Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 6.
Published in final edited form as: J Career Assess. 2018 Jul 15;27(3):544–559. doi: 10.1177/1069072718788324

Reconsidering the LGBT Climate Inventory: Understanding Support and Hostility for LGBTQ Employees in the Workplace

Elizabeth Grace Holman 1, Jessica N Fish 2,3, Ramona Faith Oswald 4, Abbie Goldberg 5
PMCID: PMC8100868  NIHMSID: NIHMS1696660  PMID: 33967571

Abstract

Workplace climate matters significantly for lesbian, gay, bisexual, queer, or other sexual minority (LGBQ) employees, given that the presence of workplace hostility or support can affect well-being. The Lesbian, Gay, Bisexual, and Transgender Climate Inventory (LGBTCI) is a measure of workplace climate for LGBQ individuals, intended to capture the full range of workplace climate from hostility to support. The purpose of this article is to provide evidence that the recommended scoring approach of the LGBTCI needs to be reconsidered. We used latent class analysis to estimate classes of work-related experiences in our sample of 442 LGBQ employees who completed the LGBTCI. A four-class solution fit the data best. Characteristics of each class were identified and consequently labeled: supportive work climate, tolerant work climate, ambiguous work climate, and hostile work climate. Findings suggest that a more accurate measure of workplace climate would include independent scales for support and hostility.

Keywords: LGBTQ, hostility, latent class analysis, minority stress theory, support, workplace climate

“Workplace climate matters” (Liddle, Luzzo, Hauenstein, & Schuck, 2004, p. 33). This statement is particularly true for employees who identify as lesbian, gay, bisexual, queer, or otherwise a sexual minority (LGBQ), a population which is significantly affected by the presence of support and hostility perceived in their work environment (Velez, Moradi, & Brewster, 2013). Thus, the development of the Lesbian, Gay, Bisexual, and Transgender Climate Inventory (LGBTCI; Liddle et al., 2004), a scale to measure workplace climate for LGBQ individuals, was necessary to help document the unique experiences of LGBQ employees in the workplace. Given that previous measures of LGBQ workplace experiences only focused on the negative aspects of climate, the major goal of the LGBTCI was to capture the full range of workplace climate including support. The authors who developed and validated the LGBTCI, however, conceptualized support and hostility as two ends of a single continuum (Liddle et al., 2004). Consequently, the recommended approaches to scoring the LGBTCI are arguably flawed. The purpose of this article is to provide evidence that the recommended scoring approaches of the LGBTCI may mask workplace experiences for LGBQ employees and, specifically, that support and hostility may in fact be co-occurring in the workplace. Our findings suggest that the scoring procedures for the LGBTCI need to be reconsidered.

LGBQ Employee Experiences

Many LGBQ employees in the United States lack workplace protections. There is currently no federal policy protecting LGBQ employees from discrimination. Although Title VII has been interpreted to protect sexual minority employees in the past, the Department of Justice recently rescinded these guidelines (Ruggiero & Park, 2017). Further, a proposed Employment Non-Discrimination Act, which would provide employment protections for LGBQ employees, has failed to pass into law since the early 1990s (Pizer, Seats, Mallory, & Hunter, 2012; Wendland, 2007). Over 25 states lack statewide protections (Movement Advancement Project, 2017), leaving smaller scale protective decisions in the hands of individual counties and employers. Yet discrimination may still occur in work environments, even with protections in place. Thus, LGBQ employees’ experiences may vary greatly in the workplace depending on the general climate of their work environment.

Climate, as it relates to LGBQ populations, has been described as “the overall level of support or hostility toward LGBQ people that is present [in a given context]” (Holman, 2016, p. 252). That is, climate describes the broader, institutional-level sense of acceptance or stigma. In residential communities, climate is typically assessed using general attitudes and beliefs about LGBQ people in a region, laws, and policies in place that support or discriminate against LGBQ people, political voting records, and religious presence (Oswald, Cuthbertson, Lazarevic, & Goldberg, 2010). Studies have similarly aimed to measure the general levels of hostility or support in a given organizational context to describe workplace climate.

Hostility

A growing body of literature documents LGBQ employees’ experiences in hostile work environments. Demeaning attitudes toward gender and sexual minorities, derogatory comments and jokes, as well as verbal and physical abuse leave LGBQ individuals feeling victimized in the workplace (Badgett, Lau, Sears, & Ho, 2007; Embrick, Walther, & Wickens, 2007; Herek, 2009; Ragins, Singh, & Cornwell, 2007; Velez et al., 2013). In their 2007 study, Badgett, Lau, Sears, and Ho (2007) found that up to 40% of LGBQ people surveyed reported some form of abuse or harassment at work related to their gender identity, expression, or sexual orientation. Prior studies also found that these types of discrimination occur frequently in the workplace (Button, 2001; Ragins & Cornwell, 2001).

Notably, hostile work environments have implications for mental and physical health and have been linked to negative outcomes among LGBQ employees. Specifically, exposure to hostility and prejudice at work is associated with higher rates of depression and anxiety (Griffith & Hebl, 2002) and increased psychological distress (Velez et al., 2013) among LGBQ employees. LGBQ employees in hostile work environments also report taking a greater number of sick days from work (Huebner & Davis, 2007), lower job satisfaction, and a stronger intention to leave their position with a given employer (Ragins et al., 2007; Velez et al., 2013).

Support

Several scholars have examined experiences of support in the workplace for LGBQ employees. A supportive workplace often embraces practices such as an organization-wide policy prohibiting discrimination based on sexual orientation, inclusion of sexual orientation in company diversity statements or diversity trainings, sexual minority resource-support groups, public support of LGBQ issues, and a general sense of acceptance such that same-sex partners are welcome at company social activities (Button, 2001; Ragins & Cornwell, 2001). Importantly, greater levels of perceived support are associated with more job satisfaction (Griffith & Hebl, 2002) and a greater likelihood of sexual identity disclosure in the work context (Huffman, Watrous-Rodriguez, & King, 2008; Ragins et al., 2007).

Assessing Workplace Climate for LGBQ Employees

The majority of research in this area has only focused on the effects of hostile work environments; there are fewer empirical studies testing associations with supportive workplaces. In an effort to study workplace climate, scholars have seemingly ignored the role that workplace support plays in the overall climate. Interestingly, Liddle and colleagues (2004) note that the LGBTCI was developed to intentionally assess more than just negative experiences in the workplace, saying,

The LGBTCI… fills an important gap in the vocational literature by providing a carefully constructed instrument to measure workplace climate for LGBT employees. Unlike previous attempts to measure LGBT climate, this instrument assesses the full range of experience from very negative to very positive rather than focusing exclusively on harassment and discrimination. (p. 46)

Yet the LGBTCI, as it is currently scored, falls short of its potential to assess the more nuanced aspects of workplace climate.

Construction and Validation of the LGBTCI

The original LGBTCI was inductively developed by Liddle and colleagues (2004) through a set of open-ended surveys sent to LGBQ employees across nine U.S. states, asking participants to provide descriptions of their workplace experiences and climate. Respondents were asked to “Please write short phrases or sentences on the following lines describing what it’s like to be a gay, lesbian, or bisexual employee at your current workplace” (Liddle et al., 2004, p. 37). Phenomenological analysis (Giorgi, 1985) of the 39 surveys returned resulted in 60 items, 33 describing workplace support and 27 describing workplace hostility toward LGBQ employees. This preliminary measure was then sent to 173 sexual minority employees to examine the scale reliability and validity. Subsequent member checking and factor analysis of the 124 surveys returned reduced the measure to 20 items— “12 positively worded and 8 reverse scored” (Liddle et al., 2004, p. 40). Example items include LGBT employees are free to be themselves and employees are expected to not act “too gay” (Liddle et al., 2004, p. 47).

After item pairing for clarity and redundancies, the resulting 20-item scale indicated adequate reliability (α = .96, n = 119). Liddle, Luzzo, Hauenstein, and Schuck (2004) then assessed the factor structure of the measure using principal component factor analysis. Findings indicated a two-structure factor model using the eigenvalue criteria of 1.0, but the prominence of the eigenvalue decline from Factor 1 to Factor 2 also suggested that a single-factor solution may have been appropriate. Subsequently, Liddle and colleagues examined the interpretability of the two-factor solution and found that 13 of the 20 items indicated relatively indistinguishable membership across factors, but that the cross-loading patterns and those items reflecting the presence or absence of negative or hostile work experiences loaded strongly onto Factor 2. This, in combination with screeplot data and high internal consistency, led the authors to favor a single-factor solution. A final pilot study was conducted (n = 93) to confirm internal reliability and construct validity, but not the factor structure of the measure.

Based on their measurement findings, Liddle et al. (2004) instructed users of the LGBTCI to reverse score the 8 negatively worded items before sum scoring all 20 items (Liddle et al., 2004). Participant scores thus range on a 60-point scale, with lower scores indicating a more hostile work climate and higher scores indicating a more positive workplace climate. Scoring data from the LGBTCI in this way implies that hostility is antithetical to support and, thus, represent two ends of a single continuum when assessing workplace climate for LGBQ employees.

Use of the LGBTCI

Nearly a dozen dissertations and published articles in the past decade have specifically used the LGBTCI (Liddle et al., 2004) to measure workplace climate. Scholars have primarily used the LGBTCI as a predictor variable to link perceptions of workplace climate to employee disclosure practices and job satisfaction. Results from several studies show that higher levels of workplace support—as measured by the LGBTCI—are related to greater disclosure of a sexual minority identity in that context (Huffman et al., 2008; Köllen, 2013). Conversely, sexual minority employees who reported a less supportive workplace climate on the LGBTCI indicated more pressure to conceal their identity and use of specific identity management strategies to avoid disclosing their sexual orientation (Boyles, 2008; Köllen, 2013, Reed & Leuty, 2015). A supportive workplace climate has also been linked to higher levels of job satisfaction (Ricci-Stiles, 2007; Velez & Moradi, 2012), a relationship that may be moderated by anticipated discrimination (Prati & Pietrantoni, 2014). That is, sexual minorities who perceive less support in their work environment may lead them to expect discriminatory experiences and thus decrease their level of satisfaction with their employment.

To date, only one study appears to have used the LGBTCI to analyze workplace climate as an outcome variable rather than the predictor. Köllen (2015) examined organization qualities and tactics that were associated with a more supportive work environment. Results showed that sexual minority employees rated their workplace climate as more supportive when the organization aimed to minimize differences between heterosexual and nonheterosexual employee experiences, for example, treating same-sex and different-sex relationships the same. Specific LGBQ marketing, both within and external to the company, was also correlated with higher levels of supportive workplace climate as measured with the LGBTCI.

In all of these studies, researchers scored the LGBTCI as recommended. Only Brewster, Velez, DeBlaere, and Moradi (2012) altered the measure in any way. First, the items were reworded to replace LGBT with transgender in an attempt to generalize the scale’s applicability to the transgender community. The resulting transgender form, the LGBTCI-TF, was presented to 263 transgender participants (86% White; mean age 38.28 years).

Analysis of the measure’s structure, however, supported a possible multidimensional approach to scoring, based on evidence that model fit was poor when imposing a one-factor structure. To address this, Brewster et al. (2012) modeled two orthogonal factors to account for the additional variance associated with the valence of the item wording: one for positive items, another for negative items. When accounting for item directionality, Brewster et al. noted a significant improvement in model fit for the single-factor structure, which suggests that there may be systematic differences in the scoring of positive and negative items that may be washed out in Liddle et al.’s (2004) recommended scoring procedures.

Analysis of the underlying structure of the LGBTCI-TF (Brewster, Velez, DeBlaere, & Moradi, 2012) calls into question the ways in which scholars have been conceptualizing and measuring support and hostility in the workplace to date. Brewster and colleagues suggest that perhaps two orthogonal factors are represented in the presumed unidimensional assessment, thus bringing to light the question of how support and hostility may be related. Are they, in fact, opposing ends of the same continuum—or are they distinct concepts, which together more fully describe the overall workplace climate?

Understanding Support and Hostility

To answer this question, we rely on minority stress theory (Meyer, 2003) to explain the relationship between support and hostility. Minority stress theory posits that sexual minorities experience additional stressors specific to a stigmatized identity, above and beyond general day-to-day stress processes (Meyer, 2003). These stressors can include experiences with direct anti-LGBQ discrimination and perceived hostility toward the LGBQ community. These unique stressors, and the potentially maladaptive coping mechanisms they warrant, can have deleterious effects on individuals’ mental, behavioral, and physical health (Goldbach, Tanner-Smith, Bagwell, & Dunlap, 2013; Lick, Durso, & Johnson, 2013; Meyer, 1995). As stated earlier, this is true for LGBQ employees in the work context as well (e.g., Ragins et al., 2007; Velez et al., 2013).

Importantly, this theory also accounts for supports an individual may receive or perceive, which can mitigate such negative effects. Meyer (2007) conceptualizes these coping mechanisms and social supports as “stress-ameliorating factors” (p. 255), meaning the aspects of the individual and their environment that make it easier to cope with discrimination, prejudice, and hostility are distinct from these stressors. Meyer asserts that stressors and protective factors can co-occur and thus may have an interactive effect on LGBQ individual’s mental health. For instance, Griffith and Hebl (2002) found that social supports from coworkers were a vital component of reported job satisfaction among sexual minority employees. In the face of hostility, positive reactions from colleagues following sexual orientation disclosure minimized negative outcomes. Thus, minority stress theory does not position hostility and support as opposing factors but as two separate constructs, each of which has a potential impact on the lived experience and health of LGBQ people and thus should be measured as such.

The Current Study

Valid measurement is a key to conducting research that can be used to improve the lives of LGBQ people. Although the LGBTCI (Liddle et al., 2004) is an empirically informed and reliable scale used to document workplace climate for LGBQ employees (e.g., Goldberg & Smith, 2013; Reed & Leuty, 2015), the recommended scoring method may not accurately capture experiences of support and hostility in the workplace for LGBQ people. Theoretically, reverse scoring hostility items unduly operationalize the absence of workplace hostility as the presence of workplace support. This assumption does not align with the original tenets of minority stress theory (Meyer, 2003) and may obscure employees’ negative workplace perceptions. Instead, we argue that workplace support and hostility may coexist and are neither theoretically nor in reality mutually exclusive.

In the current study, we interrogate the utility of the current scoring procedures of the LGBTCI. Specifically, we argue that the two subscales of the LGBTCI—Support and Hostility—do not reflect two ends of a single continuum but are distinct aspects of workplace climate that may coexist. To test this assertion, we estimated classes or profiles of participants’ work-related experiences across the 20 LGBTCI items using latent class analysis (LCA). LCA is a particularly useful statistical approach to test this assumption because it estimates patterns of responses across several variables at once, which illuminate otherwise unobservable subgroups within a population. Although traditional measurement modeling will help to establish whether items adequately capture (i.e., measure) an underlying latent construct, LCA provides a unique perspective of the data that allows researchers to uncover otherwise unobservable and qualitatively distinct subgroups within a population.

Therefore, we use LCA to help determine whether people view supportive and hostile experiences at work as a single continuum (the present of one is the absence of the other) or whether workplace climate may actually reflect a more complex experience where employees perceive both support and hostility, or a lack thereof, in their day-to-day experiences. If the latter is true, then the current scoring procedures of LGBTCI may oversimplify or simply miss the lived experiences of LGBQ people in the workplace. Subsequently, studies that use the LGBTCI may draw inaccurate conclusions about the factors associated with workplace climate for LGBQ people.

Method

Data Sources and Sample

To increase our sample size and sociodemographic diversity, we pooled data from two separate studies (Schenker & Raghunathan, 2007): Rainbow Illinois and the Transition to Adoptive Parent-hood Project(TAPP). The combination of these two smaller samples affords us more statistical power to detect potential subgroup differences in workplace climate that would be difficult to detect in smaller samples. Furthermore, the addition of TAPP increased geographic variability as Rainbow Illinois was state-specific. Because each data set measured sociodemographic characteristics with the distinct items, we are unable to statistically test sociodemographic differences across samples. We do, however, note that participants in the two samples had mean-level differences on the LGBTCI, t = −5.49, p < .001 (Rainbow Illinois, M = 2.20, SD = .60; TAPP, M = 2.56, SD = .48).

Subsample 1.

The first subsample for this study comes from Rainbow Illinois, a larger project which surveyed sexual minorities age 17 and older across the nonmetropolitan region of Illinois (Oswald & Holman, 2013). Individuals were recruited between November 30, 2010, and May 31, 2011, through a variety of snowball and purposive sampling techniques. A web link for the online survey was widely distributed throughout the area using social networks, advertisements at LGBQ-related functions, and mailing lists for LGBQ organizations. Survey participation was anonymous, although participants could choose to enter their e-mail address into a separate form at the completion of the survey for a chance to win a US$25 gift card.

Given the goal of the current study, we include only those participants who reported working full time or part time and who were not missing on all 20 LGBTCI items (n = 343; Mage = 37.79, SD = 12.88). Rainbow Illinois participants were 62.97% female, 34.40% male, and 2.62% queer or androgynous. Regarding race/ethnicity, 94.17% were White, 3.59% Latino/Latina, 2.93% Asian/Pacific Islander, 2.61% Black, and 0.98% American Indian/Alaska Native. The majority were employed full-time employment (68.51%) and 31.49% indicated part-time employment. The modal annual personal income was less than US$10,000 (19.16%), though the sample average was between US$30,001 and US$40,000. Modal household income was above US$100,000 (21.90%) with an average between US$50,001 and US$60,000. Most participants indicated that they were lesbian (42.27%), followed by gay (31.49%), bisexual (15.74%), and queer/pansexual (10.50%).

Subsample 2.

Our second subsample comes from the TAPP, a study of first-time parents in the process of adopting a child. Couples were recruited between 2005 and 2009 from adoption agencies throughout the United States, who provided study information to couples who were actively seeking to adopt their first child, but who had not yet adopted. States with a high percentage of same-sex couples (Gates & Ost, 2004) were purposively sampled in an attempt to have a balanced sample that was more heterogeneous with regard to geography and income. Researchers also partnered with LGBQ organizations to help recruit same-sex couples who were interested in adoption, but not out to adoption agencies. Once enrolled, participants completed individual telephone interviews (approximately 1–1.5 hr) at four time points: pre-adoption, 3 months after placement, 2 years postplacement, and 5–6 years following placement. Goldberg and Smith (2013) use the LGBTCI items with the TAPP data set as a single scale of perceived workplace climate to predict employees’ mental health outcomes among dual-earner parents. These authors used the original scoring recommendations of the LGBTCI (Liddle et al., 2004) with all participants in the data set.

For the purposes of the current study, we used Wave 1 data only, which occurred pre-adoption (i.e., before participants were parents). Given our interest in the experiences of LGBQ individuals in the workplace, we exclude those in the heterosexual (i.e., different-sex partnered) comparison group. Further, given that data were collected from partners, we eliminated issues of data interdependency (see Kenny, Kashy, & Cook, 2006) by only using data from the primary partner in each dyad. We excluded participants who were missing on all 20 LGBTCI items, leaving a final sample of 99 (Mage = 37.92, SD = 5.16). Participants were asked what their gender was in an open-ended question; just over half of participants identified as female (52.53%) and 47.47% as male. The majority of this sample (90%) were White, and almost all were working full time (84.54%). The average personal income was US$76,263.23 (SD = US$66,292.66), and average combined income was US$149,950.10 (SD = US$103,810.40). The majority of the sample indicated that they were “exclusively gay/homosexual” (67.35%), followed by “predominantly gay/homosexual” (25.51%) and “equally gay/homosexual and straight/heterosexual (7.14%).” All women were in female same-sex relationships, and all men were in male same-sex relationships.

Measures

LGBTCI.

Workplace climate was assessed by asking participants to indicate how well 20 different items described their primary workplace. As discussed, 12 items measured workplace support and 8 items measured negative experiences and hostility in the workplace (see Table 1). Response options include doesn’t describe at all = 1, describes somewhat or a little = 2, describes pretty well = 3, and describes extremely well = 4. In the original scoring procedures, the 8 items indicating workplace hostility were reverse coded prior to summing, so that higher scores indicate a more supportive workplace environment. Using the original scoring procedure, the LGBTCI demonstrated adequate reliability with the current sample (Cronbach’s α = .94; Liddle et al., 2004). Results from a confirmatory factor analysis of a one-factor structure also indicated adequate fit to the data, χ2(170) = 909.56, p < .001, Root Mean Square Error of Approximation (RMSEA) = .10, 95% CI [.09, .11], Comparative Fit Index (CFI) = .879, Tucker-Lewis Index (TLI) = .865, Standardized Root Mean Square Residual (SRMR) = .057.1

Table 1.

LGBT Climate Inventory Items and Recoded Participant Reponses.

Item Doesn’t Describe at All/
Describes Somewhat or a
Little (%)		 Describes Pretty Well/
Describes Extremely
Well (%)
Workplace Support subscale
1. Lesbian, gay, bisexual, and transgender (LGBT) employees are treated with respect. 17.42 82.58
3. Coworkers are as likely to ask nice, interested questions about a same-sex relationship as they are about a heterosexual relationship. 38.78 61.22
4. LGBT people consider it a comfortable place to work. 22.90 77.10
5. Non-LGBT employees are comfortable engaging in gay-friendly humor with LGBT employees (e.g., kidding them about a date). 48.17 51.83
7. LGBT employees feel accepted by coworkers. 22.48 77.52
11. My immediate work group is supportive of LGBT coworkers. 17.12 82.88
12. LGBT employees are comfortable talking about their personal lives with coworkers. 30.98 69.02
14. Employee LGBT identity does not seem to be an issue. 29.09 70.91
16. The company or institution as a whole provides a supportive environment for LGBT people. 28.08 71.92
17. LGBT employees are free to be themselves. 26.03 73.97
19. LGBT employees feel free to display pictures of same-sex partners. 39.04 60.96
20. The atmosphere for LGBT employees is improving. 42.40 57.60
Workplace Hostility subscale
2. LGBT employees must be secretive. 89.55 10.45
6. The atmosphere for LGBT employees is oppressive. 92.50   7.50
8. Coworkers make comments that seem to indicate a lack of awareness of LGBT issues. 79.00 21.00
9. Employees are expected to not act “too gay.” 84.17 15.83
10. LGBT employees fear job loss because of sexual orientation. 94.31   5.69
13. There is pressure for LGBT employees to stay closeted (to conceal their sexual orientation or gender identity/expression). 91.55   8.45
15. LGBT employees are met with thinly veiled hostility (e.g., scornful looks or icy tone of voice). 97.25   2.75
18. LGBT people are less likely to be mentored. 93.17   6.83
Open in a new tab

Given the distribution of responses across LGBTCI items with the current samples, we recoded responses to each of the 20 items as doesn’t describe at all/describes somewhat or a little = 0 and describes pretty well/describes extremely well = 1.

LCA

We used LCA to estimate classes of work-related experiences in our sample of LGBQ employees. LCA is a stochastic analytic procedure that uses categorical indicators to identify classes or “profiles” within a sample. Similar to latent factor analysis, LCA is a latent measurement model procedure but estimates the model fit of a categorical latent variable to the distributions of the data instead of a continuous latent factor (for more details, see Collins & Lanza, 2010; Kline, 2016; Masyn, 2013). The use of LCA helped identify groups defined by similar response patterns across the LGBTCI items.

Although latent class analytic procedures are often used to establish groups for comparison, we used LCA as a measurement model to help support or challenge the current scoring structure of the LGBTCI. Specifically, we used the 20 LGBTCI items to estimate latent classes, or profiles, of individuals that are characterized by their response patterns to the LGBTCI items. The identification of profiles is done through class enumeration, an iterative process that is conducted by fitting the data to a one-class (k) model and comparing subsequent fit to a k + 1 class model until models are not well identified. Each k-class model is assessed in terms of its relative fit to each subsequent model Bayesian information criterion (BIC), consistent Akaike’s information criterion (CAIC), and approximate weight of evidence criterion (AWE), where smaller values indicate better fit, Bayes factor (BF^) and correct model probability (cmP^; see Masyn, 2013). We also took into account the Lo-Mendell-Rubin likelihood ratio test (LMR-LRT), where model fit is subsequently compared to a model with k – 1 classes, andp < .05 indicates that the current model fit the data significantly better than a k – 1 class model. Finally, we assessed the interpretability of the best fitting class and compared that to the interpretability of the k – 1 and k + lclass models.

Results

We estimated and assessed the fit of one- through seven-class model on the dichotomized 20 item indicators from the LGBTCI. A review of relative fit indicators and the LMR-LRT indicated that a four-class solution fit the data best (see Table 2). Specifically, the log-likelihood value, BIC, CAIC, and AWE indicated a leveling or minimum value with a four-class model relative to three- and five-class models. Results from the LMR-LRT also indicated that a four-class model fit significantly better than a three-class model, but that a five-class model did not show an improvement in fit to the data compared to the four-class model. Finally, both our calculated BF^ and cmP^ cmP^—which provide evidence that a k-class model is the correct model relative to another (BF^) or all other possible models (cmP^ )—suggested a four-class model fit the data best. When interpreted, the four-class model profiles were empirically and theoretically sound.

Table 2.

Model Fit Indices for One-Class Through Seven-Class Model of LGBT Climate Inventory Items.

k Class LL BIC CAIC AWE LMR-LRT p Value BF^k,k+1 cmP^K
One class −4,176.67 8,475.17 8,495.17 8,656.99 <.10 .000
Two class −3,038.60 6,326.94 6,367.94 6,699.69 2,258.49 <.001 <.10 .000
Three class −2,879.45 6,136.55 6,198.55 6,700.21   315.84 <.001 <.10 .001
Four class −2,808.74 6,123.05 6,206.05 6,877.63   140.32 .050 >100.00 .999
Five class −2,761.30 6,156.09 6,260.09 7,101.58   94.15 .205 >100.00 .000
Six class −2,730.28 6,221.97 6,346.97 7,358.39   61.55 .127 >100.00 .000
Seven class −2,702.03 6,293.40 6,439.40 7,620.73   56.05 .141 .000 .000
Open in a new tab

Note. LL = log likelihood; BIC = Bayesian information criterion; CAIC = consistent Akaike’s information criterion; AWE = approximate weight of evidence criterion; LMR-LRT = Lo–Mendell–Rubin likelihood ratio test; BF^ = Bayes factor; cmP^ = correct model probability; LGBT = lesbian, gay, bisexual, and transgender.

Figure 1 displays item response probabilities for the four-class solution. Class 1 (56% of the overall sample) had a high probability of reporting workplace support and a low probability of workplace discrimination (<10% across all items); therefore, we named this class supportive work climate.

Figure 1.

Figure 1.

Open in a new tab

Model-estimated, class-specific item probabilities plot for the four-class solution of LGBTCI items.

Class 2 (21% of the sample) indicated relatively high workplace support with the exception of Items 3 and 5, which reflect coworkers’ comfort in asking questions about LGBQ employees’ relationship and LGBQ-related humor. Participants in Class 2 also reported relatively low workplace discrimination but indicated that some coworkers seemed to have a lack of awareness of LGBQ issues (Item 8, 28%) and that they were expected to not act too gay (15%). We named this class tolerant work climate.

Participants in Class 3 (17% of the sample) reported moderate to low levels of workplace support and generally moderate levels of discrimination. Nearly half of Class 3 participants, for example, indicated that LGBQ employees were treated with respect and that their immediate work group was supportive, but at the same time, these participants felt their coworkers lacked awareness of LGBQ issues and that LGBQ employees had to be secretive and stay closeted at work. Given that levels of support and hostility were similar, we called this class ambiguous work climate.

Class 4, the hostile work climate class (6% of the sample), was made up of participants who reported low levels of workplace support across items and high levels of workplace discrimination. For example, 95% of the hostile work climate class stated that they were expected to not act “too gay” and over half reported that they feared job loss because of their sexual orientation and felt pressure to stay closeted at work.

Discussion

Our findings indicate that the original scoring recommendations for the LGBTCI essentially “wash out” the presence of perceived hostility, especially for the individuals in Class 2 (the tolerant work climate) and Class 3 (the ambiguous work climate) in our sample. By reverse coding the reported markers of hostility and summing those items with reported support (Liddle et al., 2004), the LGBTCI measure presents hostility as the absence of support and vice versa. In effect, this scoring procedure oversimplifies the conceptualization of workplace climate; an organization is deemed supportive or hostile. In reality, understanding and measuring climate requires a much more nuanced analysis, given that, for some, support and hostility were co-occurring.

The current study uses LCA to investigate whether the current scoring procedure of the LGBTCI obscures the co-occurrence of workplace support and hostility for LGBQ employees. Results from this analysis show that for some employees, positive and negative workplace experiences are strongly inversely related. Their workplace climate can be described as predominantly supportive or predominantly hostile toward LGBQ employees. For others, though, their workplace climate may be simultaneously supportive and hostile. Over 20% of the current sample reported co-occurring markers of support and hostility in the workplace. Still others reported a workplace climate that was neither hostile nor overly supportive. Seventeen percent of the sample reported that despite a lack of perceived hostility in the workplace, support was also not apparent. Together, these results call into question the current scoring approach to the LGBTCI and suggest that alternative methods may be better suited to capture a more complex climate for LGBQ individuals.

For example, participants in the tolerant work climate class felt generally supported in that context, although they endorsed these items to a lesser degree than the supportive class. The response pattern of the tolerant class, however, overlapped on several items of support and hostility. That is, despite perceiving a general sense of support—that “LGBT employees are treated with respect” and “free to be themselves” —they indicated less support and more interpersonal hostility. Specifically, compared to the supportive class, those in the tolerant group were less likely to indicate that “coworkers asked nice questions about same-sex relationships” and “coworkers were comfortable engaging in gay-friendly humor” and more likely to indicate that “coworkers lacked awareness of LGBT issues.” The similar response rates of these 3 items in particular are interesting, as they all assess climate on a more microlevel. It seems those in the tolerant class perceived support in the organization as a whole, but more hostility in their proximal work groups, which can have important implications for day-to-day well-being and worker productivity.

Conceptualized in this way, tolerance becomes more than simply the middle ground between support and hostility. Instead, tolerance is a unique contextual climate (Chojnacki & Gelberg, 1994), one that is defined by a general sense of support, but also very clear limitations to that support. Tolerance may be better described as providing some level of institutional support, despite discomfort or misunderstanding among coworkers. Even though a business or state has employment protections for LGBQ employees, for example, this population may still perceive discrimination and hostility in the work environment, particularly microaggressions that are often less preventable by policy. For LGBQ employees, a tolerant work climate communicates support and at the same time discomfort with their sexual orientation or a lack of cultural competency that makes the workplace climate unwelcoming or exclusionary.

On the other hand, an ambiguous climate is distinct from this conceptualization. Ambiguity in the work context is instead characterized by relatively low levels of support and relatively low levels of hostility. LGBQ employees may be left wondering whether their identity will be met with discrimination or inclusion. Participants in the ambiguous class responded with the most support for the items: “LGBT employees are treated with respect,” “My immediate work group is supportive of LGBT employees,” and also “LGBT employees are expected to not act ‘too gay’” and “LGBT employees must be secretive.” The similar levels of agreement for these items indicate that sexual minority employees may not perceive direct stigmatization but also may be unsure of their ability to be out in the workplace and be supported. LGBQ employees in an ambiguous work climate could be left not knowing where they stand in terms of support and hostility.

Encouragingly, our largest class was the supportive work climate—characterized largely by the presence of workplace support and lack of workplace hostility. This suggests that many LGBQ employees feel positively about their workplace climate. The presence of merely tolerant, ambiguous, and even hostile work climates, however, highlights the need for policies and programs that improve workplace climate for all LGBQ people. For those perceiving hostility in the workplace, policies explicitly enumerating protections for LGBQ people could, if similar to the process of policies protecting LGBQ youth in schools (Hatzenbuehler, Schwab-Reese, Ranapurwala, Hertz, & Ramirez, 2015), help decrease experiences of hostility in the workplace for LGBQ employees. In situations where hostility is present, nondiscrimination laws provide LGBQ employees’ actionable recourse for combating hostility in the workplace. Similarly, laws protecting LGBQ employees would likely increase employers’ proactive engagement of employees around cultural competency training, which may help to increase support for those employees who perceive their work climate to be more ambiguous or tolerant.

Interestingly, scholars who assess climate for LGBQ individuals in other contexts, namely, family and community, also note the simultaneous presence of similar levels of support and hostility. For example, one case study described a young lesbian mother who experienced conflicting messages of support and hostility in her rural community (Mendez, Holman, Oswald, & Izenstark, 2016). Similar to participants in the tolerant work climate class, this woman received messages of support for her sexual minority identity a a macrolevel and yet reported microaggressions and hostile comments from those in her direct environment (including coworkers). Similarly, Glass and Few-Demo (2013) found that Black lesbian couples were likely to report co-occurring experiences of support and hostility within their families and church communities. Thus, it is important that researchers examine workplace climate in ways that allow for complex categorizations that move beyond a unidimensional measure.

Results from the current study indicate that a more accurate measure and analysis of workplace climate is needed, specifically one that includes separate scales for support and hostility. Measuring climate in this way would more accurately capture the nuances of workplace climate for LGBQ employees. Further, this operationalization of workplace climate would more closely align research methodology with the tenets of minority stress theory (Meyer, 2003). Continued use of the current scoring system could bias future work in this area. Accurate measures of climate are needed in order to truly assess the effects of the work environment on workplace retention and productivity, as well as LGBQ individuals’ well-being. Although Brewster et al. (2012) offer a practical solution to improve model fit when using the recommended scoring procedure of the LGBTCI with their latent orthogonal approach, this solution is specific to latent variable modeling and does not adequately capture the co-occurring support and hostility that we illustrate with our classes here. Therefore, we recommend that future studies use support and hostility as independent measures of workplace climate when assessing correlates and outcomes of LGBQ employees’ experiences. The positively and negatively worded items of the LGBTCI (see Table 1) can, and should, be used as separate subscales.

Limitations

Although the sample for this analysis was drawn from two different data sets in order to increase sociodemographic diversity, we recognize the relative homogenous nature of the samples. In particular, the samples used for this analysis were predominantly White, female, and middle to upper middle class, thus limiting the generalizability of our findings. Sexual minorities’ experiences in the workplace can vary as a function of their unique social locations (e.g., the intersections of gender, race, and socioeconomic standing, beyond simply sexual orientation), and it is important for scholars to recognize this diversity within LGBQ populations. Ideally, we would have been able to statistically test the sociodemographic differences between the two data sets; however, the lack of consistent measures across studies precluded this assessment. As such, this analysis should be replicated with more diverse samples, including population-based, nationally representative data, in order to further examine the distribution of LGBQ employees’ responses regarding workplace climate across the four classes identified in this study.

Similarly, transgender employees’ experiences in the workplace warrant scholarly attention. Experiences around gender identity and gender expression are uniquely different from those related to sexual orientation (e.g., Schilt, 2006). Indeed, transgender employees report various levels of discrimination and support in their work environments in relation to their gender (Hartzell, Frazer, Wertz, & Davis, 2009). Despite Liddle and colleagues’ (2004) intention to measure work climate for both sexual and gender minorities, little attention has been paid to the unique experiences of transgender employees—aside from Brewster and colleagues’ (2012) assessment of the LGBTCI-TF. However, the samples used for this analysis also focused specifically on sexual minority populations and therefore did not include transgender participants. While it is outside the scope of the current analysis to discuss how this measure may be used with transgender employees, the future work in this area is needed.

Conclusion and Future Directions

Our findings lend credence to the idea that support and hostility are not two sides of the same coin and should not be operationalized as opposing ends of a single continuum. Instead, when assessing climate, scholars need to independently measure levels of support and hostility in the workplace environment as they may be co-occurring or dually absent. This study showcases the benefits of utilizing more descriptive approaches to assessing workplace experiences of LGBQ people by using LCA. Such approaches may prove useful when attempting to assess differences in specific supportive and hostile experiences and how these nuanced differences may relate to occupational differences as well as individual health and well-being. More work is needed to test outcome variables with these newly identified classes of workplace climate using the updated scoring recommendations. It may be that LGBQ employees’ well-being is affected differently by a context of tolerance or ambiguity. Furthermore, research should examine how these classes of workplace climate affect diverse LGBTQ populations differently.

As researchers look to inform supportive workplace policies at the local, state, and national level, capturing an accurate picture of workplace climate for LGBQ employees is vital—particularly given the shifting social and political landscape. As widespread workplace protections for LGBQ people continue to be debated (and denied), it is important that researchers can more fully and accurately understand the workplace experiences for this population and its impact on health and well-being. More work is needed to classify work environments in this nuanced way. By aligning theory and research methodology, scholars and researchers can more precisely measure and assess workplace climate for LGBQ employees and understand meaningful ways to intervene. With the onus of creating supportive environments for marginalized populations resting on the organizations and institutions, being able to capture the nuances of workplace support and hostility will allow family scholars and advocates to better serve as catalysts for change in this arena.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded in part by the National Institute on Alcohol Abuse and Alcoholism (awarded to Fish) grant number F32AA023138. This research was also supported by grant, P2CHD042849, Population Research Center, awarded to the Population Research Center at The University of Texas at Austin by the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

1

Given findings from Brewster et al. (2012) regarding the improvement of model fit with orthogonal factors (one for positively worded items, one for negatively worded items), we also tested the fit of a two-factor structure, χ2(169) = 634.14,p < .001, RMSEA = .08, 95% CI [.07, .09], CFI = .924, TLI = .915, SRMR = .040. χ2 difference test indicated that the two-factor model statistically fit the data better than the one-factor model, Δχ2(1) = 275.42,p < .001, where, similar to Brewster et al. (2012), positively worded items strongly loaded onto one factor and negatively worded items strongly loaded onto a second.

References

  1. Badgett MVL, Lau H, Sears B, & Ho D (2007). Bias in the workplace: Consistent evidence of sexual orientation and gender identity discrimination. Los Angeles, CA: Williams Institute. Retrieved from http://williamsinstitute.law.ucla.edu/wp-content/uploads/Badgett-Sears-Lau-Ho-Bias-in-the-Workplace-Jun-2007.pdf [Google Scholar]
  2. Boyles P (2008). “Thank you for letting me be myself”: Exploring the effects of identity management strategies on engagement levels of lesbian, gay and bisexual employees (Unpublished doctoral dissertation). Virginia Polytechnic Institute and State University, Blacksburg, VA. [Google Scholar]
  3. Brewster ME, Velez B, DeBlaere C, & Moradi B (2012). Transgender individuals’ workplace experiences: The applicability of sexual minority measures and models. Journal of Counseling Psychology, 59, 60–70. doi: 10.1037/a0025206 [DOI] [PubMed] [Google Scholar]
  4. Button SB (2001). Organizational efforts to affirm sexual diversity: A cross-level examination. Journal of Applied Psychology, 86, 17–28. doi: 10.1037//0021-9010.86.1.17 [DOI] [PubMed] [Google Scholar]
  5. Chojnacki JT, & Gelberg S (1994). Toward a conceptualization of career counseling with gay/lesbian/bisexual persons. Journal of Career Development, 21, 3–10. [Google Scholar]
  6. Collins LM, & Lanza ST (2010). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. Hoboken, NJ: Wiley. [Google Scholar]
  7. Embrick DG, Walther CS, & Wickens CM (2007). Working class masculinity: Keeping gay men and lesbians out of the workplace. Sex Roles, 56, 757–766. doi: 10.1007/s11199-007-9234-0 [DOI] [Google Scholar]
  8. Gates GJ, & Ost J (2004). The gay and lesbian atlas. Washington, DC: Urban Institute Press. [Google Scholar]
  9. Giorgi A (1985). Sketch of a psychological phenomenological method. In Giorgi A (Ed.), Phenomenology and psychological research (pp. 12–28). Pittsburgh, PA: Duquesne University Press. [Google Scholar]
  10. Glass VQ, & Few-Demo AL (2013). Complexities of informal social support arrangements for black lesbian couples. Family Relations, 62, 714–726. doi: 10.1111/fare.12036 [DOI] [Google Scholar]
  11. Goldbach JT, Tanner-Smith EE, Bagwell M, & Dunlap S (2013). Minority stress and substance use in sexual minority adolescents: A meta-analysis. Prevention Science, 15, 350–363. [DOI] [PubMed] [Google Scholar]
  12. Goldberg AE, & Smith JZ (2013). Work conditions and mental health in lesbian and gay dual-earner parents. Family Relations, 62, 727–740. doi: 10.1111/fare.12042 [DOI] [Google Scholar]
  13. Griffith KH, & Hebl MR (2002). The disclosure dilemma for gay men and lesbians: “Coming out” at work. Journal of Applied Psychology, 87, 1191–1199. doi: 10.1037/0021-9010.87.6.1191 [DOI] [PubMed] [Google Scholar]
  14. Hartzell E, Frazer MS, Wertz K, & Davis M (2009). The state of transgender California: Results from the 2008 California transgender economic health survey. San Francisco, CA: Transgender Law Center. [Google Scholar]
  15. Hatzenbuehler ML, Schwab-Reese L, Ranapurwala SI, Hertz MF, & Ramirez MR (2015). Associations between antibullying policies and bullying in 25 states. The Journal of the American Medical Association Pediatrics, 169, e152411. doi: 10.1001/jamapediatrics.2015.2411 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Herek GM (2009). Hate crimes and stigma-related experiences among sexual minority adults in the United States: Prevalence estimates from a national probability sample. Journal of Interpersonal Violence, 24, 54–74. doi: 10.1177/0886260508316477 [DOI] [PubMed] [Google Scholar]
  17. Holman EG (2016). Community climate. In Goldberg AE (Ed.), The SAGE encyclopedia of LGBTQ studies (pp. 252–255). Los Angeles, CA: Sage. doi: 10.4135/9781483371283.n91 [DOI] [Google Scholar]
  18. Huebner DM, & Davis MC (2007). Perceived antigay discrimination and physical health outcomes. Health Psychology, 26, 627–634. doi: 10.1037/02786133.26.5.627 [DOI] [PubMed] [Google Scholar]
  19. Huffman AH, Watrous-Rodriguez KM, & King EB (2008). Supporting a diverse workforce: What type of support is most meaningful for lesbian and gay employees? Human Resource Management, 47, 237–253. doi: 10.1002/hrm.20210 [DOI] [Google Scholar]
  20. Kenny DA, Kashy DA, & Cook WL (2006). Dyadic data analysis. New York, NY: Guilford Press. [Google Scholar]
  21. Kline RB (2016). Principles and practice of structural equation modeling (4th ed.). New York, NY: Guilford Press. [Google Scholar]
  22. Köllen T (2013). Bisexuality and diversity management—Addressing the B in LGBT as a relevant ‘sexual orientation’ in the workplace. Journal of Bisexuality, 13, 122–137. [Google Scholar]
  23. Köllen T (2015). Lessening the difference is more—The relationship between diversity management and the perceived organizational climate for gay men and lesbians. The International Journal of Human Resource Management, 27, 1967–1996. doi: 10.1080/09585192.2015.1088883 [DOI] [Google Scholar]
  24. Lick DJ, Durso LE, & Johnson KL (2013). Minority stress and physical health among sexual minorities. Perspectives on Psychological Science, 8, 521–548. doi: 10.1177/1745691613497965 [DOI] [PubMed] [Google Scholar]
  25. Liddle BJ, Luzzo DA, Hauenstein AL, & Schuck K (2004). Construction and validation of the lesbian, gay, bisexual, and transgendered climate inventory. Journal of Career Assessment, 12, 33–50. doi: 10.1177/1069072703257722 [DOI] [Google Scholar]
  26. Masyn KE (2013). Latent class analysis and finite mixture modeling. In Little TD (Ed.), The Oxford handbook of quantitative methods in psychology: Vol. 2: Statistical analysis (pp. 551–611). New York, NY: Oxford University Press. [Google Scholar]
  27. Mendez SN, Holman EG, Oswald RF, & Izenstark D (2016). Minority stress in the context of rural economic hardship: One lesbian mother’s story. Journal of GLBT Family Studies, 28, 214–230. doi: 10.1080/1550428X.2015.1099493 [DOI] [Google Scholar]
  28. Meyer IH (1995). Minority stress and mental health in gay men. Journal of Health & Social Behavior, 36, 38–56. [PubMed] [Google Scholar]
  29. Meyer IH (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychology Bulletin, 129, 674–697. doi: 10.1037/0033-2909.129.5.674 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Meyer IH (2007). Prejudice and discrimination as social stressors. In Meyer IH & Northridge ME (Eds.), The health of sexual minorities: Public health perspectives on lesbian, gay, bisexual, and transgender populations (pp. 242–267). New York, NY: Springer. [Google Scholar]
  31. Movement Advancement Project. (2017). Non-discrimination laws. Retrieved from http://www.lgbtmap.org/equality-maps/non_discrimination_laws
  32. Oswald RF, Cuthbertson C, Lazarevic V, & Goldberg AE (2010). New developments in the field: Measuring community climate. Journal of GLBT Family Studies, 6, 214–228. doi: 10.1080/15504281003709230 [DOI] [Google Scholar]
  33. Oswald RF, & Holman EG (2013). Rainbow Illinois: How downstate LGBT communities have changed 2000-2011. Champaign, IL: Author. [Google Scholar]
  34. Pizer JC, Seats B, Mallory C, & Hunter ND (2012). Evidence of persistent and pervasive workplace discrimination against LGBT people: The need for federal legislation prohibiting discrimination and providing for equal employment benefits. Loyola of Los Angeles Law Review, 45, 715–779. [Google Scholar]
  35. Prati G, & Pietrantoni L (2014). Coming out and job satisfaction: A moderated mediation model. The Career Development Quarterly, 62, 358–371. doi: 10.1002/j.2161-0045.2014.00088.x [DOI] [Google Scholar]
  36. Ragins BR, & Cornwell JM (2001). Pink triangles: Antecedents and consequences of perceived workplace discrimination against gay and lesbian employees. Journal of Applied Psychology, 86, 1244–1261. doi: 10.1037/0021-9010.86.6.1244 [DOI] [PubMed] [Google Scholar]
  37. Ragins BR, Singh R, & Cornwell JM (2007). Making the invisible visible: Fear and disclosure of sexual orientation at work. Journal of Applied Psychology, 92, 1103–1118. doi: 10.1037/0021-9010.92.4.1103 [DOI] [PubMed] [Google Scholar]
  38. Reed L, & Leuty ME (2015). The role of individual differences and situational variables in the use of workplace sexual identity management strategies. Journal of Homosexuality, 63, 985–1017. doi: 10.1080/00918369.2015.1117900 [DOI] [PubMed] [Google Scholar]
  39. Ricci-Stiles N (2007). Predictors of job satisfaction for gay, lesbian, and bisexual employees (Unpublished doctoral dissertation). Walden University, Minneapolis, MN. [Google Scholar]
  40. Ruggiero D, & Park M (2017). DOJ files amicus brief that says title VII does not protect sexual orientation. CNN. Retrieved from https://www.cnn.com/2017/07/26/politics/doj-amicus-brief-title-vii-sexual-orientation/index.html [Google Scholar]
  41. Schenker N, & Raghunathan TE (2007). Combining information from multiple surveys to enhance estimation of measures of health. Statistics in Medicine, 26, 1802–1811. doi: 10.1002/sim.2801 [DOI] [PubMed] [Google Scholar]
  42. Schilt K (2006). Just one of the guys? How transmen make gender visible at work. Gender & Society, 20, 465–490. doi: 10.1177/0891243206288077 [DOI] [Google Scholar]
  43. Velez BL, & Moradi B (2012). Workplace support, discrimination, and person-organization fit: Tests of the theory of work adjustment with LGB individuals. Journal of Counseling Psychology, 59, 399–407. doi: 10.1037/a0028326 [DOI] [PubMed] [Google Scholar]
  44. Velez BL, Moradi B, & Brewster ME (2013). Testing the tenets of minority stress theory in workplace contexts. Journal of Counseling Psychology, 60, 532–542. doi: 10.1037/a0033346 [DOI] [PubMed] [Google Scholar]
  45. Wendland J (2007). A new beginning for ENDA. The Williams Institute. Retrieved from https://web.archive.org/web/20080125092049/http://www.law.ucla.edu/williamsinstitute/press/ANewBeginningforENDA.html [Google Scholar]

RESOURCES