A Pilot Evaluation of Sexual and Gender Minority Identity Measures in a Treatment-Engaged Military Veteran Sample

Abstract

Sexual and gender minority (SGM) military veterans have endured a history of discriminatory policies and hetero- and cis-sexist related military culture that can negatively impact identity and mental health. The present pilot evaluation examined measure characteristics of the Lesbian, Gay, and Bisexual Identity Scale (LGBIS) and Lesbian, Gay, and Bisexual-Positive Identity Measure (LGB-PIM) in a clinical sample of SGM military veterans in order to assess the potential use of these instruments in understanding identity and mental health in the context of program implementation. A cross-sectional pilot survey of 83 SGM veterans was conducted in ten Veterans Affairs sites. Self-report data were collected as part of a quality improvement project across 2018 and 2019. Results showed that the sample was characterized by low internalized prejudice and identity uncertainty, as well as generally high positive aspects of identity (e.g., identity affirmation, authenticity, social justice beliefs). LGB-PIM subscale internal consistency values were acceptable (α range = .89-.92), whereas LGBIS subscale values varied (α range = .51-.87). Acceptance concerns, identity uncertainty, and social justice beliefs distinguished mental health symptom severity levels. Higher identity uncertainty and social justice beliefs were associated with worse symptoms of depression, anxiety, and suicide risk. Preliminary results support further application and study of the LGB-PIM and some LGBIS subscales as possible tools in program development and improvement within military veteran samples.

Sexual and gender minority (SGM) identity broadly refers to a process of self-labeling, evaluation of SGM identity (e.g., internalized stigma), and assigning importance to one’s identity (e.g., centrality; Mohr & Kendra, 2011). The term SGM captures both sexual minority (e.g., lesbian, gay, bisexual) and gender minority (e.g., transgender, non-binary) identities (Centers for Disease Control and Prevention, 2019). Minority stress due to one’s sexual orientation and/or gender identity (e.g., discrimination, rejection, harassment) limits positive SGM identity development (Riggle et al., 2014) and negatively impacts mental health (e.g., depression, anxiety) and well-being (Meyer, 2003; Riggle et al., 2014; Testa et al., 2015). Further, stigma and rejection in the military present additional threats to SGM identity among veterans; for example, now rescinded military policies such as Don’t Ask Don’t Tell (DADT) and the transgender military ban contributed to a culture of SGM identity concealment that persists today (McNamara et al., 2020). Indeed, SGM veterans report greater concealment and lower levels of outness compared to non-veteran SGM persons (Moradi, 2009), which is associated with negative mental health outcomes, including depression and posttraumatic stress disorder symptoms (Cochran et al., 2013).

Sexual and Gender Minority Identity Measures

The Lesbian, Gay, and Bisexual Identity Scale (LGBIS) assesses sexual minority identity across eight dimensions: acceptance concerns, concealment motivation, identity uncertainty, internalized prejudice, difficult process, identity superiority, identity affirmation, and identity centrality (α range = .76 to .89; Mohr & Kendra, 2011). Mean scores in the original development samples varied. The LGBIS subscales are scored as item-level means (1–6 scale range; 1 = disagree strongly, 6 = agree strongly; Mohr & Kendra, 2011). Assessment of item-level means can lend insight into how groups of SGM persons tend to score on negative and positive aspects of identity. For instance, scores well above the mid-point (3.5 on a 1–6 scale) may reflect self-perceived negative identity on certain subscales (e.g., concealment motivation) and positive self-concept on others (e.g., identity affirmation). Acceptance concerns and concealment motivation were about equal to the scale mid-point, whereas difficult process, identity uncertainty, internalized prejudice, and identity superiority were below the scale mid-point. Positive aspects of identity (i.e., identity affirmation and identity centrality) were above the scale mid-point (Mohr & Kendra, 2011) in general, non-military samples.

Psychometric studies support this eight-factor model among sexual minority college students (Mohr & Kendra, 2011) and a Turkish sample (Kemer et al., 2017). Support for a six-factor model, inclusive of a negative identity higher-order factor (comprised of concealment motivation, internalized prejudice, difficult process, and acceptance concerns), was observed among a sample of sexual minority adults, although internal consistency was low for four subscales: identity superiority (α = .64), difficult process (α = .52), acceptance concerns (α = .60), and identity centrality (α = .62; Cramer et al., 2017). Further support was observed for both the six- and eight-factor models among sexual minority adults (Cramer et al., 2018). LGBIS item means varied by subscale among low income urban-dwelling sexual minority adults (Cramer et al., 2017) and sexual minority members of a sexuality special interest group (Cramer et al., 2018). Across these samples, the negative identity higher-order factor, as well as specific subscales of identity uncertainty, identity superiority, acceptance concerns, difficult process, concealment motivation, and internalized prejudice, were below the scale mid-point (Cramer et al., 2017, 2018). Identity affirmation was above the scale mid-point, whereas identity centrality was about equal to the scale mid-point (Cramer et al., 2018).

Among the college student sample, the following subscales demonstrated significant small-to-moderate associations with depression: acceptance concerns, internalized prejudice, and difficult process (Mohr & Kendra, 2011). Later LGBIS studies showed that concealment motivation (Cramer et al., 2018), and identity uncertainty and the higher-order negative identity factor (Cramer et al., 2017), demonstrated significant small correlations with mental health indicators (e.g., anxiety). The LGBIS has been used in non-clinical and non-military samples in the study of mental health, physical health, and coping (Cramer et al., 2017, 2018; Denton et al., 2014); however, use of the LGBIS with military and clinical samples is scant. In a study with SGM veterans, LGBIS-based analyses showed that identity centrality was associated with older age among veterans (Cortes et al., 2019); however, reliability and other important information were not reported, and the study did not examine associations between LGBIS subscales and health outcomes among veterans. Overall, the available data on the LGBIS suggests varying internal consistencies and subscale mean scores, with little application in military or clinical populations.

The Lesbian, Gay, and Bisexual Positive Identity Measure (LGB-PIM) assesses positive sexual minority identity across five dimensions: self-awareness, authenticity, community, intimacy, and social justice (α range = .85 to .95; Riggle et al., 2014). Among the development sample of sexual minority adults recruited through community listservs and websites, all subscale means were well above the mid-point (Riggle et al., 2014), suggesting that participants generally regarded their sexual orientations positively. Authenticity and community involvement displayed small positive associations with depression (Riggle et al., 2014). The LGB-PIM has been used in the study of relationship satisfaction (Mark et al., 2020), psychological well-being (Riggle et al., 2017), and sexual minority group differences in well-being and identity (Petrocchi et al., 2020). Psychometric studies of the LGB-PIM are limited to the initial development and validation studies with non-clinical sexual minority adult samples (Riggle et al., 2014).

Little published research uses the LGB-PIM with military or clinical samples. Two recent exceptions were application of the LGB-PIM in veteran samples to understand basic group differences in aspects positive identity. Kauth and colleagues (2019) found that SGM veterans who used Veterans Affairs care, compared to those using non-Veterans Affairs care, were lower in social justice beliefs. Kauth et al. (2017) reported a significant pattern among SGM veterans in which gay men in rural areas reported lower scores on the community subscale compared to their suburban/urban counterparts. Notably, these studies did not report comprehensive psychometric properties of the LGB-PIM, leaving a gap in the literature in understanding such measurement characteristics among SGM veterans. Further, neither study examined the associations between SGM identity characteristics and mental health outcomes among veterans, highlighting the potential for the current study to address a gap in the literature of importance to clinical providers.

The Present Study

The unique minority stressors faced by SGM veterans, namely the additional layer of minority stress stemming from the military, warrants examination of SGM identity measures among SGM veteran populations. Exploration of the utility of SGM identity measures among SGM veterans engaged in treatment may inform such instruments’ use in future program development and clinical practice. To the extent psychometric properties are acceptable and consistent with prior literature, SGM identity measures can be appropriately applied in future research and program evaluation. For instance, the LGB-PIM may be a promising tool for inclusion in outcome assessment for SGM veteran-focused interventions where a current gold standard has not yet been identified. Further, to the degree LGBIS or LGB-PIM subscales may be related to mental health, measurement and targeting of identity aspects may be warranted in future intervention development for SGM veterans engaged in care. By identifying potential linkages between identity-related factors and mental health outcomes, the findings of the current study may also inform best practices outlined in VA policy related to inclusion and affirmative SGM care (e.g., VHA Directive 1340: Provision of Healthcare for Veterans who Identify as Lesbian, Gay, or Bisexual). As such, the current study sought to examine the measure characteristics of the LGBIS and LGB-PIM in a clinical sample of SGM military veterans.

Consistent with recent federal strategic planning in SGM health (National Institutes of Health [NIH], 2020), we examined two aims regarding the LGBIS and LGB-PIM among a pilot clinical sample of SGM veterans: (1) To evaluate descriptive data and internal consistency; and (2) To examine associations with indicators of mental health.

Methods

Procedure.

Ten VA Medical Centers (VAMC) participated in the Pride in All Who Served 10-week health education group for SGM veterans in 2018–2019 (Lange et al., 2020). The initial feasibility and implementation pilot (e.g., manual development, trialability at new sites, Lange et al., 2020) and a program to spread the practice to additional VA medical centers (Hilgeman et al., in press) were both supported by the VA Spark–Seed–Spread Innovation Investment Program, which identifies and spreads innovative ideas that improve care quality for veterans (Vega & Kizer, 2020). SGM veterans were offered a voluntary, anonymous questionnaire during the first and last sessions of the group as part of the clinical program evaluation; no participant compensation was offered. Outcome measures were selected to be consistent with a Minority Stress Model framework (e.g., Meyer, 2003; Testa et al., 2015). Only baseline assessments are included in this analysis. The pilot procedures were reviewed by the Tuscaloosa VA Institutional Review Board and determined to be non-research, consistent with VA policy and Innovation Network goals (Vega & Kizer, 2020). Investigators complied with APA ethical standards in the treatment of veterans in this pilot program.

Participants.

A total of 83 SGM adult military veterans completed surveys during session one. Table 1 contains full demographic information. The sample was of middle-adult average age; diverse with respect to race, sexual orientation, gender identity, and military branch; and ethnicity was predominantly non-Hispanic/Latinx.

Table 1.

Sample Descriptive Statistics.

Variable (N=83)	n (%)	M (SD)	Skewness (SE)	Kurtosis (SE)
Age (years)	-	47.55 (14.05)	0.09 (0.27)	−1.20 (0.53)
Depressive Symptomsa	-	10.19 (6.77)	0.35 (0.28)	−0.76 (0.54)
Anxiety Symptomsb	-	8.73 (6.12)	0.33 (0.28)	−1.10 (0.58)
12-Month Suicidal Ideationc	-	2.28 (1.47)	0.78 (0.26)	−0.80 (0.52)
Future Suicide Attempt Likelihoodc	-	1.62 (1.56)	0.85 (0.26)	0.12 (0.52)
Lifetime Suicidal Behaviorc	-	8.35 (3.71)	0.52 (0.26)	−0.41 (0.52)
Lesbian, Gay and Bisexuality Identity Scaled:
Negative Identity	-	3.28 (0.95)	0.09 (0.26)	−0.54 (0.52)
Acceptance Concerns	-	3.64 (1.40)	−0.26 (0.26)	−0.80 (0.52)
Concealment Motivation	-	3.82 (1.29)	−0.35 (0.26)	−0.72 (0.52)
Identity Uncertainty	-	1.90 (0.98)	1.28 (0.26)	1.16 (0.52)
Internalized Prejudice	-	2.27 (1.51)	0.85 (0.26)	−0.70 (0.52)
Difficult Process	-	3.40 (1.25)	0.07 (0.26)	−0.61 (0.52)
Identity Superiority	-	1.57 (0.69)	1.65 (0.26)	3.91 (0.52)
Identity Affirmationd	-	4.55 (1.31)	−0.85 (0.26)	0.02 (0.52)
Identity Centralityd	-	3.91 (1.03)	−0.46 (0.26)	−0.50 (0.52)
Lesbian, Gay and Bisexual Positive Identity Measuree:
Self-Awareness	-	5.22 (1.34)	−0.90 (0.26)	0.84 (0.52)
Authenticity	-	5.50 (1.40)	−1.14 (0.26)	1.07 (0.52)
LGBT Community Involvement	-	4.24 (1.47)	−0.15 (0.26)	−0.34 (0.52)
Intimacy	-	4.72 (1.65)	−0.62 (0.26)	−0.24 (0.52)
Social Justice Beliefs	-	5.32 (1.28)	−0.86 (0.26)	0.40 (0.52)
Race:
White	44 (53.0)	-	-	-
Black	27 (32.5)	-	-	-
Filipino	1 (1.2)	-	-	-
Other Singular Racial Minority	4 (4.8)	-	-	-
Multiracial	7 (8.4)	-	-	-
Ethnicity:
White/Non-Hispanic or Latinx	78 (94.0)	-	-	-
Hispanic or Latinx	3 (3.6)	-	-	-
Missing	2 (2.4)	-	-	-
Sexual Orientation:
Gay	25 (30.1)	-	-	-
Lesbian	14 (16.9)	-	-	-
Queer	4 (4.8)	-	-	-
Straight	4 (4.8)	-	-	-
Questioning	2 (2.4)	-	-	-
Pansexual	8 (9.6)	-	-	-
Bisexual	6 (7.2)	-	-	-
Don’t Know	5 (6.0)	-	-	-
Other (e.g., Asexual)	12 (14.5)	-	-	-
Missing	1 (1.2)	-	-	-
Gender Identity:
Man	28 (33.7)	-	-	-
Woman	25 (30.1)	-	-	-
Transgender Man	8 (9.6)	-	-	-
Transgender Woman	15 (18.1)	-	-	-
Gender Queer (non-inclusive male or female)	2 (2.4)	-	-	-
Multiple	3 (3.6)	-	-	-
Missing	2 (2.4)	-	-	-
Military Branch:
Army	46 (55.4)	-	-	-
Air Force	11 (13.3)	-	-	-
Navy	16 (19.3)	-	-	-
Marines	7 (8.4)	-	-	-
Coast Guard	3 (3.6)	-	-	-

Notes: n = Subsample size; M = Mean; SD = Standard deviation; SE = Standard error; LGBT = Lesbian, gay, bisexual, and transgender

^a= PHQ-9 (Kroenke et al., 2001), score denotes sample average in moderate depression range

^b= GAD-7 (Spitzer et al., 2006), score denotes sample average is in mild anxiety range

^c= SBQ-R item 2 (ideation; 1–5 range), item 4 (attempt likelihood; range 0–6), and total score (score denotes sample average is above the clinically elevated risk cut-off)

^d= subscales and/or latent factor from modified LGBIS (subscale score range 1–6; Mohr & Kendra, 2011)

^e= subscales from LGB-PIM (subscale score range 1–7; Riggle et al., 2014).

Measures.

Demographics.

A form requested demographic information, including age, race, sexual orientation, gender identity, ethnicity, and military branch.

SGM Identity.

Participants completed the 27-item LGBIS (Mohr & Kendra, 2011) and the 25-item LGB-PIM (Riggle et al., 2014). The LGBIS originally contained the following 8 subscales, scored using item-level averages: acceptance concerns, concealment motivation, identity uncertainty, internalized prejudice, difficult process, identity superiority, identity affirmation, and identity centrality (Mohr & Kendra, 2017). Subsequent research (Cramer et al., 2017, 2018) identified a latent negative identity factor subsuming the following subscales: acceptance concerns, concealment motivation, difficult process, and internalized prejudice. All subscales, as well as the latent factor, were used in the present study. Because the context of this data collection was in a clinical setting at VA Medical Centers (Hilgeman et al., in press; Lange et al., 2020), survey instructions for the LGBIS were modified to retain focus on sexual orientation while communicating gender-focused inclusivity (i.e., “For those identifying as heterosexual and gender diverse, please respond to the items on this page only as you feel comfortable in how they may apply to you”). The LGB-PIM comprises the following five subscales, all of which were used in the present study: self-awareness, authenticity, community, intimacy, and social justice beliefs (Riggle et al., 2014).

Mental Health.

The Generalized Anxiety Disorder-7 (GAD-7; Spitzer et al., 2006) is a seven-item questionnaire assessing recent symptoms of anxiety (e.g., worrying). The GAD-7 provides a total score and cut-scores to demarcate symptom severity level (e.g., ≥10=moderate or worse symptoms). Internal consistency is consistently acceptable to high. The Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001) is a nine-item screener of depressive symptoms (e.g., depressed mood). The PHQ-9 provides a total score and cut-scores differentiating symptom severity levels (e.g., ≥10=moderate or worse symptoms). The total PHQ-9 score consistently has high reliability across samples. The Suicidal Behaviors Questionnaire-Revised (SBQ-R; Osman et al., 2001) is a four-item screener of lifetime suicidal behavior. Items can be used separately, and the SBQ-R provides a total score and clinical cut-score (i.e., ≥7=elevated suicide risk). All three measures possess strong psychometric properties in clinical and military samples (Cunningham et al., 2020; Franks et al., 2021; Kroenke et al., 2001; Osman et al., 2001; Spitzer et al., 2006).

Data Analysis.

Four persons from a total pilot sample (N = 87) were dropped due to missing data, leaving an analyzable sample of 83 SGM military veterans. Multiple imputation (Enders, 2017) was used to estimate missing items where entire scales were not missing. Specifically, regression-based imputation (10 imputations, maximum 2 case draws) was used with complete identity and mental health items as predictors to supplant missing values for these same measures.

To examine study aims, we used the following LGBIS and LGB-PIM subscales (see online supplement Figure 1a and 1b for a visual depiction). All eight original LGBIS subscales were used, namely: acceptance concerns, concealment motivation, identity uncertainty, internalized prejudice, difficult process, identity superiority, identity affirmation, and identity centrality (Mohr & Kendra, 2011). Additionally, we calculated a score for the negative identity latent factor identified by Cramer et al. (2017), which is comprised of the following subscales: acceptance concerns, concealment motivation, difficult process, and internalized prejudice. All LGB-PIM subscales were used to test study aims as well, namely: self-awareness, authenticity, community, intimacy, and social justice beliefs (Riggle et al., 2014). Aim 1 (LGBIS and LGB-PIM descriptives and internal consistency) was assessed with descriptive statistics, Cronbach’s α, and Pearson r intercorrelations. Aim 2 (explore associations with mental health indicators) was assessed with additional Pearson r values between identity subscales and mental health symptoms. Independent-samples t-tests with Cohen’s d effect sizes tested identity subscale variation in mental health risk categories using cut-scores.

Results

Missing data.

Missing data on LGBIS and LGB-PIM items ranged from 1.1–10.3%. Missing data on mental health items ranged from 4.6–10.3%.

Aim 1. LGBIS and LGB-PIM descriptive information and internal consistency.

Notable LGBIS descriptive findings included (see Table 1): (1) identity uncertainty, internalized prejudice, and identity superiority were below the scale mid-point, suggesting overall disagreement or low levels of these identity aspects; (2) identity affirmation and identity centrality were above the scale mid-point, indicating overall endorsement of positive identity aspects; and (3) negative identity, acceptance concerns, concealment motivation, and difficult process approximated the scale mid-point, a pattern indicating overall middle ground views of one’s identity. LGB-PIM subscales of self-awareness, authenticity, and social justice beliefs were well above scale mid-point, suggesting overall positive views of one’s identity. Community involvement and intimacy more closely approximated the scale mid-point, a pattern indicating neither low nor high levels of agreement.

LGBIS reliability values varied. Half of the subscales (i.e., acceptance concerns, concealment motivation, internalized prejudice, and identity affirmation; α range = .70 to .87), and the underlying negative identity factor (α = .81), possessed acceptable internal consistency (see Table 2). Identity uncertainty, identity superiority, and identity centrality demonstrated marginal internal consistency (α range = .60 to .69), and difficult process (α = .51) reliability was poor. All LGB-PIM subscales demonstrated acceptable to high internal consistency (α range = .82 to .92).

Table 2.

Sexual and Gender Identity Measure Internal Consistency, and Correlations with Mental Health and Coping Beliefs

Var.	Dep	Anx	SI	Sui	SA	NI	AC	CM	IU	IP	DP	IS	IA	IC	SA	AU	CI	IT	SJ
Dep	-	.83	.36	.43	.40	.03	.21	.00	.33	−.01	−.15	.04	.03	−.00	.08	−.01	.04	−.04	.24
Anx		-	.41	.53	.50	.02	.29	−.06	.26	−.01	−.17	.02	.04	−.00	.07	−.03	.05	.03	.25
SI			-	.81	.64	.06	.12	.16	.19	−.04	−.06	−.04	−.03	−.03	.08	−.05	−.02	−.15	.11
Sui				-	.88	.03	.18	.02	.22	−.13	.01	.01	.10	−.04	.10	.03	.07	−.11	.25
SA					-	−.05	.10	−.11	.16	−.13	−.02	.04	.12	−.11	.07	.09	.11	−.07	.25
NI						-	.78	.63	.58	.72	.65	.24	−.49	−.09	.06	−.55	−.26	−.32	−.33
AC							-	.31	.43	.43	.43	.25	−.31	.00	.18	−.45	−.31	−.18	−.07
CM								-	.39	.28	.21	.06	−.18	−.05	.05	−.17	−.16	−.26	−.34
IU									-	.51	.26	.37	−.34	−.17	.07	−.33	−.12	−.38	−.16
IP										-	.23	.15	−.74	−.27	−.22	−.65	−.26	−.42	−.46
DP											-	.21	−.06	.11	.19	−.23	.03	.00	−.01
IS												-	.04	.13	.27	−.06	−.07	.13	.06
IA													-	.53	.44	.73	.42	.48	.56
IC														-	.50	.34	.24	.39	.36
SA															-	.30	.37	.35	.32
AU																-	.38	.41	.42
CI																	-	.37	.35
IT																		-	.28
SJ																			-
α	.88	.91	-	.74	-	.81	.82	.70	.68	.87	.51	.66	.87	.61	.90	.89	.92	.92	.82

Notes: Bold font denotes p < .05; Dep = Depression; Anxiety = Anxiety; SI = 12-month suicidal ideation; Sui = Lifetime suicidal behavior; SA = Future likelihood of suicide attempt; NI = LGBIS negative identity latent factor; AC = LGBIS acceptance concerns; CM = LGBIS Concealment motivation; IU = LGBIS Identity uncertainty; IP = LGBIS internalized prejudice; DP =LGBIS difficult process; IS = LGBIS identity superiority; IA = LGBIS identity affirmation; IC = LGBIS identity centrality; SA = LGB-PIM self-awareness; AU = LGB-PIM authenticity; CI = LGB-PIM LGBT community involvement; IT = LGB-PIM intimacy; SJ = LGB-PIM social justice beliefs; LGBIS = Lesbian, Gay and Bisexual Identity Scale (Mohr & Kendra, 2011); LGB-PIM = Lesbian, Gay and Bisexual Positive Identity Measure (Riggle et al., 2014).

Aim 2. Explore associations between identity subscales and mental health indicators.

First, with few exceptions, LGBIS negative identity subscales were positively correlated with one another (small-to-moderate range; see Table 2). Second, identity centrality was mostly unrelated to LGBIS negative identity subscales (e.g., the negative identity higher-order factor, acceptance concerns, concealment motivation). Third, identity superiority was positively associated with negative identity, acceptance concerns, and identity uncertainty (small-to-moderate sizes). Yet, identity superiority was also positively associated with self-awareness. LGB-PIM subscales were all positively correlated with one another (small-to-large range). About half of the correlations between LGB-PIM and LGBIS negative identity subscales were negative (small-to-moderate effects), whereas identity centrality and identity affirmation were consistently positively associated with LGB-PIM subscales (small-to-large effects).

Regarding mental health, identity uncertainty was positively correlated with depression and anxiety (see Table 2). Acceptance concerns were positively correlated with anxiety. Social justice beliefs displayed positive associations with depression, anxiety, lifetime suicidal behavior, and future suicide attempt likelihood. However, all significant identity-mental health correlations were small in terms of effect size.

Table 3 contains statistics comparing identity subscales by levels of depressive symptoms, anxiety symptoms, and suicide risk level. Acceptance concerns differentiated levels of anxiety (moderate effect), such that SGM veterans reported higher acceptance concerns in the moderate or higher anxiety symptom category. Identity uncertainty differentiated all three mental health outcomes (moderate effects), such that SGM veterans reported higher identity uncertainty in the worse mental health symptom categories (i.e., moderate or worse depressive symptoms, moderate or worse anxiety symptoms, and high suicide risk). Social justice beliefs differentiated all three mental health outcomes (small-to-moderate effects), such that SGM veterans reported higher social justice beliefs in the worse mental health symptom categories.^I

Table 3.

Sexual Identity Measure Subscale Scores by Mental Health Symptom Category

	Depressive Symptoms				Anxietv Symptoms				Suicide Risk Level
	T(df)a	Low	Mod or higher	d	T(df)b	Low	Mod or higher	d	T(df)c	Low Risk	High Risk	d
LGBIS:
NI	−0.01 (74)	3.28 (0.90)	3.28 (1.00)	0.00	0.01 (73)	3.27 (1.02)	3.28 (0.86)	0.01	0.26 (81)	3.24 (1.04)	3.30 (0.91)	0.06
AC	1.72 (74)	3.35 (1.32)	3.90 (1.44)	0.40	2.56 (73)	3.28 (1.33)	4.09 (1.40)	0.60	0.64 (81)	3.50 (1.45)	3.71 (1.39)	0.16
CM	−0.11 (74)	3.82 (1.33)	3.79 (1.30)	0.02	−0.72 (73)	3.90 (1.34)	3.68 (1.28)	−0.17	0.19 (81)	3.78 (1.45)	3.84 (1.21)	0.04
IU	2.62 (74)	1.60 (0.70)	2.17 (1.15)	0.59	2.32 (73)	1.67 (0.78)	2.19 (1.18)	0.54	2.73 (81)	1.50 (0.76)	2.10 (1.02)	0.64
IP	−0.57 (74)	2.38 (1.47)	2.20 (1.56)	0.12	−0.88 (73)	2.42 (1.47)	2.10 (1.59)	−0.21	−0.17 (81)	2.31 (1.55)	2.25 (1.50)	−0.04
DP	−1.09 (74)	3.57 (1.30)	3.25 (1.25)	0.25	−0.88 (73)	3.50 (1.32)	3.24 (1.24)	−0.20	0.09 (81)	3.38 (1.17)	3.41 (1.29)	0.02
IS	0.14 (74)	1.54 (0.60)	1.56 (0.79)	0.03	0.94 (73)	1.48 (0.61)	1.63 (0.82)	0.21	0.44 (81)	1.53 (0.60)	1.59 (0.73)	0.09
IA	0.38 (74)	4.43 (1.29)	4.54 (1.36)	0.08	0.91 (73)	4.36 (1.27)	4.64 (1.40)	0.21	0.90 (81)	4.37 (1.34)	4.64 (1.29)	0.20
IC	−0.25 (74)	3.90 (1.08)	3.84 (1.05)	−0.06	0.51 (73)	3.83 (1.17)	3.96 (0.91)	0.12	−0.21 (81)	3.95 (126)	3.90 (0.90)	−0.04
LGB-PIM:
SA	0.13 (74)	5.13 (1.12)	5.17 (1.42)	0.03	1.24 (73)	5.01 (1.34)	5.38 (1.16)	0.29	0.33 (81)	5.15 (1.42)	5.25 (1.31)	0.07
AU	0.23 (74)	5.42 (126)	5.50 (1.57)	0.06	0.59 (73)	5.39 (1.29)	5.59 (1.59)	0.14	0.54 (81)	5.38 (1.45)	5.56 (1.39)	0.13
CI	−1.08 (74)	4.38 (1.49)	4.01 (1.52)	−0.24	0.31 (73)	4.14 (1.49)	4.26 (1.57)	0.08	1.32 (81)	3.95 (162)	4.40 (1.38)	0.30
IT	−0.99 (74)	4.85 (1.44)	4.48 (186)	−0.22	0.91 (73)	4.51 (1.58)	4.86 (1.80)	0.21	−0.44 (81)	4.83 (1.52)	4.66 (1.72)	−0.10
SJ	1.98 (74)	5.00 (1.45)	5.57 (1.03)	0.45	2.01 (73)	5.04 (1.44)	5.62 (0.95)	0.46	2.01 (81)	4.91 (1.63)	5.53 (1.02)	0.49

Notes: Bold font p ≤ .05

^a:Low n=37, moderate or higher symptoms n=39

^b:Low n=43, moderate or higher symptoms n=32

^c:risk n=28, elevated risk n=55; NI = LGBIS negative identity latent factor; AC = LGBIS acceptance concerns; CM = LGBIS concealment motivation; IU = LGBIS identity uncertainty; IP = LGBIS internalized prejudice; DP = LGBIS difficult process; IS = LGBIS identity superiority; IA = LGBIS identity affirmation; IC = LGBIS identity centrality; SA = LGB-PIM self-awareness; AU = LGB-PIM authenticity; CI = LGB-PIM LGBT community involvement; IT = LGB-PIM intimacy; SJ = LGB-PIM social justice beliefs; LGBIS = Lesbian, Gay and Bisexual Identity Scale (Mohr & Kendra, 2011); LGB-PIM = Lesbian, Gay and Bisexual Positive Identity Measure (Riggle et al., 2014).

Discussion

The first aim of the present pilot investigation was to assess descriptive patterns and internal consistencies of the LGBIS (adapted for gender-inclusive instructions; Mohr & Kendra, 2011) and LGB-PIM (Riggle et al., 2014). The sample was characterized by primarily scale mid-point level negative identity, with the exceptions of low internalized prejudice and identity uncertainty. Identity superiority was also below the mid-point, whereas positive identity elements of both measures were consistently above scale mid-points. These patterns dovetail well with prior studies in several ways. For example, internalized prejudice and identity uncertainty have consistently been below scale mid-points across samples (Cramer et al., 2017, 2018; Mohr & Kendra, 2011). However, several negative aspects of identity (e.g., concealment motivation) were higher in the present sample than prior studies, although not concerningly above the scale mid-point. Higher levels of identity concealment may be, in part, a function of navigating one’s identity within the stigmatizing culture of the military (McNamara et al., 2020) or related to completing these measures in a VA treatment setting. Further, consistent with prior scale validation studies (Mohr & Kendra, 2011; Riggle et al., 2014), almost all aspects of positive identity were above scale mid-points.

In line with sexual minority adult validation data (e.g., Cramer et al. 2017), several LGBIS subscales (identity uncertainty, difficult process, identity superiority) demonstrated poor internal consistency. Several explanations exist for these low LGBIS subscale internal consistency values. First, the relatively small sample size may have affected internal consistency. Statistically speaking, the low number of items on many subscales may result in low internal consistency (Tavakol & Dennick, 2011). Alternatively, the mere variety of SGM identities reflected in the sample may explicate poor internal consistency values. A growing body of literature (e.g., Van et al., 2018; Van Gilder, 2019; Williams et al., 2014) shows that the lived experiences of SGM persons with differing identities (e.g., gay vs. bisexual vs. pansexual) vary and may uniquely contribute to, or be affected by, military culture, mental health, and discrimination. That is, varied life experience may contribute to differential interpretation of and response to identity measure items. The underlying negative identity factor, however, demonstrated good internal consistency, offering further support for a global underlying identity dimension (Cramer et al., 2017). Mirroring the original LGB-PIM development (Riggle et al. 2014), all subscales demonstrated high internal consistency.

Identity measure subscale inter-correlations also lend insight into the possible nature of SGM veteran identity by illuminating one-to-one associations of adaptive and maladaptive aspects of one’s self-concept as reflect within and across the LGBIS and LGB-PIM. For instance, while LGBIS intercorrelation patterns were largely consistent with expected directions based on literature, identity superiority was positively correlated with negative identity subscales. This finding is consistent with non-clinical and non-military findings (Cramer et al., 2017, 2018), and contributes to ongoing question about the nature of identity superiority, as it has demonstrated psychometric qualities consistent with aspects of both adaptive and maladaptive aspects of identity (Cramer et al. 2017). Second, the underlying negative identity factor, as well as the internalized prejudice score, were largely negatively correlated with LGB-PIM subscales. Consistent with minority stress and resilience perspectives (e.g., Meyer, 2003; Riggle et al. 2014), this pattern may suggest that positive identity, as reflected in the LGB-PIM, may be negatively influenced by minority stress. Given the nature of these clinical pilot findings, future work may seek to test prospective pathways between aspects of identity. Finally, identity centrality and identity affirmation (LGBIS subscales) displayed small positive correlations with LGB-PIM subscales. These associations suggest the LGBIS may not be a comprehensive measure of positive identity, requiring additional consideration of centrality and affirmation as well. While a valuable start to understanding how specific pairs of subscales are related, the present study did not formally test the LGBIS or LGB-PIM factor structure model fit. Future research should employ confirmatory factor analysis in larger samples of SGM veterans in order to understand how well entire subscale factor structures hold together and to increase confidence of measurement recommendations to providers in the field.

Observed variation in LGBIS subscale descriptive statistics, internal consistency, and inter-correlation patterns informs clinical practice and research for veteran and non-veteran SGM persons. Given the mounting evidence of poor internal consistency of certain subscales, especially identity superiority, the field utility of the entire LGBIS in program evaluation or other quality improvement work may be limited. Particular subscales such as concealment motivation or internalized prejudice, as well the higher-order negative identity factor, may be of most utility with SGM adults in applied settings. A potential next step for researchers may be to examine LGBIS and LGB-PIM measurement invariance and internal consistency across sexual orientation, gender identity, racial, military, and other identities. This may lend insight into the extent to which LGBIS subscales in particular may or may not possess acceptable psychometric properties for certain groups (but not others).

The second aim was to evaluate SGM veteran identity subscale scores with mental health. Consistent with prior literature (Cramer et al., 2017, 2018; Denton et al. 2014) and minority stress theory (Meyer, 2003), identity uncertainty and acceptance concerns were associated, albeit modestly so, with worse mental health. Social justice beliefs were positively associated with depression and anxiety. This pattern is contrary to the LGB-PIM development study (Riggle et al., 2014), but consistent with human rights literature. Historically, SGM individuals acting as social activists have endured extra strains from the responsibility of having to strengthen personal knowledge about political and legal matters in order to advocate for civil rights (Knauer, 2012). Moreover, campaigning for reform and inclusivity can contribute to activist burnout among professionals (Chen & Gorski, 2015). Many SGM individuals find purpose in advocating for social justice, while others do so out of necessity. The positive association between social justice beliefs and mental health symptoms (i.e., depression and anxiety) may be an extension of burnout, suggesting that the onus of a social justice orientation may be burdensome to members of a marginalized group. Within large systems such as VA, our findings may suggest it is paramount to prioritize development of affirmative care services to proactively address both identity-related resilience and mental health symptoms that could buffer SGM individuals from shouldering primary responsibility for advocating for health equity.

Enhancing the novelty and importance of the present pilot data, several identity subscales showed potential for identifying clinically relevant mental health groups. Identity uncertainty and social justice may differentiate symptoms levels of depression, anxiety, and suicide risk, whereas acceptance concerns may differentiate anxiety symptom severity. Such aspects of SGM identity have been hypothesized to explain SGM mental health (Moleiro & Pinto, 2015). Components of SGM veteran identity may possess relevance in clinical screening and treatment as determinants of, or moderating influences on, mental health outcomes. However, especially given the multitude of analyses conducted and risk of Type I error, additional work is needed to assess potential clinical utility of SGM identity measures.

This pilot is not without limitations. First, while the heterogeneity of the groups was by design clinically (Lange et al., 2020), we were unable to rule in or out differences in SGM identity subgroups as a source of variation in findings. Second, the LGBIS was originally developed for use only with sexual minority persons. The LGBIS instructions were altered to be gender inclusive due to 1) the goal of the VA Pride in All Who Served program (Lange et al., 2020) to serve both sexual and gender minority persons and 2) time constraints in the clinical setting. This deviation from standard administration introduces a source of potential error that may have impacted findings. Further work is needed to examine the possible impact of gender-inclusive instructions on the LGBIS. These limitations are offered in the context that excluding the four heterosexual persons from analyses did not impact findings.

Statistically, the small sample size limits statistical power, particularly for Aim 2 analyses. As such, hypothesis testing (e.g., t tests) were employed to identify promising targets for future research Also, we opted not to implement a Bonferroni or other statistical correction, leaving open the possibility of elevated Type I error. We elected this approach because (a) the statistical literature is mixed on the impact of Bonferroni corrections on Type I error (e.g., Feise, 2002; Perneger, 1998); (b) use of such corrections exacerbates Type II error in small sample studies with multiple comparisons and may worsen publication bias (e.g., Nakagawa, 2004; VanderWeele & Mathur, 2019), and (c) SGM military veterans represent a marginalized minority group difficult to enroll in research – as such, we did not want this small sample pilot study of marginalized persons to suffer a heightened risk of failing to detect potential associations. Findings should be viewed as preliminary, however, given the statistical and sample size limitations. Despite limitations, this article is the first to provide initial information on two commonly used SGM identity measures among treatment-engaged SGM veterans in VA healthcare settings. Although low internal consistency for several LGBIS subscales was observed and a small sample limited the scope of analyses, both scales warrant additional examination. Low internal consistency for three LGBIS subscales limits confidence in the aim 2 associations with mental health. As such, observed associations with mental health should be viewed with as preliminary, especially with regard to identity superiority, identity uncertainty, and difficult process. Our findings concerning acceptable reliability and expected associations with mental health for LGB-PIM and most LGBIS subscales point to opportunities for replication and extension in the future.

In conclusion, these pilot data illuminate promising next steps regarding inquiry into SGM identity related to military programming and clinical practice. First, the LGB-PIM in particular can be applied in larger clinical military (active duty and veteran) samples. Doing so may provide more detailed testing, including prospective associations with mental health over time. Second, the LGB-PIM may be useful in tracking the impact of affirmative programs and development of quality improvement. Given the VA has a history of stigma for SGM veterans (e.g., Mattocks et al., 2015), the LGB-PIM, as well as specific LGBIS subscales, may offer new tools to clinicians and SGM care coordinators working with veterans who may benefit from innovative programs like the Pride in All Who Served group (e.g., Lange et al., 2020, Hilgeman et al., in press) to address internalized stigma and other negative outcomes. The administration of these measures may also function as a sign of inclusivity and affirmation for SGM veterans. Finally, future work should expand focus on LGBIS and LGB-PIM measurement invariance, as well as the assessment of social justice, identity uncertainty, and concealment motivation for possible clinical utility in mental health screening and treatment.

Public Policy Relevance:

The ability to track the impact of healthcare services on patient-reported outcomes is critical for clinical providers when communicating treatment goals and determining when to alter their treatment approach. The ability to demonstrate the impact of an affirmative care program can accelerate its adoption and support within a measurement-based care environment like the VA. This paper provides initial information on two commonly used SGM identity measures offering preliminary support for their use in future SGM military veteran program development.