or provide inaccurate information regarding their sexual orientation or identity (Schrimshaw et al., 2013). A growing body of research suggests that the prevalence and consequences of SOC vary by racial identity, sexual orientation, gender identity, and geographical region (e.g. Glynn et al., 2024; Hitch & Brown, 2024; Pachankis & Bränström, 2018; Pachankis, Mahon, et al., 2020), highlighting the need to examine this construct across identity groups and sociocultural contexts. For instance, a recent study with predominantly white participants showed that cis women were more likely to conceal their identities relative to trans women, and bisexual people were more likely to conceal them than pansexual people (Feinstein et al., 2020). Moreover, a scoping review by Hitch and Brown (2024) reported differences in associations between SOC and mental health difficulties across racial/ethnic groups; for instance, greater concealment was associated with increased depression, anxiety, and poorer psychological well-being among Asian men in one study, but there were no significant associations between these variables among Latino sexual minority men in other studies.

Scholars have aptly identified complex ways that cultural and structural factors shape identity concealment and disclosure processes. Szymanski and Sung (2013), for example, highlighted that greater adherence to Asian cultural values is associated with lower levels of sexual orientation disclosure. Pachankis and Bränström (2018) found that greater country-level structural stigma (e.g. laws and policies restricting LGBTQ+ rights) is associated with higher levels of sexual orientation concealment. Although SOC has been increasingly examined in scholarship focused on correlates of cis LGBTQ+ people’s mental health (with a welcome focus on variability across cultures, genders, and other identity factors), we are not aware of any studies that have examined SOC and mental health outcomes among nonbinary people.

Being open about one’s LGBTQ+ sexual orientation and/or having a visible LGBTQ+ identity can open individuals up to discrimination, profiling, incarceration, expulsion from family and friends, violence, and/or death (DeChants et al., 2022; Flores et al., 2022; Meyer et al., 2017; Pachankis & Bränström, 2018). In other words, LGBTQ+ people often conceal their sexual orientation for a variety of reasons that stem from very real and reasonable concerns about their safety and well-being. SOC may be adaptive and necessary in certain situations (see Pachankis, 2007; Pachankis & Bränström, 2018) and the reasons for concealing and the prevalence of concealment vary by context.

SOC, however, seems to have concerning psychological costs. A breadth of data indicates that, consistent with minority stress theory, SOC is associated with adverse mental health outcomes. A recent meta-analysis of nearly 200 studies reported a small, positive association between SOC and internalising problems (e.g. depressive symptoms, anxiety symptoms), general distress, and problematic eating behaviours (Pachankis, Mahon, et al., 2020). Although this meta-analysis was robust, some studies have highlighted that the associations between concealment and mental health outcomes are quite complex (Huang & Chan, 2022) or non-significant (Kase & Mohr, 2023). Thus, it is important for researchers to continue expanding this body of knowledge, including how, when, and for whom SOC is especially corrosive.

Limitations of the existing sexual orientation concealment literature

Despite the value of the existing empirical literature, there are critical drawbacks. One limitation is the widespread conceptual confusion surrounding this construct, which has led to inconsistent operationalisation across studies (Hitch & Brown, 2024; Jackson & Mohr, 2016; Pachankis, Mahon, et al., 2020). For instance, many researchers have conceptualised SOC (like the concealment of other stigmatised identities) as the opposite of disclosure (e.g. Le Forestier et al., 2022). Yet scholars such as Jackson and Mohr (2016) have contended that concealment entails active attempts to prevent disclosure and, thus, is a distinct variable. Another notable limitation of this research is that it has almost exclusively focused on cis populations. In many studies that have included trans and nonbinary participants, they have represented extremely small proportions, precluding a robust examination of their experiences. For instance, Huang and Chan (2022) recruited a sample comprising 1.3% trans or nonbinary participants. To our knowledge, only one study (Feinstein et al., 2020) has specifically examined variables related to nonbinary people’s SOC. Using a single-item measure of concealment, Feinstein et al. (2020) found that nonbinary people were not more likely to conceal their sexual orientation than cis bi+ women; however, they had significantly higher levels of interpersonal motivations for concealment (e.g. concerns about being judged, treated negatively, or rejected) than cis bi+ women.

Measures of sexual orientation concealment

Given that SOC is associated with critical mental health outcomes, it is vital to have psychometrically sound tools to assess it. Unfortunately, the pervasive confusion about how to conceptualise this construct is paralleled by confusion about how to measure it. Scholars who have (mis)understood SOC as simply the lack of disclosure, for instance, have assessed it using measures of disclosure (e.g. equating higher levels of concealment with explicitly disclosing one’s sexual minority identity to fewer people; Ragins et al., 2007). The inconsistent operationalisation of SOC across studies may obscure meaningful associations between concealment and mental health outcomes (Pachankis, Mahon, et al., 2020), and is methodologically imprecise.

One possible reason that such minimal scholarship has investigated SOC and its correlates among nonbinary people is that no previous studies have validated SOC scales with this population. This caution seems to be warranted, as recent research on LGBTQ+ people (e.g. Sullivan et al., 2025) has shown that commonly used measures of minority stressors (e.g. outness, discrimination) operate differently across gender groups (e.g. cis men, transmasculine people, gender expansive people). Broadening this body of work to validate measures of SOC with nonbinary people would highlight avenues for understanding their concealment processes with more nuance.

Currently, the most commonly used measures of SOC appear to be the modified Self-Concealment Scale, developed by Schrimshaw et al. (2013), and the Sexual Orientation Concealment Scale (SOCS), developed by Jackson and Mohr (2016). The first (Schrimshaw et al., 2013) is a modified general concealment scale that focuses specifically on SOC. Schrimshaw et al. (2013) reported that the seven items on this scale loaded onto a single factor (with loadings ranging from .57 to .79) and demonstrated strong internal consistency (α = .81) among a sample of non-gay men who have sex with men. To our knowledge, no studies have attempted to validate this scale for use with nonbinary people.

Jackson and Mohr (2016) developed and validated the SOCS several years later. The SOCS is a six-item Likert-type scale that assesses lesbian, gay, and bisexual (LGB) individuals’ efforts to hide their sexual orientation from others over the previous two weeks. In their validation study, Jackson and Mohr (2016) conducted exploratory and confirmatory factor analyses (CFAs). The exploratory factor analysis demonstrated that a one-factor solution fit the data well; therefore, it was retained in the CFA. Both factor analyses showed a good fit to the data, and convergent and discriminant validity were supported. Since the publication of Jackson and Mohr (2016), the utility of this brief scale has been realised in the field more generally; it has been widely used in empirical research focused on predominantly cis LGBTQ+ people, much of which has made important conceptual contributions (e.g. T. A. Brown et al., 2024; Csoti & McLaren, 2025; Dunn et al., 2025). Notably, nearly all participants (97%) in the validation study were cis, and thus, it remains unknown whether the SOCS is an appropriate measure for use with nonbinary people.

An additional psychometric property that would further bolster confidence in the use of the SOCS with nonbinary people is measurement invariance. Measurement invariance means that score comparisons between groups indicate true score differences rather than measurement differences (Meuleman et al., 2023). There is utility in examining measurement invariance of the SOCS across many groups, as doing so would help researchers ensure they are drawing accurate conclusions. Moreover, given research findings that LGBTQ+ people’s experiences of SOC may vary by racial/ethnic identity (Pachankis, Mahon, et al., 2020; see also Hitch & Brown, 2024) and sexual orientation (Glynn et al., 2023), and the preliminary evidence for noninvariance across sexual orientations on measures of other minority stressors (Sullivan et al., 2025), studying invariance across these identity groups is critical. We are not aware of previous research that has studied the invariance of the SOCS. Demonstrating measurement invariance, in this case, would mean that the SOCS assesses the same latent construct and operates similarly across nonbinary people with different racial/ethnic identities and sexual orientations.

The present study

A robust body of research has documented associations between increased SOC and adverse mental health outcomes among predominantly cis LGBTQ+ people, thus reinforcing the rationale for including SOC as a main minority stressor (Meyer, 1995, 2003). It is important to expand research in this area to also capture the experiences of nonbinary people, who are uniquely marginalised and chronically underrepresented in LGBTQ+ health research (Matsuno & Budge, 2017). However, doing so will require a measure of SOC that is validated for use with this population, as research suggests their experiences of minority stress differ meaningfully from cis LGBTQ+ people (Sullivan et al., 2025). The SOCS is a brief measure of SOC that has shown promising psychometric properties among cis LGBTQ+ people but has not been validated for use with nonbinary people. To this end, the present study aimed to examine the extent to which the factor structure and internal consistency of the SOCS as reported by Jackson and Mohr (2016) for predominantly (97%) cis people are consistent for nonbinary people. Furthermore, to enhance our understanding of the validity of the SOCS among subgroups of nonbinary people, we assessed measurement invariance by racial/ethnic identity (i.e. white versus Black, Indigenous, and People of Colour [BIPOC]) and sexual orientation (i.e. monosexual, bi+, and queer).

Method

Procedure

The study’s research protocol was approved by the Charles Sturt University Human Research Ethics Committee (reference number H24359). Study recruitment took place online electronically via snowball sampling methods, such as through social media platforms and referrals, and forwarding by LGBTQ+ groups. The online survey was hosted by Qualtrics© and participants accessed the survey via a web link or QR code. Before beginning the survey, participants viewed an Information Statement. This statement: invited people who had a minority sexual orientation and were aged 18 years or older to participate in the study; indicated that participation was voluntary and could be discontinued at any time; provided basic information about the survey; and displayed information for support services if distress arose during the study. Participants were then required to click ‘I consent to participate’ to continue. Those who did not meet the study requirements (e.g. age, sexual orientation) were exited from the study at this point. Participants were not required to answer any items to proceed, leaving them free to respond as desired. All who completed the survey in full were shown a debriefing statement.

Participants

Participants in this study are a subset of a larger sample of LGBTQ+ people, as only participants who identified as nonbinary were included in the present analyses. An international sample of 340 nonbinary people completed the survey in full, and there were no missing data. Participants’ ages ranged from 18 to 62 (M = 28.15, SD = 8.75). They lived in 29 countries, with the largest representation from the US (n = 177, 52.1%), the UK (n = 40, 11.8%), Australia (n = 38, 11.2%), Canada (n = 19, 5.6%), and Germany (n = 13, 3.8%). The vast majority of participants were white, plurisexual (e.g. bisexual, pansexual) or queer, were employed part- or full-time, and earned a below-average income. Half of the sample was partnered, had a tertiary education, and out to most people they knew. Full sample demographic characteristics are shown in Table 1.

Table 1.

Sample demographic information.

Characteristic n % of sample
Sexual Orientation
 Lesbian 58 17.1
 Gay 11 3.2
 Bisexual 106 31.2
 Queer 93 27.4
 Pansexual 53 15.6
 Omnisexual 19 5.6
Racial/Ethnic Identity
 white 267 78.5
 BIPOC 73 21.5
  First Nations 8 2.4
  Black/African/African American 11 3.2
  Asian 17 5.0
  Latinx 14 4.1
  Multiracial 23 6.8
Education
 Primary/Secondary/Vocational 176 51.8
 Tertiary 164 48.2
Employment
 Working full-time 123 36.2
 Working part-time 142 41.8
 Not Working 75 22.1
Income Status
 Low 224 65.9
 Middle 103 30.3
 High 13 3.8
Relationship Status
 Partnered 171 50.3
 Unpartnered 169 49.7
Sexual Orientation Disclosure
 0–25% 82 24.1
 26–50% 64 18.8
 51–75% 87 25.6
 76–100% 107 31.5
Open in a new tab

Note: BIPOC = Black, Indigenous, and People of Colour.

Measures

Demographic characteristics

Participants provided basic sociodemographic information to describe the sample. These variables included age, gender, sexual orientation, relationship status, education level, income status (answer options were low, middle, or high income earner), employment status, country of residence, and racial/ethnic identity. Lastly, participants were asked to report their level of sexual orientation disclosure (i.e. the percentage of people one knows that are aware of their sexual orientation [0%−25%, 26%−50%, 51%−75%, or 76%−100%]).

Sexual orientation concealment

The SOCS (Jackson & Mohr, 2016) is a brief, 6-item self-report measure assessing the extent to which LGB individuals hide their sexual orientation from other people, based on their experiences over the past two weeks. A sample item from the measure is: ‘In the last 2 weeks, I have concealed my sexual orientation by avoiding contact with other LGB individuals’. Participants were asked to indicate their endorsements for each item on a Likert-type scale with answer options ranging from 1 (Not at all) to 5 (All the time). Higher scores signify higher levels of SOC. As reported in its initial validation study, this scale has a unidimensional factor structure and demonstrated convergent and discriminant validity. Moreover, the SOCS has shown good internal consistency across studies (α = .77 or higher; e.g. Csoti & McLaren, 2025; Huang & Chan, 2022; Jackson & Mohr, 2016; Ramos et al., 2021).

Data analysis

Groupings

We combined participants into groups for our invariance analyses due to insufficient cell sizes across some sexual orientations and racial/ethnic groups (see Table 1). Regarding sexual orientation groupings, queer was retained as its own group and we created two new groups: monosexual (lesbian, gay) and bi+ (bisexual, pansexual, and omnisexual). Regarding race/ethnicity, we had a relatively small sample of BIPOC participants and thus, created a single BIPOC group composed of First Nations, Black/African/African American, Asian, Latinx, and Multiracial participants. We recognise that grouping BIPOC participants together helped us to answer our broad measurement invariance research question, but may obfuscate within-group diversity. We do not intend to equate BIPOC experiences (Nielsen, 2000).

Analytic plan

The hypotheses were tested in R (R Core Team, 2021) and Statistical Package for the Social Sciences (SPSS) Version 28. Given that the SOCS has previously been validated with other populations (i.e. cis LGBTQ+ people; Jackson & Mohr, 2016), we assessed its validity for nonbinary people using a CFA. Below, we report various model fit indices, including chi-square, comparative fit index (CFI), root mean square error of approximation (RMSEA), Tucker-Lewis index (TLI), and standardised root mean square residual (SRMR). The following standard cut-offs guided decisions about model fit: CFI > .90; RMSEA < .08; SRMR < .08; and TLI > .90 (Browne et al., 1993; Hu & Bentler, 1999). An acceptable average variance extracted was considered to be .50 or higher, and good construct reliability was .80 or higher. In line with Hair et al. (2009), standardised item loadings of .50 or higher were considered adequate. We used multiple pieces of evidence, such as model fit statistics and standardised factor loadings, to determine the adequacy of the models. All item loadings described herein are standardised.

To assess for measurement variance, we planned to use an increasingly restrictive set of steps beginning with configural invariance and, when applicable, proceeding to metric (weak) invariance and scalar (strong) invariance by racial/ethnic identity (white, BIPOC) and sexual orientation (monosexual, bi+, and queer). As with our base model analysis, we used multiple sources of evidence to make determinations about invariance. We drew from conventional cut-off recommendations (e.g. Chen, 2007) for loading invariance analyses with unequal sample sizes; specifically, we determined that a change of ≥.01 in CFI paired with a change of ≥.01 in RMSEA or a change of ≥.025 in SRMR indicated noninvariance. We considered a significant Δχ2 as further evidence of noninvariance.

Results

Preliminary and descriptive analyses

Descriptive analyses were performed, and the means and standard deviations for each SOCS item are displayed in Table 2.

Table 2.

SOCS Item means, standard deviations, and standardised factor loadings.

Item Mean Standard deviation Standardised factor loading
1: In the past two weeks, I have concealed my sexual orientation by telling someone that I was straight or denying that I was LGB 1.629 1.161 .822
2: In the past two weeks, I have concealed my sexual orientation by avoiding contact with other LGB individuals 1.174 0.653 .527
3: In the past two weeks, I have avoided the subjects of sex, love, attraction, or relationships to conceal my sexual orientation 1.874 1.162 .762
4: In the past two weeks, I have allowed others to assume I am straight without correcting them 2.568 1.453 .702
5: In the past two weeks, I have altered my appearance, mannerisms, or activities in an attempt to ‘pass’ as straight 1.782 1.165 .660
6: In the past two weeks, I have remained silent while witnessing anti-gay remarks, jokes, or activities because I did not want to be labelled as LGB by those involved 1.553 1.086 .669
Open in a new tab

Base model confirmatory factor analysis

The model demonstrated mixed, but acceptable fit for the one-factor solution, χ2(9) = 36.571, p = .001, CFI = .964, TLI = .939, RMSEA = .095, SRMR = .041. The fit statistics met their respective cut-offs except RMSEA, which slightly exceeded its cut-off. The average variance extracted was acceptable (.511), internal consistency was high (Cronbach’s α = .839), and item loadings were uniformly acceptable (see Table 2).

Measurement invariance by racial/ethnic identity

The fit indices, ΔCFI, ΔRMSEA, ΔSRMR, and Δχ2 for the model testing invariance by racial/ethnic identity are shown in Table 3. Results indicated that the SOCS configural factor structure was the same for white people and BIPOC; all fit statistics met the specified cut-offs, except for RMSEA (.083) which slightly exceeded its cut-off. The CFI, TLI, and SRMR supported a strong fit and thus the model demonstrated configural invariance.

Table 3.

SOCS measurement invariance by racial/ethnic identity and sexual orientation.

Type of Invariance χ2 (df) p-value RMSEA (90% CI) TLI CFI SRMR ΔRMSEA ΔCFI ΔSRMR Δχ2(p-value)
Racial/Ethnic identity (white versus BIPOC)
 Configural 38.932
(18)		 .003 .083
(.047, .118)		 .953 .972 .036
 Metric 58.806
(23)		 <.001 .096
(.066, .126)		 .937 .951 .067 .013 .021 .031 19.874
(.001)
 aPartial metric 44.974
(22)		 .003 .078
(.045, .111)		 .957 .969 .048 .005 .003 .012 6.041
(.196)
 aPartial scalar 49.987
(27)		 .005 .071
(.039, .101)		 .965 .969 .052 .007 .000 .004 5.014
(.414)
Sexual identity (monosexual versus bi+ versus queer)
 Configural 72.393
(27)		 <.001 .122
(.088, .156)		 .903 .942 .044
 Metric 99.069
(37)		 <.001 .122
(.093, .151)		 .903 .920 .088 .00 .022 .044 26.676
(.003)
 aPartial metric 84.454
(35)		 <.001 .112
(.081, .142)		 .918 .936 .065 .001 .023 .026 12.060
(.149)
 aPartial scalar 101.380
(45)		 <.001 .105
(.078, .133)		 .928 .928 .072 .007 .008 .007 16.927
(.076)
Open in a new tab

Note: BIPOC = People of Colour; SOCS = Sexual Orientation Concealment Scale; RMSEA = root mean square error of approximation; ΔRMSEA = root mean square error of approximation difference; CFI = comparative fit index; ΔCFI = comparative fit index difference; χ2 = Chi-squared; Δχ2 = Chi-squared difference; TLI = Tucker-Lewis index; SRMR = standardised root mean square residual; ΔSRMR = standardised root mean square residual difference;

a

model was performed with the loading for Item 2 unconstrained.

At the metric level, model fit was acceptable as indicated by the fit indices, again except for RMSEA. However, ΔCFI (.021), ΔRMSEA (.013), and ΔSRMR (.031) exceeded their cut-offs, and Δχ2 was significant, suggesting noninvariance at the metric level. Upon inspection and exploration of modification indices, the factor loading for Item 2 (‘In the past two weeks, I have concealed my sexual orientation by avoiding contact with other LGB individuals’) appeared to differ across groups. In the configural model, the standardised loading for this item was .698 for the BIPOC group (M = 1.342, SD = .961) and .433 for the white group (M = 1.127, SD = .533). Thus, Item 2 loaded onto the latent factor more strongly for the BIPOC group. We then performed partial metric and partial scalar invariance analyses with the loading for Item 2 unconstrained. The fit change statistics were acceptable for both models, and the Δχ2 values were nonsignificant for both models. These results suggest partial scalar invariance based on racial/ethnic identity and demonstrate that Item 2 should be carefully considered when making group comparisons between white and BIPOC groups. Below, we also perform exploratory CFAs with Item 2 removed from the scale.

Given that our analyses demonstrated partial scalar invariance, it is appropriate to compare mean SOCS scores between racial/ethnic groups. However, the results of an independent-samples t-test showed no significant differences in SOC between white people and BIPOC (see Table 4). In addition, the item-level means and standard deviations across white and BIPOC participants are displayed in Table 5.

Table 4.

Comparing SOCS means by racial/ethnic identity and sexual orientation.

Sociodemographic variable n M (SD) t/F statistic p
Racial/ethnic identity .866 .387
 BIPOC 73 1.727 (.794)
 white 267 1.822 (.921)
Sexual identity 2.244 .108
 Monosexual 69 1.763 (.974)
 Queer 93 1.613 (.755)
 Bi+ 178 1.842 (.835)
Open in a new tab

Table 5.

SOCS Item-level means by racial/ethnic identity and sexual orientation.

Sociodemographic variable Item 1
M (SD)		 Item 2
M (SD)		 Item 3
M (SD)		 Item 4
M (SD)		 Item 5
M (SD)		 Item 6
M (SD)
Racial/ethnic identity
 BIPOC 1.767 (1.318) 1.342 (.961) 2.027 (1.236) 2.603 (1.534) 1.904 (1.293) 1.740 (1.302)
 white 1.592 (1.115) 1.127 (.533) 1.831 (1.140) 2.558 (1.433) 1.749 (1.128) 1.502 (1.016)
Sexual identity
 Monosexual 1.507 (1.106) 1.246 (.864) 1.942 (1.271) 2.348 (1.474) 1.884 (1.182) 1.652 (1.281)
 Queer 1.473 (1.038) 1.151 (.607) 1.720 (1.077) 2.398 (1.392) 1.602 (1.044) 1.333 (.864)
 Bi+ 1.758 (1.232) 1.157 (.580) 1.927 (1.160) 2.742 (1.462) 1.837 (1.213) 1.629 (1.098)
Open in a new tab

Measurement invariance by sexual orientation

The fit indices, ΔCFI, ΔRMSEA, ΔSRMR, and Δχ2 for the second invariance model (testing invariance by sexual orientation) are also displayed in Table 3. In general, fit statistics met the specified cut-offs, except for RMSEA (.122) which exceeded the cut-off of .08. However, the CFI, TLI, and SRMR supported good fit, and we therefore determined that the model demonstrated configural invariance; the factor structure was equivalent across monosexual, bi+, and queer participants.

Moving on to metric invariance, model fit was similar to the configural output except that SRMR (.088) slightly exceeded the specified cut-off. ΔCFI and ΔSRMR also exceeded their respective cut-offs and Δχ2 was significant, showing that metric invariance was not supported. As in the previous invariance analyses, the factor loading for Item 2 appeared to be problematic. In the configural model, the standardised loading for this item was .413 for the bi+ group (M = 1.157, SD = .580), .581 for the queer group (M = 1.151, SD = .607), and .708 for the monosexual group (M = 1.246, SD = .864). We ran a partial metric invariance analysis with the loading for Item 2 unconstrained. Δχ2 was no longer significant (p = .149) and ΔRMSEA was acceptable (.001), but ΔCFI exceeded its cut-off (.023) and ΔSRMR marginally exceeded its cut-off (.026). Given that ΔSRMR was so close to its cut-off of .025, we tentatively considered this an indicator of invariance at the metric level and elected to proceed with partial scalar invariance models (with the loading for Item 2 unconstrained). In these models, Δχ2 was not significant (.076) and all fit change statistics were acceptable, thus suggesting partial scalar invariance by sexual orientation.

As with the analyses focused on invariance by racial/ethnic identity, we ran a one-way ANOVA to compare mean SOCS scores across sexual orientations. This analysis indicated no significant differences in concealment across sexual orientations (see Table 4).

Follow-up exploratory analyses

As noted, several findings across our analyses suggested that Item 2 may be somewhat problematic. This item had the lowest standardised loading in the base CFA model. In addition, results supported partial metric and scalar invariance (when unconstraining the factor loading and intercept for Item 2). Thus, as an exploratory goal, we re-ran each of our analyses with Item 2 removed to examine whether this would improve the psychometric properties of the SOCS and enhance measurement invariance.

Revised base model confirmatory factor analysis

The revised 5-item base model fit the data well, χ2(5) = 8.228, p = .144, CFI = .995; TLI = .990, RMSEA = .044, SRMR = .018. The fit statistics met their respective cut-offs. Compared to the six-item base model, CFI was higher (.995 versus .964), TLI was higher (.990 versus .939), RMSEA was lower (.044 versus .095), and SRMR was lower (.018 versus .041). The average variance extracted was acceptable (.528) and internal consistency was high (Cronbach’s α = .841). The standardised factor loadings for the remaining items were comparable to those from the original model, each well above the specified .50 cut-off: Item 1 = .824; Item 3 = .764; Item 4 = .723; Item 5 = .645; Item 6 = .655.

Revised model: measurement invariance by racial/ethnic identity

The fit indices, ΔCFI, ΔRMSEA, ΔSRMR, and Δχ2 for the revised invariance by racial/ethnic identity model are displayed in Table 6. Beginning with configural invariance, all fit statistics met the specified cut-offs and demonstrated good fit, signalling that the factor structure was equivalent across white people and BIPOC. Moving on to metric and scalar invariance, the model fits were similarly adequate. The ΔCFIs were less than .01 and Δχ2 values were nonsignificant, and thus both metric and scalar invariance were demonstrated. The meanings of the SOCS items did not vary between white participants and participants of colour following the removal of Item 2.

Table 6.

Revised (5-item) SOCS measurement invariance by racial/ethnic identity and sexual orientation.

Type of invariance χ2 (df) p-value RMSEA (90% CI) TLI CFI SRMR ΔRMSEA ΔCFI ΔSRMR Δχ2 (p-value)
Racial/Ethnic Identity (white versus BIPOC)
 Configural 10.972
(10)		 .360 .024
(.000, .089)		 .997 .998 .018
 Metric 17.527
(14)		 .229 .038
(.000, .088)		 .992 .994 .040 .014 .004 .022 6.555
(.161)
 Scalar 19.826
(18)		 .343 .024
(.000, .074)		 .997 .997 .042 .014 .003 .002 2.299
(.681)
Sexual Orientation (monosexual versus bisexual versus queer)
 Configural 35.677
(15)		 .002 .110
(.064, .157)		 .937 .969 .031
 Metric 49.777
(23)		 .001 .101
(.062, .140)		 .947 .959 .065 .009 .010 .034 14.100
(.079)
Open in a new tab

Note: BIPOC = Black, Indigenous, and People of Colour; SOCS = Sexual Orientation Concealment Scale; RMSEA = root mean square error of approximation; ΔRMSEA = root mean square error of approximation difference; CFI = comparative fit index; ΔCFI = comparative fit index difference; χ2 = Chi-squared; Δχ2 = Chi-squared difference; TLI = Tucker-Lewis index; SRMR = standardised root mean square residual; ΔSRMR = standardised root mean square residual difference.

Revised model: measurement invariance by sexual orientation

As with the previous measurement invariance analysis, the fit indices, ΔCFI, and Δχ2 for the revised invariance model (by sexual orientation) are shown in Table 6. Starting with configural invariance, all fit statistics fell within the specified cut-offs and demonstrated good fit except RMSEA (.110). However, given that all other fit statistics exhibited good fit, we determined that configural variance was established and thus that the SOCS factor structure was equivalent across monosexual, bi+, and queer participants. In metric invariance analyses, ΔCFI met its cut-off (.01) and ΔSRMR slightly exceeded its cut-off (.034); thus, metric invariance was not supported.

Discussion

A rich literature has documented associations between SOC and a wide array of mental health challenges among cis LGBTQ+ people (e.g. depressive symptoms, anxiety symptoms, distress, problematic eating behaviours; Pachankis, Mahon, et al., 2020). It is essential to expand this research base to include a focus on the experiences of nonbinary people, particularly given that they face unique and insidious forms of marginalisation (that undoubtedly shape their concealment processes), and are frequently excluded from LGBTQ+ health research (Matsuno & Budge, 2017). However, a valid measure of SOC is required to do so, especially given the distinctness of minority stressors experienced by nonbinary people (Matsuno et al., 2026). The SOCS (Jackson & Mohr, 2016) is a brief measure of SOC that has shown promising psychometric properties with predominantly cis LGBTQ+ people, but we are aware of no studies investigating its appropriateness for use with nonbinary people. The present study evaluated whether the SOCS (Jackson & Mohr, 2016) demonstrates strong psychometric properties with nonbinary individuals, including invariance by racial/ethnic identity and sexual orientation.

The results supported the internal structure of the SOCS, showing that the original one-factor model demonstrated acceptable fit and high internal consistency. The only fit index that exceeded its cut-off in the six-item CFA was RMSEA, which is common in models with few degrees of freedom and small to moderate sample sizes (Kenny et al., 2015). We also investigated measurement invariance across racial/ethnic groups (white, BIPOC) and sexual orientation groups (monosexual, bi+, queer), and found evidence of configural invariance and partial metric and scalar invariance. To pursue an exploratory aim, we removed a problematic (noninvariant) item (Item 2) and re-ran our analyses. This improved fit indices and further supported scalar invariance across racial/ethnic identities, and supported configural invariance across sexual orientations. Metric invariance for the revised scale model based on sexual orientation was found based on some fit statistics, but slightly missed for others; thus, the practical implications of noninvariance may be negligible, but further research is warranted to confirm this. Overall, these findings lend preliminary support to the SOCS as an appropriate tool for assessing SOC among diverse nonbinary populations.

We could not ultimately discern the reason Item 2 (i.e. ‘In the past two weeks, I have concealed my sexual orientation by avoiding contact with other LGB individuals’) was noninvariant. However, we would like to highlight a few possibilities. Relative to the other items, it is the only one focused on avoiding contact with LGB people and thus appears conceptually distinct. Perceptions of what it means to be in contact with LGB people may vary across cultures, such that contact with LGB people is viewed as strongly indicative of an LGB identity in some cultures and less so in others. It is also certainly possible that responses to Item 2 across white and BIPOC participants were shaped by intersectional oppression and violence (Crenshaw, 1989, 1991). For instance, some BIPOC may conceal their sexual orientations to eschew contact with other LGB individuals due to experiences of racism and/or xenophobia within the LGBTQ+ community (e.g. colourist remarks, explicit whitewashing of LGBTQ+ culture). Indeed, Parmenter et al. (2026) highlighted that identity fragmentation, safety concerns, and exclusion often characterise BIPOC experiences in LGBTQ+ spaces; moreover, intense ‘pressures to assimilate to dominant narratives of Whiteness’ (p. 7) repeatedly alienate BIPOC. In contrast, as discussed in literature on white homonormativity (e.g. Vo, 2021), white LGB people find their experiences spotlighted and centred across LGBTQ+ spaces and they are equipped with immense racial power and privilege. These divergent racialised experiences shape how LGB people negotiate (or avoid) interactions with one another (Parmenter et al., 2026). Regarding sexual orientation, it is worth considering that bi+ people experience unique forms of marginalisation within the LGBTQ+ community (e.g. identity invalidation; McLean, 2008; Oswald et al., 2024) relative to gay, lesbian, and many queer people; this may impact their approaches to SOC via disengagement from LGBTQ+ community members.

Study limitations and considerations

This study is characterised by several notable limitations, many of which pertain to the sample composition. Although we sought to recruit a representative participant group, this was a non-random community sample and a majority of participants were white and lived in western countries (e.g. US, UK). Sample diversity may have been hindered by the questionnaire solely being offered in English. Moreover, although we assessed invariance by racial/ethnic identity and sexual orientation, other invariance analyses that were not conducted (e.g. by assigned sex, by age cohort) may be important. Another methodological consideration was our decision to allow participants to endorse only one sexual orientation. These identities frequently overlap (e.g. people may identify as both queer and bi+), and thus, the present analyses may be most accurately interpreted as invariance by primary sexual orientation.

Although the findings demonstrated invariance of the SOCS by racial/ethnic identity, they should be generalised with caution to nonbinary BIPOC because there were relatively few BIPOC participants in the study and very small numbers of participants from specific racial/ethnic groups under the BIPOC umbrella (e.g. Black people, Asian people). The decision to examine BIPOC experiences unidimensionally was a methodological necessity due to power concerns, and it is thus vital to interpret the present data as only preliminary evidence of measurement invariance across racial/ethnic groups. This approach was not an attempt to conflate the experiences of BIPOC or dismiss meaningful differences between them (Nielsen, 2000). Furthermore, we understand that (even tacitly) reinforcing a white vs. BIPOC binary can unfortunately limit the nuance brought to dialogues about racial power, privilege, and domination (Al-Yagout, 2017). We hope this work prompts other scholars’ interest in studying invariance of the SOCS (among other measures of minority stress) across racial/ethnic identities further, which will help scholars avoid making erroneous claims about group differences that are ultimately produced by measurement artefacts (Sullivan et al., 2025) and resist the whitewashing of LGBTQ+ health science.

Consistent with research centred on predominantly cis populations (e.g. Jackson & Mohr, 2016), the participants reported low levels of SOC on average. A pervasive challenge with studies focused on SOC is that individuals with moderate to high concealment behaviours are unlikely to participate in research about their sexual orientations; doing so requires them to disclose an LGBQ+ identity to researchers, with which many are uncomfortable (Pachankis, Mahon, et al., 2020). Although this study took place online, which may reduce closeted or highly concealed individuals’ concerns about participating (Glynn et al., 2024), it is unsurprising that the representation of moderate to highly concealed individuals was low.

Implications for future research

Researchers should consider widely using the SOCS to understand the experiences of nonbinary people knowing, with a modest degree of confidence, that it accurately captures them. We hope this study ignites an interest in studying the consequences of concealment among nonbinary people, which will clarify whether, how, and to what degree they are harmful to nonbinary people’s health. To our knowledge, there is only one study in this area to date (Feinstein et al., 2020). In particular, we hope forthcoming research aims to confirm whether the correlates of SOC for predominantly cis populations are also correlates for nonbinary people. For instance, a meta-analysis by Pachankis, McConocha, et al. (2020) reported a small positive association between SOC and symptoms of both depression and anxiety. Studies aiming to replicate this finding with nonbinary people – particularly using an intersectional approach – would enhance the cohesion of this literature and may have important implications for clinical practice.

Some methodological limitations of this study, too, point to critical avenues for future investigation. For instance, it would be helpful for scholars to replicate this study with a more highly concealed sample, perhaps taking extra measures to clarify how participants’ data will be anonymised and protected. Scholars may also choose to test the psychometric properties and measurement invariance of the SOCS using both the original six-item scale as well as the revised five-item scale (i.e. removing Item 2), given that this item was problematic in invariance tests across identity groups. Further investigation of both the six-item and five-item scales would help to confirm which version is optimal. Future work may also consider investigating other forms of validity evidence of the SOCS for nonbinary people, such as predictive, convergent, and discriminant validity.

Although we tested measurement invariance by racial/ethnic identity, this analysis was preliminary and limited by the moderately small sample size of BIPOC (n = 73). Forthcoming work would also benefit from recruiting larger samples of nonbinary BIPOC and test the invariance of the SOCS across specific racial/ethnic groups under the BIPOC umbrella (e.g. Black people, Hispanic and/or Latinx people). Such studies would also further clarify how well RMSEA performs on the SOCS, as the present sample was fairly small, which may have artificially inflated RMSEA (Kenny et al., 2015). In addition, investigating the invariance of the SOCS across other identity factors (e.g. sexual orientations, age cohorts, assigned sex) would further refine our understanding of its validity for nonbinary and other LGBTQ+ populations.

Implications for clinical practice

Given that the SOCS is a brief, valid tool for assessing SOC among nonbinary people, practitioners (e.g. psychologists, primary care providers) should consider incorporating it into their practice with this population. As per both theoretical (e.g. Meyer, 1995) and empirical research (Pachankis, Mahon, et al., 2020), SOC is a notable minority stressor; although research in this area is limited at present, SOC seems to have pressing health outcomes for nonbinary people. In a primary care setting, medical professionals (e.g. physicians) can administer the SOCS to nonbinary patients and use high scores to flag those who may benefit from psychotherapy to address SOC and any co-occurring mental health concerns. In addition, psychotherapists may consider integrating the SOCS into their intake paperwork to inform treatment plans. Nonbinary clients who are more highly concealed, for instance, may have higher levels of shame about their LGBQ+ identities (see Mohr & Kendra, 2011), thus highlighting a useful area for intervention in psychotherapy. Psychotherapists may consider following or incorporating strategies from trans-affirming psychotherapy programmes, such as AFFIRM (Austin & Craig, 2015). They may also draw from Effective Skills to Empower Effective Men (ESTEEM; Pachankis, 2014; Pachankis et al., 2019), an LGBTQ+-affirming programme that incorporates a focus on concealment and has been tailored for use with people of various genders including non-binary people (Pachankis, McConocha, et al., 2020).

Conclusion

The present study is the first to provide empirical support for the validity of the SOCS with nonbinary populations, including partial scalar invariance between nonbinary white people and BIPOC, and sexual orientation groups. We reason that this is an important development in the literature, as no SOC measures have been validated for use with nonbinary people. Findings may pave the way for rigorous scholarship on the correlates and consequences of SOC in this population. Practitioners working with nonbinary people should consider incorporating the SOCS into their service delivery.

Funding

Peter D. Goldie acknowledges salary support from grant 5T32HL161270-05 from the National Heart, Lung, and Blood Institute.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, Suzanne McLaren, upon reasonable request.

References

  1. Al-Yagout N. (2017). Who is a ‘person of color?’ Humboldt Journal of Social Relations, 39(39), 5–7. 10.2307/90007864 [DOI] [Google Scholar]
  2. Austin A, & Craig SL. (2015). Transgender affirmative cognitive behavioral therapy: Clinical considerations and applications. Professional Psychology, Research and Practice, 46(1), 21–29. 10.1037/a0038642 [DOI] [Google Scholar]
  3. Blondeel K, Say L, Chou D, Toskin I, Khosla R, Scolaro E, & Temmerman M. (2016). Evidence and knowledge gaps on the disease burden in sexual and gender minorities: A review of systematic reviews. International Journal for Equity in Health, 15(1), Article 16. 10.1186/s12939-016-0304-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bradford NJ, & Catalpa JM. (2019). Social and psychological heterogeneity among binary transgender, non-binary transgender and cisgender individuals. Psychology & Sexuality, 10(1), 69–82. 10.1080/19419899.2018.1552185 [DOI] [Google Scholar]
  5. Brooks VR. (1981). Minority stress and lesbian women. Lexington Books. [Google Scholar]
  6. Brown A. (2022, June). About 5% of young adults in the U.S. say their gender is different from their sex assigned at birth. Pew Research Center. https://www.pewresearch.org/fact-tank/2022/06/07/about-5-of-young-adults-in-the-u-s-say-their-gender-is-different-from-their-sex-assigned-at-birth/ [Google Scholar]
  7. Brown TA, Klimek-Johnson P, Siegel JA, Convertino AD, Douglas VJ, Pachankis J, & Blashill AJ. (2024). Promoting resilience to improve disordered eating (PRIDE): A case series of an eating disorder treatment for sexual minority individuals. International Journal of Eating Disorders, 57(3), 648–660. 10.1002/eat.24150 [DOI] [PubMed] [Google Scholar]
  8. Browne MW, & Cudeck R. (1993). Alternative ways of assessing model fit. In A Bollen K. & Long JS. (Eds.), Testing structural equation models (pp. 136–162). Sage. [Google Scholar]
  9. Chen FF. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504. 10.1080/10705510701301834 [DOI] [Google Scholar]
  10. Crenshaw K. (1989). Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. University of Chicago Legal Forum, 140, 139–167. https://chicagounbound.uchicago.edu/uclf/vol1989/iss1/8/?utm_source=chicagounbound.uchicago.edu/uclf/vol1989/iss1/8&utm_medium=PDF&utm_campaign=PDFCoverPages [Google Scholar]
  11. Crenshaw K. (1991). Mapping the margins: Intersectionality, identity politics and the violence against women of color. Stanford Law Review, 43(6), 1241–1299. 10.2307/1229039 [DOI] [Google Scholar]
  12. Csoti JP, & McLaren S. (2025). Sexual orientation concealment, hope, and depressive symptoms among gay and bisexual men: A moderated-moderation model. Psychology & Sexuality, 16(1), 72–85. 10.1080/19419899.2024.2348717 [DOI] [Google Scholar]
  13. DeChants JP, Shelton J, Anyon Y, & Bender K. (2022). “It kinda breaks my heart”: LGBTQ young adults’ responses to family rejection. Family Relations, 71(3), 968–986. 10.1111/fare.12638 [DOI] [Google Scholar]
  14. Dunn IB, Power E, Casey LJ, & Wootton BM. (2025). Efficacy of videoconferencing-delivered cognitive behavioural therapy to reduce anxiety disorder severity in LGBTQ+ people: An exploratory trial protocol. PLOS ONE, 20(1), Article e0316857. 10.1371/journal.pone.0316857 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Feinstein BA, Xavier Hall CD, Dyar C, & Davila J. (2020). Motivations for sexual identity concealment and their associations with mental health among bisexual, pansexual, queer, and fluid (bi+) individuals. Journal of Bisexuality, 20(3), 324–341. 10.1080/15299716.2020.1743402 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fiani CN, & Han HJ. (2019). Navigating identity: Experiences of binary and non-binary transgender and gender non-conforming (TGNC) adults. The International Journal of Transgenderism, 20(2–3), 181–194. 10.1080/15532739.2018.1426074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Flores AR, Stotzer RL, Meyer IH, & Langton LL. (2022). Hate crimes against LGBT people: National crime victimization survey, 2017–2019. PLOS ONE, 17(12), e0279363. 10.1371/journal.pone.0279363 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Glynn SA, McLaren S, & Goldie PD. (2024). Sexual orientation concealment, hope, and depressive symptoms among sexual minority women: The moderating role of sexual orientation. Journal of Homosexuality, 71(11), 2594–2616. 10.1080/00918369.2023.2250500 [DOI] [PubMed] [Google Scholar]
  19. Hair JF, Black WC, Babin BJ, & Anderson RE. (2009). Multivariate data analysis (7th ed.). Prentice Hall. [Google Scholar]
  20. Hammack PL, Hughes SD, Atwood JM, Cohen EM, & Clark RC. (2022). Gender and sexual identity in adolescence: A mixed-methods study of labeling in diverse community settings. Journal of Adolescent Research, 37(2), 167–220. 10.1177/07435584211000315 [DOI] [Google Scholar]
  21. Hendricks ML, & Testa RJ. (2012). A conceptual framework for clinical work with transgender and gender nonconforming clients: An adaptation of the minority stress model. Professional Psychology, Research and Practice, 43(5), 460–467. 10.1037/a0029597 [DOI] [Google Scholar]
  22. Hitch AE, & Brown JL. (2024). Sexual orientation concealment and mental health among black, indigenous, and people of color men who have sex with men living in the US: A scoping review. Journal of Gay and Lesbian Mental Health, 28(4), 525–548. 10.1080/19359705.2023.2239732 [DOI] [Google Scholar]
  23. Hu L-T, & Bentler PM. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  24. Huang Y-T, & Chan RCH. (2022). Effects of sexual orientation concealment on well-being among sexual minorities: How and when does concealment hurt? Journal of Counseling Psychology, 69(5), 630–641. 10.1037/cou0000623 [DOI] [PubMed] [Google Scholar]
  25. Ipsos. (2021). LGBT+ pride 2021 global survey: A 27-country ipsos survey. https://www.ipsos.com/sites/default/files/ct/news/documents/2021-06/lgbt-pride-2021-global-survey-ipsos.pdf
  26. Jackson SD, & Mohr JJ. (2016). Conceptualizing the closet: Differentiating stigma concealment and nondisclosure processes. Psychology of Sexual Orientation and Gender Diversity, 3(1), 80–92. 10.1037/sgd0000147 [DOI] [Google Scholar]
  27. Kase CA, & Mohr JJ. (2023). Reciprocal associations between sexual orientation concealment and mental health among LGBQ college students. Journal of Gay and Lesbian Mental Health. Advance online publication. 29(1), 69–85. 10.1080/19359705.2023.2258819 [DOI] [Google Scholar]
  28. Kenny DA, Kaniskan B, & McCoach DB. (2015). The performance of RMSEA in models with small degrees of freedom. Sociological Methods & Research, 44(3), 486–507. 10.1177/0049124114543236 [DOI] [Google Scholar]
  29. Le Forestier JM, Page-Gould E, & Chasteen AL. (2022). Concealment stigma: The social costs of concealing. Journal of Experimental Social Psychology, 101, Article 104340. 10.1016/j.jesp.2022.104340 [DOI] [Google Scholar]
  30. Matsuno E, Bricker NL, Collazo E, Huynh KD, Mohr R Jr., Colson AE, & Balsam KF. (2026). Development and validation of the nonbinary distal minority stressors, proximal minority stressors, and resilience scales. Psychology of Sexual Orientation and Gender Diversity, 13(1), 35–53. 10.1037/sgd0000719 [DOI] [Google Scholar]
  31. Matsuno E, & Budge SL. (2017). Non-binary/genderqueer identities: A critical review of the literature. Current Sexual Health Reports, 9(3), 116–120. 10.1007/s11930-017-0111-8 [DOI] [Google Scholar]
  32. McLean K. (2008). Inside, outside, nowhere: Bisexual men and women in the gay and lesbian community. Journal of Bisexuality, 8(1–2), 63–80. 10.1080/15299710802143174 [DOI] [Google Scholar]
  33. Meuleman B, Żółtak T, Pokropek A, Davidov E, Muthén B, Oberski DL, Billiet J, Schmidt P. (2023). Why measurement invariance is important in comparative research. A response to Welzel et al. (2021). Sociological Methods & Research, 52(3), 1401–1419. 10.1177/00491241221091755 [DOI] [Google Scholar]
  34. Meyer IH. (1995). Minority stress and mental health in gay men. Journal of Health and Social Behaviour, 36(1), 38–56. 10.2307/2137286 [DOI] [PubMed] [Google Scholar]
  35. Meyer IH. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129(5), 674–697. 10.1037/0033-2909.129.5.674 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Meyer IH, Flores AR, Stemple L, Romero AP, Wilson BD, & Herman JL. (2017). Incarceration rates and traits of sexual minorities in the United States: National inmate survey, 2011–2012. American Journal of Public Health, 107(2), 267–273. 10.2105/AJPH.2016.303576 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mohr JJ, & Kendra MS. (2011). Revision and extension of a multidimensional measure of sexual minority identity: The lesbian, gay, and bisexual identity scale. Journal of Counseling Psychology, 58(2), 234–245. 10.1037/a0022858 [DOI] [PubMed] [Google Scholar]
  38. Nielsen LB. (2000). Situating legal consciousness: Experiences and attitudes of ordinary citizens about law and street harassment. Law and Society Review, 34(4), 1055–1090. 10.2307/3115131 https://www.jstor.org/stable/3115131 [DOI] [Google Scholar]
  39. Oswald F, Walton KA, Orlova L, Ayad J, Hooper A, & Pedersen CL. (2024). “I had already reconciled with being a slut when I came out as bisexual”: Experiences of binormativity, prototypicality, and marginalization within the bisexual community. Journal of Bisexuality, 24(2), 150–184. 10.1080/15299716.2024.2310159 [DOI] [Google Scholar]
  40. Pachankis JE. (2007). The psychological implications of concealing a stigma: A cognitive-affective-behavioral model. Psychological Bulletin, 133(2), 328–345. 10.1037/0033-2909.133.2.328 [DOI] [PubMed] [Google Scholar]
  41. Pachankis JE. (2014). Uncovering clinical principles and techniques to address minority stress, mental health, and related health risks among gay and bisexual men. Clinical Psychology: Science and Practice, 21(4), 313–330. 10.1111/cpsp.12078 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Pachankis JE, & Bränström R. (2018). Hidden from happiness: Structural stigma, sexual orientation concealment, and life satisfaction across 28 countries. Journal of Consulting & Clinical Psychology, 86(5), 403–415. 10.1037/ccp0000299 [DOI] [PubMed] [Google Scholar]
  43. Pachankis JE, Mahon CP, Jackson SD, Fetzner BK, & 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]
  44. Pachankis JE, McConocha EM, Clark KA, Wang K, Behari K, Fetzner BK, Brisbin CD, Scheer JR, & Lehavot K. (2020). A transdiagnostic minority stress intervention for gender diverse sexual minority women’s depression, anxiety, and unhealthy alcohol use: A randomized controlled trial. Journal of Consulting & Clinical Psychology, 88(7), 613–630. 10.1037/ccp0000508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pachankis JE, McConocha EM, Reynolds JS, Winston R, Adeyinka O, Harkness A, Burton CL, Behari K, Sullivan TJ, Eldahan AI, Esserman DA, Hatzenbuehler ML, & Safren SA. (2019). Project esteem protocol: A randomized controlled trial of an LGBTQ-affirmative treatment for young adult sexual minority men’s mental and sexual health. BMC Public Health, 19(1), Article 1086. 10.1186/s12889-019-7346-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Parmenter JG, Lê D, Barrita A, Muskrat T, Easton A, Staton D, & Trenary J. (2026). “I think I take on the whole system”: Experiences of intersectional proximal minority stressors among queer, transgender, and nonbinary people of color. Cultural Diversity and Ethnic Minority Psychology. Advance online publication. 10.1037/cdp0000789 [DOI] [PubMed] [Google Scholar]
  47. Ragins BR, Singh R, & Cornwell JM. (2007). Making the invisible visible: Fear and disclosure of sexual orientation at work. Journal of Applied Psychology, 92(4), 1103–1118. 10.1037/0021-9010.92.4.1103 [DOI] [PubMed] [Google Scholar]
  48. Ramos SR, Lardier DT Jr., Opara I, Turpin RE, Boyd DT, Gutierrez JI Jr., Williams CN, Nelson LE, & Kershaw T. (2021). Intersectional effects of sexual orientation concealment, internalized homophobia, and gender expression on sexual identity and HIV risk among sexual minority men of color: A path analysis. Journal of the Association of Nurses in AIDS Care, 32(4), 495–511. 10.1097/JNC.0000000000000274 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. R Core Team. (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ [Google Scholar]
  50. Reisner SL, Poteat T, Keatley J, Cabral M, Mothopeng T, Dunham E, Holland CE, Max R, & Baral SD. (2016). Global health burden and needs of transgender populations: A review. Lancet, 388(10042), 412–436. 10.1016/S0140-6736(16)00684-X [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Schrimshaw EW, Siegel K, Downing MJ Jr., & Parsons JT. (2013). Disclosure and concealment of sexual orientation and the mental health of non-gay-identified, behaviorally bisexual men. Journal of Consulting & Clinical Psychology, 81(1), 141–153. 10.1037/a0031272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Sullivan TJ, Huang H, Vilkin E, Brandt SA, & Davila J. (2025). A measurement invariance/equivalence study of sexual minority stress measures across sexual orientation, gender identity, and race/ethnicity. Psychology of Sexual Orientation and Gender Diversity, 12(3), 468–476. 10.1037/sgd0000681 [DOI] [Google Scholar]
  53. Szymanski DM, & Sung MR. (2013). Asian cultural values, internalized heterosexism, and sexual orientation disclosure among Asian American sexual minority persons. Journal of LGBT Issues in Counseling, 7(3), 257–273. 10.1080/15538605.2013.812930 [DOI] [Google Scholar]
  54. Vo TD. (2021). Gaps in safety within LGBTQ+ leisure spaces for diverse LGBTQ+ people: White homonormativity and considerations for inclusion in safe spaces. Social Work & Policy Studies: Social Justice, Practice and Theory, 4(1), 1–24. [Google Scholar]
  55. Wittgens C, Fischer MM, Buspavanich P, Theobald S, Schweizer K, & Trautmann S. (2022). Mental health in people with minority sexual orientations: A meta-analysis of population-based studies. Acta Psychiatrica Scandinavica, 145(4), 357–372. 10.1111/acps.13405 [DOI] [PubMed] [Google Scholar]
  56. Zeeman L, Sherriff N, Browne K, McGlynn N, Mirandola M, Gios L, Davis R, Sanchez-Lambert J, Aujean S, Pinto N, Farinella F, Donisi V, Niedźwiedzka-Stadnik M, Rosińska M, Pierson A, Amaddeo F, Network H, Toskin I, Jonas K, … Troussier T. (2019). A review of lesbian, gay, bisexual, trans and intersex (LGBTI) health and healthcare inequalities. European Journal of Public Health, 29(5), 974–980. 10.1093/eurpub/cky226 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author, Suzanne McLaren, upon reasonable request.

RESOURCES