Introduction

Ten percent of Europeans identify themselves as non-heterosexual (Deveaux, 2016). Thus, LGB + employees (lesbian, gay, bisexual and other non-heterosexual individuals) already represent a relatively large part of the European workforce. Nevertheless, many of LGB + employees choose not to disclose their sexual orientation at work (Frohn et al., 2017). Thus far, there is relatively little research on factors related to LGB + employees choosing not to disclose at their workplace. Therefore, this study aims at investigating the role of personality traits on disclosing a non-heterosexual orientation at work. Specifically, the five-factor personality traits as well as impulsivity, locus of control (LoC), and self-esteem were addressed, as will be outlined below.

Disclosure of a non-heterosexual orientation at work

In a recent German study, 30.6% of lesbian and gay as well as 55.5% of the bisexual participants reported that they talk about their sexual orientation to no one or only very few people at work (Frohn et al., 2017). The most common reasons for LGB + employees not to disclose their sexual orientation are fear of discrimination at their workplace, possible disadvantages in career opportunities or loss of employment (Frohn et al., 2017).

However, disclosing a non-heterosexual orientation of LGB + employees has been shown to be positively associated with well-being and negatively with feelings of depression, anger, and anxiety (Badgett et al., 2013; Pachankis et al., 2015). In line with these findings, LGB + employees who disclose their non-heterosexual orientation also show lower stress levels and lower job-related anxiety (Day & Schoenrade, 2000). Moreover, disclosing at work relates to higher job satisfaction, higher satisfaction with co-workers and higher organizational commitment (Lloren & Parini, 2017).

Previous research has focused mainly on work environmental factors that are potentially related to disclosure at the workplace, such as a LGB + supportive policy or work climate (Lloren & Parini, 2017; Webster et al., 2018). A few studies also examined the relationship between disclosure behavior at work and inter-individual differences such as age, gender or ethnicity (Frohn et al., 2017; Ragins et al., 2003). However, to our knowledge, the potential relationship between the disclosure of a non-heterosexual orientation at work and personality factors of LGB + employees has not been systematically investigated yet. In addition, our study provides the opportunity to replicate previous findings, such as the role of age, the presence of an intimate partner, and a LGB + supportive work climate in disclosing a non-heterosexual orientation at the workplace.

The five-factor model of personality and disclosure behavior at work

The Five-Factor Model of personality (FFM; McCrae & John, 1992) describes a higher-order structure of personality based on five basic personality dimensions (McCrae & Costa, 1997), which in turn might be related to disclosure behavior. Extraversion is associated with positive emotions and optimism (Canu et al., 2007). Thus, extraverted LGB + employees may tend to expect more positive responses from co-workers, superiors and subordinates after disclosing their sexual orientation. Additionally, extraversion is associated with being communicative and sociable. So, extraverted LGB + employees may be more inclined to reveal their sexual orientation and may assume that they are able to build stronger and more authentic social relationships to others (Costa & McCrae, 1992), when they decide to disclose (Ragins et al., 2007).

Neuroticism is a tendency to experience negative emotions such as anxiety, depression or self-consciousness (Costa & McCrae, 1992; Schulte et al., 2004). LGB + employees who score higher in neuroticism could worry more about potential negative consequences and might feel more anxious about disclosing their non-heterosexual orientation. Thus, neuroticism might be negatively related to disclosure at work.

Conscientiousness is positively associated with achievement striving (Schmidt & Hunter, 1998), and conscientious individuals are less likely to take career risks (Costa & McCrae, 1992), such as quitting a job without having another one (Nicholson et al., 2005). As noted above, one of the most common reasons why LGB + employees choose not to disclose at work is their fear of being disadvantaged regarding career opportunities (Frohn et al., 2017). Therefore, disclosing at work may seem to be viewed as a career risk, which, presumably, a more conscientious LGB + employee could be less willing to take on. On the other hand, however, conscientiousness has been strongly associated with integrity and honesty (Horn et al., 2004) as well as authenticity (Wood et al., 2008), traits that may render disclosure of a non-heterosexual orientation more likely.

Openness is linked with independence of judgement, and open individuals tend to be unconventional and to oppose social or cultural norms, authority and tradition (Costa & McCrae, 1992; Schulte et al., 2004). LGB + employees with higher openness might thus be less concerned about violating potential norms with disclosing at work. Moreover, similar to conscientiousness, open individuals are considered more authentic (Wood et al. 2008).

Individuals who describe themselves as agreeable show more altruism and compliance, but also sociotropy, which refers to excessive concern about what others think (Bagby et al., 2001; Schulte et al., 2004). Agreeable individuals also tend to avoid conflicts (Antonioni, 1998). Therefore, LGB + employees who score higher in agreeableness could possibly decide not to disclose.

Impulsiveness, self-esteem, locus of control and disclosure at work

General self-esteem is considered an individual’s positive or negative attitude toward the self as a whole (Rosenberg et al., 1995). Self-esteem is positively related to self-acceptance, whereas low self-esteem seems to be related to internalized homophobia (Peterson & Gerrity, 2006). Thus, one may assume that LGB + employees possessing a lower self-esteem could potentially be less inclined to disclose themselves at work.

LoC (Rotter, 1966) reflects the extent to which individuals believe that positive reinforcements are either a consequence of their own actions (internal LoC) or are unrelated to their behavior (external LoC). Internal LoC is related to better adaptation to new work settings (Van der Horst et al., 2017), and better coping with stressful situations (Groth et al., 2019). LGB + employees possessing a higher internal LoC may thus expect to be better able to adapt to challenges that might be associated with disclosure at work and may feel more competent to cope with potential negative consequences. In turn, external LoC has been associated with a lower ability to cope with stressful situations (Groth et al., 2019), which may make such LGB + employees less willing to disclose at work.

Additionally, we examined impulsiveness, which is the predisposition to rapid, unplanned reactions to internal or external events without adequately considering possible negative consequences (Moeller et al., 2001). Hence, LGB + employees possessing a higher impulsivity may be more inclined to disclose their sexual orientation because they might not think extensively about possible negative consequences or risks in advance.

Moreover, studies suggest that gender may moderate the relationship between personality (including FFM traits, impulsivity, self-esteem, and LoC) and various behavioral outcomes (including work-, health-, and risk-related behaviors; e.g., Spector & Zhou, 2014; Tallman & Bruning, 2008; Stoltenberg et al., 2008). Therefore, interaction effects of gender and personality on disclosure behavior were exploratively investigated and reported in the present study.

Methods

Participants and procedure

The data for this cross-sectional study were collected via an online survey for German-speaking participants that was conducted in 2019 at the Medical School Berlin. Before data collection a power analysis using G*Power 3.1.9.2 (Faul et al., 2009) indicated that with a desired power of 0.90, an alpha error probability of 0.05 and an estimated effect size of about 0.03, and concerning all relevant predictors, 353 participants are needed. The online survey platform Unipark (Questback, Cologne, Germany) was used for data collection and the participants were recruited online by sharing the link to the survey in LGB + groups on Facebook as well as the German social media platform Jodel. The study was conducted in accordance with the Declaration of Helsinki (revised form). The data collection was absolutely anonymous and the study participation was voluntary. Participants were carefully informed by an instruction page prior to the survey that they could withdraw from the study at any time and without giving any reason. All participants were at least 18 years of age and were informed that they would give their consent to study participation with continuing the online questionnaire. There was no compensation for participation, except for students who received course credit using SONA, a web-based system at the Medical School Berlin. Participants who identified themselves as LGB+, were 18 years or older and had a job in which they work at least 5 h per week were eligible to participate in the study. Of a total of 377 participants five had to be excluded, because they did not meet the selection criteria. From the remaining 372 participants (females: 52.7%; males: 43.8%; other: 3.5%) the majority identified themselves as either homosexual (n = 230) or bisexual (n = 104). The remaining participants identified themselves as polysexual (n = 3), pansexual (19), asexual (n = 10) or other (n = 6). Age ranged between 18 and 48 (M = 24.64; SD = 5.27) and participants worked between 5 and 60 h per week (M = 28; SD = 14.05). A more detailed description of participants’ characteristics is given in Table 1.

Table 1 Sample’s socio-demographic characteristics
Full size table

Measures

Disclosure of a non-heterosexual orientation at work

We listed co-workers (close and less co-workers), superiors and subordinates to assess disclosure at work. Additionally, we added the category friends as disclosure to heterosexual friends was shown to be significantly associated with disclosure behavior at work (Griffith & Hebl, 2002) and thus, this variable might be relevant to separate work disclosure from general disclosing behavior. The participants had to rate these groups on a 4-point Likert scale ranging from 1 = ‘This group of people is definitely not aware of my sexual orientation’ to 4 = ‘This group of people is definitely aware of my sexual orientation’. Cronbach’s α (four items) was 0.90.

Five-factor model

The FFM was assessed with the BFI-K, a German short version of the FFM Inventory (Rammstedt & John, 2005), consisting of 21 items, five items for openness and four each for the other traits. Cronbach’s α was 0.86 for extraversion, 0.79 for neuroticism, 0.71 for openness, 0.67 for conscientiousness and 0.60 for agreeableness, which corresponds to the validation study of the BFI-K (Rammstedt & John, 2005).

Impulsiveness

Impulsiveness was assessed with the BIS-15, a German short version of the Barratt Impulsiveness Scale (Meule et al., 2011), consisting of 15 items, five items each for the three subscales non-planning, motor- and attention-based impulsiveness. Higher scores indicate higher impulsiveness. Cronbach’s α was 0.81.

Self-esteem

Self-esteem was assessed with the German form of the Rosenberg scale (von Collani & Herzberg, 2003), consisting of 10 items. Higher scores indicate higher self-esteem. Cronbach’s α was 0.92.

LoC

LoC was assessed with a German short form of the Internal-External LoC (IE-4) scale (Kovaleva, 2012), consisting of four items, two each for internal and external LoC. Higher scores on the LoC subscales indicate higher internal and external LoC, respectively. Cronbach’s α was 0.75 for internal and 0.44 for external LoC.

Potential covariates

In previous studies, age, gender, perceived sexual orientation of co-workers, and having a romantic/intimate partner were associated with disclosure at work (Frohn et al., 2017; Pew Research Center, 2013; Ragins et al., 2007). Thus, these variables were assessed as potential covariates. Furthermore, being part of another minority group may influence disclosure behavior of LGB + individuals because they may face further challenges and prejudice as part of two minority group (Miller, 2018; Ragins et al., 2003; Rosario et al., 2004; Shakespeare, 1999). So, the participants were also asked if they have experienced discrimination at work due to their ethnicity, their age, their religion or a mental or physical disorder/disability (response options were ‘Yes’ and ‘No’). About 20% of the participants reported having such experiences, and thus this variable was considered as covariate in the analyses. Moreover, an LGB + supportive workplace climate and relationships have been shown to be positively associated with disclosure at work (Webster et al., 2018). According to this literature, we developed two items each to assess LGB + supportive climate (e.g., ‘The climate at my workplace is pleasant for LGB+’) and LGB + supportive relationships at work (e.g., ‘All employees at my workplace accept and feel comfortable around LGB+’). The response format was a 5-point Likert scale ranging from 1 = ‘Strongly Disagree’ to 5 = ‘Strongly agree’. The Spearman Brown coefficient for the two-item scale LGB + supportive climate was .93 and .88 for supportive relationships. The means were used for data analysis. Finally, the number of work hours was included as covariate as individuals who spend more time at work might be more inclined to show disclosing behavior at work.

Data analysis

The data was analyzed using IBM SPSS Statistics 25. Correlation and hierarchical regression analysis was used to examine the relationship between personality and disclosing at work. Participants in the gender category ‘other’ were excluded from analyses because we also examined the relationship between gender with disclosing at work and this subgroup was too small (n = 13) to produce meaningful results. The data was checked for all relevant prerequisites for statistical analyses (Casson & Farmer, 2014). Those potential covariates along with gender that showed significant correlations (p < .05) with disclosure at work were entered in the first step of the regression analysis. Subsequently, personality traits that were significantly associated with disclosure behavior (p < .05) were entered in the second step to examine whether their bivariate associations observed in the correlation analyses showed incremental validity and whether personality explains variance over and above the already known predictors (e.g., presence of an intimate partner) in disclosure behavior at work.

Similar to Griffith and Hebl (2002), disclosure at work was strongly associated with disclosure to friends (r = .485, p < .001), which may suggest that both variables are manifestations of the same latent construct (i.e., disclosure behavior). Because we specifically aimed to investigate the role of personality in workplace disclosure behavior, we conducted all analyses with a residualized variable of workplace disclosure, where the variable disclosure to friends was partialled out and which therefore should be more specific to work-related disclosure behavior. Additionally, the results for the unresidualized variable of workplace disclosure are given in the electronic supplementary material (see below).

In addition to our central research question of whether personality additively explains variance in workplace disclosure behavior, we exploratively examined the moderating role of gender in the relationship between the personality factors and workplace disclosure. A hierarchical moderated regression analysis was conducted. Because moderator effects do not require that predictor, moderator, and criterion variables are bivariate related, all personality variables (except self-esteem) were included in the regression model. To avoid spurious results due to multicollinearity for example by suppression effects (Morrow-Howell, 1994), self-esteem was excluded because of it’s strong correlation with neuroticism (r=-.692, p < .001) and in view of the larger association with disclosure behavior for neuroticism (see Table 3). In a first step, disclosure at work was regressed on the gender variable, the (centered) personality variables, and those covariates that showed significant effects on disclosure behavior in our main regression analysis. In a second step, the interaction terms of gender and each personality factor were added.

Results

Research data are available as electronic supplementary material (see below). The overall mean of disclosure at work (disclosure towards close and less close co-workers, superiors and subordinates) was M = 2.18 (SD = 0.96). Disclosure was the highest towards close co-workers (M = 2.79; SD = 1.14) followed by superiors (M = 2.04; SD = 1.15), subordinates (M = 2.02; SD = 1.10), and less close co-workers (M = 1.87; SD = 0.94). The relative frequencies of reported disclosure at work of all groups are shown in Table 2.

Table 2 Frequencies of reported disclosure at work
Full size table

Regarding the covariates, age (r = .241, p = < 0.001), having an intimate partner (r = .155, p = .003), LGB + supportive relationships (r = .277, p = < 0.001), LGB + supportive climate at work (r = .279, p = < 0.001), and work hours per week (r = .202, p = < 0.001) showed significant correlations with disclosure at work. Out of all personality variables, conscientiousness exhibited the strongest correlation with disclosure at work (r = .232, p = < 0.001), followed by neuroticism (r=-.193, p = < 0.001), extraversion (r = .183, p = < 0.001), and internal LoC (r = .157, p = .003). Agreeableness, openness, self-esteem, external LoC, and impulsiveness were not significantly correlated with disclosure at work (all p > .05). Intercorrelations of all study variables are given in Table 3.

Table 3 Intercorrelations of all study variables
Full size table

Table 4 shows the results of the hierarchical regression analysis. Due to their significant correlation with disclosure at work (see above), age, gender, intimate partner, LGB + supportive climate and relationships at work, and work hours per week were entered in the first step of the regression analysis. Subsequently, the personality factors conscientiousness, extraversion, neuroticism, and internal LoC were entered in the second step. The R2 for our overall model was 0.235 and the Adjusted R2 was 0.213, F (10, 355) = 10.583, p < .001. The R2 change between step 1 and 2 was 0.036, which is statistically significant, F (4, 345) = 4.068, p = .003. Age (\(\beta\)=0.208, p < .001), intimate partner (\(\beta\)=0.149, p = .002) and work hours per week (\(\beta\)= 0.131, p=.01) incrementally predicted disclosure at work. Regarding personality, only conscientiousness (\(\beta\) =0.151, p = .004) showed incremental validity in explaining disclosure at work in the hierarchical regression model. Thus, in contrast to the bivariate correlations, the other personality variables did not incrementally predict disclosure at work (all p > .05).

Table 4 Hierarchical regression predicting disclosure behavior at work
Full size table

Additionally, a moderated regression analysis was run to test whether gender moderates the influence of the personality traits on disclosure behavior at work. As outlined above in more detail, in a first step disclosure at work was regressed on gender, all personality variables (except self-esteem), and the significant covariates in our main regression analysis. Including the interaction terms of personality and gender in the second step, the model fit did not statistically improve. R2 change from step 1 to 2 is 0.022, which is statistically non-significant, F (8, 335) = 1.118, p = .350. Moreover, none of the interaction terms reached significance (all p > .05).

Discussion

The main aim of the present study was to examine whether and how basic personality traits are associated with the disclosure of a non-heterosexual orientation at work. First of all, as in previous studies our results presented in Table 2suggest that a major part of LGB + employees are not fully disclosed at work. Further, disclosure to friends was substantially associated with disclosure behavior at work, which aligns with the results by Griffith and Hebl (2002). In addition, our results replicate previous findings on the role of age and the presence of an intimate partner (Frohn et al., 2017; Pew Research Center, 2013; Ragins et al., 2007), suggesting that increasing age and the presence of an intimate partner increase the likelihood of disclosing a non-heterosexual orientation at the workplace. Moreover, our results suggest that the more time individuals spend at work, the more likely they are to disclose.

In contrast to previous results, our study showed no significant role of gender in disclosure at work (Frohn et al., 2017), which also applies to a possible role of interactions of gender and personality. However, although the findings of Frohn et al. (2017) for example suggests that lesbian woman are significantly more likely to disclose at work than gay employees, the effect size of this difference was only weak. Furthermore, Ragins et al. (2003) observed no significant direct effect of gender, which is consistent with the results of the present study. Nonetheless, gender may play a role in workplace disclosure behavior, depending on certain demographic, individual, or work-related variables (e.g., the type of work), which should be addressed in future studies (see e.g. Ragins et al., 2003). However, other factors previously observed to influence disclosure behavior, namely LGB + supportive relationships and work climate (Webster et al., 2018) also showed significant positive associations with disclosure behavior at a bivariate correlation level in our sample. Nonetheless, subsequent results from a hierarchical regression analysis did not provide incremental validity for these variables.

To our knowledge, there has been no study on the relationship between personality and disclosure behavior at work, so far. However, there is some knowledge on the association with personality variables and disclosure behavior in other areas of life. First of all, the results from the bivariate correlation analyses of the present study suggest that neuroticism, extraversion, conscientiousness, and internal LoC seem to be associated with disclosure behavior at work. Indeed, this would be in line with previous studies suggesting that general disclosure behavior is associated with lower anxiety and higher positive affectivity (Halpin & Allen, 2004). Since neuroticism is related to anxiety and extraversion to positive affectivity (Costa & McCrae, 1992; Schulte et al., 2004), our results on disclosing behavior at work may concur with these findings on general disclosing behavior. Previous results demonstrated that a higher internal LoC is associated with a better coping in stressful situations, whereas a higher external LoC is related to a lower ability to cope with stressful situations (Groth et al., 2019; Van der Horst et al., 2017). The present results might suggest that this would be also true for disclosure at work as a potentially stressful situation, because internal LoC was positively (and external LoC descriptively negatively) associated with disclosure at work.

It is important to note, however, that although neuroticism, extraversion, and internal LoC are significantly associated with disclosure behavior at work in the bivariate correlation analyses, these traits did not incrementally explain variance in our regression model. That is, these significant associations disappeared when controlling for the shared variance with the other (personality) factors potentially influencing disclosing behavior. Indeed, regression analyses or similar approaches were not used in the studies described in context of general disclosing behavior (e.g. Halpin & Allen, 2004; Peterson & Gerrity, 2006). Thus, we were able to support such previous results when a comparable method was used, but not when a more conservative approach focusing on incremental validity was applied.

The only personality trait that explained incremental variance in disclosure at work was conscientiousness. As far as we know previous research has not yet systematically examined the role of conscientiousness on disclosure behaviour at work. Our results imply that conscientiousness is significantly positively related to disclosure behavior at work. One might have expected that conscientiousness and disclosure at work would potentially be negatively associated because conscientious individuals seem to be less inclined to take career risk (e.g., Costa et al., 1995) and might be more afraid that disclosing a non-heterosexual orientation at work may negatively affect their career (Frohn et al., 2017). On the other hand, for example, Horn et al. (2004) showed that integrity and honesty are key characteristics of conscientious individuals, which also applies to authenticity (e.g., Wood et al., 2008). Therefore, it may be more likely that conscientious LGB + employees tend to disclose their sexual orientation at their workplace in order to be honest and authentic, possibly facilitated by the perception that disclosure is increasingly less likely to be associated with a direct career risk. These aspects should be further addressed in future work.

Agreeableness, openness, self-esteem, and impulsiveness were not significantly associated with disclosing a non-heterosexual orientation at work in our sample. Because compliancy and sociotropy are linked to agreeableness, a negative association between agreeableness and disclosure at work might have been conceivable (Bagby et al., 2001; Schulte et al., 2004). With regard to openness, as outlined above, a positive relationship with disclosure at work has been expected, primarily because open people are less inclined to follow social or cultural norms, authority, and tradition (Costa & McCrae, 1992; Schulte et al., 2004). So, it remains an open and interesting question why openness was not significantly associated with disclosing a non-heterosexual orientation at work. Future research may investigate whether and how the different facets of the agreeableness and openness constructs might still be related to disclosure behavior, because the BFI-K (Rammstedt & John, 2005) used in the present study only refers to the domains of the Five-Factor Model, which might be a limitation.

Previous research also suggested an association between self-esteem and general disclosure behavior (Halpin & Allen, 2004; Peterson & Gerrity, 2006) reported that self-esteem and LGB + identity development are moderately positively associated. For our specific workplace related disclosure variable (were general disclosure behavior was partialled out) we could not replicate such an association. However, future research may be more interested in examining the causal relationship between self-esteem and LGB + identity development and exploring the question of whether LGB + identity development has a positive impact on self-esteem and whether this positively influences LGB + employees’ disclosure behavior in the workplace.

There was also no significant association between impulsiveness and disclosure at work as measured with the BIS (Meule et al., 2011). The BIS covers aspects of non-planning, motor-, and attention-based impulsiveness. Although studies show that the BIS predicts risk-taking behavior in various contexts (Martins et al., 2004), it does not specifically include the risk-taking aspect of impulsiveness. Because disclosure at work may still be a risk in western countries (Frohn et al., 2017), future research should examine the relationship between disclosure of a non-heterosexual orientation at the workplace and risk-taking behavior using more specific measurements of the risk taking aspect of impulsivity.

A further potential limitation of our study is that our participants likely live under similar legal regulations on discrimination at the workplace. Thus, future research concerning the role of personality in disclosure at work may also address other countries or regions. Furthermore, our sample is relatively young on average, which might limit the generalizability of our results. Because LGB + employees may consider disclosure as a risk, as outlined above, future research may also investigate whether the relationship between personality traits and disclosure at work varies with the extent of anticipated negative consequences of disclosure.

To conclude, given that personality is a key factor to how we feel, think, and act, the present study examined the role of personality in disclosing a non-heterosexual orientation at the workplace, which, to our knowledge has not been done systematically so far. Correlation analyses indicated that some basic personality traits appear to play a role in workplace disclosure behavior. However, applying a hierarchical regression approach they did not significantly explain unique variance in disclosure at work, with the exception of conscientiousness. Because integrity, honesty, and authenticity are important characteristics of conscientiousness, this could explain the positive association with disclosing a non-heterosexual orientation at work and add to existing knowledge about disclosing behavior.

However, as outlined above, personality variables are thought to play a crucial role in shaping our thinking, feeling, and behavior. In this regard, it would have been expected that the overall influence of personality on workplace disclosure behavior would be greater. Interestingly, however, the proportion of additional explained variance of about 3% over and above contextual/situational variables, age, or having an intimate relationship is comparatively small, as shown by the adjusted R² in the hierarchical regression model. This result emerged after controlling for important sources of variance that have already been observed to substantially increase disclosure behavior at work. Nonetheless, the rather small but significant incremental effect of personality factors on disclosure behavior may also suggest that other contextual/situational variables and individual factors may play a critical role in disclosing a non-heterosexual orientation at work. Indeed, our study replicated previous findings regarding the presence of an intimate partner and age in incrementally predicting disclosure behavior. However, future studies could also focus on other individual, social, and workplace factors that may influence whether individuals disclose or conceal their sexual orientation, such as to avoid perceived or actual exclusion, marginalization, and career risks at the workplace.