Start Submission

Reading: Openness about Sexual Orientation and Exposure to Workplace Bullying


A- A+
Alt. Display

Original Article

Openness about Sexual Orientation and Exposure to Workplace Bullying


Helge Hoel ,

University of Manchester, GB
X close

Duncan Lewis,

Edith Cowan University, AU
X close

Anna Einarsdottir,

University of York, GB
X close

Guy Notelaers

University of Bergen, NO
X close


Previous studies of workplace bullying have not investigated whether Lesbian, Gay and Bisexual (LGB) employees experience bullying in similar or different ways to their heterosexual counterparts. This study reports on how and to what extent sexuality or sexual orientation influences the experience of workplace bullying and whether openness about sexual orientation elevates risks and shapes exposure to bullying. Using a large and rigorously compiled sample of the British working population comprising 500 non-heterosexuals and 722 heterosexuals (N = 1,222) and applying latent Class Cluster Analysis, a similar behavioural pattern of bullying for LGB employees emerged as for heterosexuals, although LGB employees were 1.34 times more likely to be bullied, and not being open about their sexual orientation elevated the risk of bullying. LGB employees were also more likely to be exposed to intrusive, sexualized behaviours and behaviours of an exclusionary nature. Altogether, this suggests that prejudices and stereotyping towards LGB people persist. Whilst being open about their sexual orientation did not make LGB people more likely to become a target of bullying as hypothesized, those who only reveal their sexual orientation when asked, were significantly more likely to be exposed to negative acts than those who were totally open. This indicates that non-disclosure does not prevent others at work making assumptions of sexual orientation, indicating that stereotyping of LGBs plays a greater part in disclosure than has previously been acknowledged.

How to Cite: Hoel, H., Lewis, D., Einarsdottir, A. and Notelaers, G., 2022. Openness about Sexual Orientation and Exposure to Workplace Bullying. Scandinavian Journal of Work and Organizational Psychology, 7(1), p.9. DOI:
  Published on 13 Jun 2022
 Accepted on 06 Apr 2022            Submitted on 19 Jul 2021


The global surge in interest in workplace bullying (e.g., D’Cruz, 2012; Einarsen, Hoel, Zapf & Cooper, 2020; Pinkos Cobb, 2017) can largely be explained by its perceived magnitude and impact, affecting targets (Nielsen & Einarsen, 2012), bystanders (e.g., Emdad, Alipour, Hagberg & Jensen, 2013), the organisation (Hoel, Cooper & Einarsen, 2020) and society (Di Martino, Hoel & Cooper, 2003). Workplace bullying is understood as persistent exposure to negative behaviours in situations where a power imbalance exists between the protagonists (Einarsen et al., 2020). Interlinked with establishing the prevalence of bullying as a rationale for organisational efforts to prevent and control it (e.g. Hoel, Cooper & Faragher, 2001; Lewis, Giga & Hoel, 2011; Lutgen-Sandvik, Tracy & Alberts, 2007; Zapf, Escartin, Scheppa-Lahyani, Einarsen, Hoel & Vartia, 2020), researchers have endeavoured to map those particularly at risk of bullying. Whilst the impact of gender has been investigated at length (e.g. Salin & Hoel, 2013; Simpson & Cohen, 2004), less attention has been afforded to ethnic minorities (Lewis & Gunn, 2007; Fox & Stallworth, 2005) and disabled employees (Fevre, Robinson, Lewis & Jones, 2013). Although research into the impact of sexual orientation and bullying is rare, evidence suggests that lesbians, gays and bisexuals (LGB) are especially vulnerable to workplace mistreatment of the type that underpins bullying (e.g., Fevre, Nichols, Prior & Rutherford, 2009; Grainger & Fitzner, 2007). This paucity of evidence is surprising, as we might expect constituents of an historically socially stigmatised minority (Ragins, 2004; 2008) such as LGBs would be more widely reported. Moreover, since non-heterosexuality is often described as an invisible or concealable social stigma (e.g., Clair, Beatty & MacLean, 2005; Ragins, 2004; Rumens & Broomfield, 2012), it is assumed that LGB employees would disclose their sexual orientation for it to become known among work colleagues. This could subsequently increase risks of adverse treatment including stigmatisation, homophobia and bullying (e.g., Acas, 2007; Ragins, 2004; Wax, Coletti & Ogaz 2018); despite the possible positive personal outcomes of removing social stigmatisation through acts of disclosure (Croteau, Anderson & VanderWal, 2008; Ragins, 2008; Wax et al., 2018).

Bullying research has been criticised for ‘theoretical under specification’ (Aquino and Thau, 2009), being more focused on providing evidence of the problem and its consequences (Einarsen et al., 2020; Ramsey, Troth & Branch, 2011). Responding to such criticism, we locate our investigation within social identity theory (e.g., Tajfel, 1972; Turner, Hogg, Oakes, Reicher & Whetherellm, 1987), emphasising the role of self-categorisation and intergroup social comparison and related notions of in-group and outgroup categorisation (Ashforth & Mael, 1989; Hogg & Terry, 2000). Yet, unlike Ramsey et al., (2011), who theorised bullying in terms of being victimised by a group, we apply these theories more broadly to account for bullying of LGB people whether targeted by an individual or a group.

The contribution of the article is thus threefold: First, it identifies how and the extent to which the experiences of LGB bullying differs to bullying of heterosexuals whilst at the same time providing reliable estimates of LGB bullying, which have hitherto been underreported. Second, with respect to disclosure of non-heterosexuality at work, the relationship between relative openness and adverse workplace treatment is explored. Whilst acknowledging negative factors associated with individual sexual orientation or identity, stereotypical images of homosexuality frequently appear to play central roles in scenarios of bullying of LGBs, despite largely being overlooked in research. Third, by locating our exploration within a social identity theoretical framework, we respond to previous generic criticisms directed at bullying research for being theory-light. Altogether, we approach the issues of design, execution and analysis in a manner combining scientific innovation and rigour with considerations of the considerable sensitivities required in studying sexual orientation and bullying in workplace settings.

We commence our article by reviewing literatures on both workplace bullying and sexual orientation leading to an exploration of our theoretical framework and its relevance to scenarios of bullying associated with non-heterosexuality. We then attend to disclosure of non-heterosexuality at work and how this links to experiences of bullying. Our analytic strategy deploys a latent class modelling method to investigate, and test identified hypotheses. Our discussion utilises the findings to highlight the article’s theoretical contributions, indicating both practical implications and methodological limitations.

Workplace bullying, sexual orientation and social identity theory

Since Leymann’s seminal study of workplace bullying in Sweden (Leymann, 1992), researchers have attempted to ascertain the levels and relative risks of exposure to bullying. Whilst some, like Leymann, have attempted to establish nationally representative estimates, (Nielsen et al., 2009) or approximations based upon large heterogeneous samples (Lutgen-Sandvik et al., 2007; Niedhammer, Chastang & David, 2007; Notelaers, Baillien, Vermunt, De Witte & Einarsen, 2011), others have focused attention on particular populations, including employment sectors, industries or occupations (see Zapf et al., 2020). However, estimates of bullying need to be understood in light of the measurement approach applied, as two measurement approaches dominate research approaches: namely the self-labelling and behavioural measurement methods (Hoel et al., 2001; Nielsen, Notelaers & Einarsen, 2020). Whilst self-labelling measures prevalence by providing participants with an accepted definition of bullying, behavioural measurement estimates bullying using an inventory of negative acts associated with bullying. These approaches can be deployed in tandem, with self-labelling validating the behavioural experience method findings and vice versa, thus providing an investigation of the behavioural nature of bullying experiences whilst recognising that perceptions of being a target of bullying is essential to the overall experience (Nielsen et al., 2020). Researchers have identified three types of bullying behaviours, such as work-related bullying; personal-related bullying; or social exclusion (Einarsen, Hoel & Notelaers, 2009; Nielsen et al., 2020; Notelaers, Van der Heijden, Hoel & Einarsen, 2018; Zapf et al., 2020). However, latent class analyses defeats such a dimensional approach arguing for a patterned approach (Keasley & Jagatic, 2011) that is more appropriate in understanding the phenomena and its behavioural expressions.

Acknowledging such differences in measurement methods and shortcomings with sampling in a review of published empirical studies, Nielsen et al. (2020) and Zapf et al. (2020) concluded that 3–4% of employees may experience serious bullying with negative encounters of a weekly or more frequent occurrence, while approximately 9–15% of the population would experience less severe or occasional bullying. Also, rates in the UK vary: from 10% (e.g., Cowie et al., 2000; Hoel, Cooper & Faragher, 2001) to 5% (Fevre et al., 2009) dependent upon method and sampling.

For risk of exposure, with some exceptions (e.g. Björkqvist et al., 1994; Rosander et al., 2020) most studies report little or no gender difference (Hoel et al., 2001; Zapf et al., 2020). However, where research has focused on other protected classes, including people from ethnic minorities and people with disabilities, they have often showed elevated risks of bullying (e.g., Fevre, Robinson, Lewis, & Jones., 2013; Grainger & Fitzner, 2007; Lewis & Gunn, 2007). The limited data available on the non-heterosexual working population paints a bleak picture of the realities facing LGB employees. A UK survey by Stonewall (2007) suggested that nearly 20% of lesbians and gay men had experienced some degree of homophobic bullying from colleagues. Similarly, according to a survey for the UK’s Equality and Human Rights Commission by Ellison & Gunstone (2009) a total of 39% of gay men, 31% of lesbians, and 11% and 16% of bisexual men and bisexual women respectively had experienced bullying, although these figures included experiences outside work.

Research on LGB people often appears to suffer from methodological shortcomings, particularly with respect to sampling (e.g., Croteau, 1996; Martin & Knox, 2000), typically relying on small, self-selected samples, often involving urban and “out” members of the LGB community (Lewis, Hoel & Einarsdottír, 2013) or according to Price (2011, p.15), over-represented by “younger, male, urban dwelling, white, middle-class participants”. When more robust sampling has been deployed using face-to-face interviews (N = 4,010), evidence pointed to LGB employees being 2.71 times more likely to report bullying in their workplaces than heterosexual respondents (Fevre, Nichols, Prior & Rutherford, 2009). However, this focus of the Fevre et al. (2009) study was on ill treatment at work and not sexuality and thus had a relatively small sample of LGB employees.

On this basis we put forward the following hypothesis:

Hypothesis 1 (H1): LGB employees will report more workplace bullying than non-LGB employees.

With regard to the behavioural experiences of bullying, previous studies of other protected groups including gender (e.g., Salin & Hoel, 2013; Simpson & Cohen, 2004) and disabled workers (Fevre et al., 2013; Fevre et al., 2009) report some discrepancies between sub- groups. Women, for example, are more exposed to social manipulation than men (Salin & Hoel; 2013; Salin, 2001), whilst disabled and chronically ill workers experience more physical violence than people without such conditions (Jones, Robinson, Fevre & Lewis, 2011). Similarly, ethnic minority groups report more personalised and offensive forms of bullying (Giga, Hoel & Lewis, 2008; Lewis & Gunn, 2007). As for LGB employees, the limited evidence available supports anecdotal evidence (e.g., Stonewall, 2007) where LGB people appear to be exposed to different behaviours than heterosexual colleagues, being particularly vulnerable to exclusionary and disrespectful acts from co-workers and supervisors (Stonewall, 2007; Minton, Dahl, O’Moore & Tuck, 2008). Similarly, data from UK Employment Tribunals (labour courts) into discrimination and unfair dismissal reveals that LGB claimants are particularly vulnerable to sexualised practical jokes and intrusive sexualised behaviour as well as acts of homophobia (Acas, 2007) (homophobia was exemplified by threats, physical abuse and humiliating acts such as being spat at, as well as social exclusion through numerous means). Altogether, and supported by evidence emerging from organisational case studies (Colgan, Wright, Creegan & McKearney, 2009; Ward & Winstanley, 2006), it appears that LGB employees’ negative workplace experiences diverge from heterosexuals.

To account theoretically for LGB employees’ negative workplace experiences, Social Identity Theory (SIT), which can be seen as “an interface between psychological and societal explanations for prejudice and discrimination” (Brewer & Miller, 1984, p. 282), and its more recent extended variant, self-categorisation theory (Ashforth & Mael, 1989), may offer valuable guidance. SIT and self-categorisation theory explain how individual self-esteem is achieved through social comparison of groups, generating the idea of in-groups (or put simply – us) and out-groups (them), where members of the ingroup and their characteristics are assessed favourably over members of the outgroup (Ashforth & Mael, 1989; Tajfel, 1972; Turner et al., 1987). The social categories in question that may be cognitively activated and used to identify with a preferred group depend upon what appears to be important in the given context/situation (category salience) including job type, professional or departmental affiliation, hierarchical status and demographic group membership (Ramsey et al., 2011), including gender and sexual orientation. Furthermore, according to Hogg and Terry (2000), central to social identity dynamics are notions of group prototypicality and “depersonalisation”. Here, prototypicality or group prototype refers to features of group membership associated with exemplary group members (“ideal types”) who best represent the group in terms of perceptions, behaviours, feelings and values. Through categorising of self and others into ingroup and outgroup, similarity and difference are emphasised between respective prototypes generating a series of outcomes including stereotypes, behavioural norms, attitudes and group cohesion (Hogg & Terry, 2000). Consequently, the uniqueness of individuals within the respective group disappears, with group members becoming either interchangeable (Wenzel, Mumandey & Walduz, 2007) or simply emerging as “embodiment of the relevant prototype” (Hogg & Terry, 2000, p. 123). Group members not conforming with prototype may be considered deviant and rejected from the group, especially in situations when they are perceived to bear the marks of a salient outgroup, because marginality could be seen to undermine “the distinctiveness and prototypical clarity and integrity of the group” (Hogg & Terry, 2000, p. 127). This represents a particular problem for what Brewer and Miller (1984) refer to as “clear minorities”, for example, being the only black person or “out” LGB person in the group.

Mostly, ingroup identification is seen to be relatively harmless, providing there are acts of compliance with group norms, while the outgroup is viewed with indifference or suspicion. Differentiation to the outgroup can be satisfied by making clear distinctions between the groups (Brewer, 1999), but where distrust to the outgroup is emphasised, and the outgroup is portrayed as second-rate or inferior to the ingroup, particularly with respect to those aspects of identity being compared (Ullrich, 2009), this may trigger aggression, including bullying of atypical members who are seen as breaching the group’s normative attributes (Ramsey et al., 2011). This is particularly the case when aggression can be justified through what Brewer (1999) refers to as “moral authority” when, for example, behaviour breaches religious or moral beliefs. Thus, where the moral order is seen as absolute rather than relative, for example when any sexual orientation other than heterosexuality is considered deviant, moral superiority becomes incompatible with tolerance for difference. This might result in denigration and contempt as an outcome, particularly when the outgroup fails to observe or subscribe to dominant moral codes. Whilst this may give rise to derogatory behaviour, it is argued that contempt is more likely to be associated with avoidance rather than outright hostility, hence leading to segregation and social exclusion (Brewer, 1999).

This above empirical evidence and theoretical discussion give rise to the following hypothesis:

Hypothesis 2 (H2): The behavioural nature of the bullying experience of LGB individuals will differ from non-LGB individuals.

LGB employees’ experiences of bullying must nevertheless be considered in connection with their relative openness about their sexual orientation at work, and to what extent they believe that work colleagues know of their sexual orientation. It is to these issues that we now turn.

Workplace bullying and degree of openness about sexuality at work

With some notable exceptions (e.g., Einarsdottír, Hoel & Lewis, 2015; Froyum, 2007; Rumens & Broomfield, 2012), research portrays non-heterosexuality as an invisible entity and, therefore, needs to be disclosed to become known by others (Clair et al., 2005; Ragins, 2004; Tilcsik, Antby & Knight, 2015). By contrast, for most heterosexuals, disclosure is a non-issue as their sexuality is seen or perceived as given (Röndahl, Innala & Carlsson, 2007; Ward & Winstanley, 2003) and therefore not questioned, as it is considered the norm (Ng and Rumens, 2017), as accurately articulated in the concept of heteronormativity (Jackson, 2006). Furthermore, for LBG employees, disclosure is increasingly seen as an ongoing and repeated process rather than a single event, and is largely considered as being under the control of the individual (Croteau et al., 2008; Ragins, 2008). Decisions about disclosure are seen as strategic choices, albeit not necessarily planned (Colgan, Creegan, McKearney & Wright, 2008), as one may be left to respond to colleagues’ personal queries including questions about partners or family arrangements (Bowring & Brewis, 2009). Furthermore, in some circumstances, the control over the process is entirely taken away from the LGB person, where an individual’s non-heterosexuality is made public against their own will (for example, being “outed”) (e.g., Ragins, 2004).

Disclosure decisions are seen as products of conscious cost-benefit evaluations (Clair et al., 2005; Ragins, 2004) in which LGB people assess the pros and cons of being open about their sexual orientation to those around them. According to Tilcsik et al., (2015), the awareness of the need to navigate potentially hazardous social situations effectively, often from adolescence, and concealing sexual orientation if necessary, has contributed to making LGB people more socially perceptive. Fear of bullying, violence and discrimination at work and outside it would, in this respect, be factors influencing LGB employees’ strategies and decisions as to whether or not to “come out” and affect their overall level or degree of personal disclosure (Day & Schoenrade, 1997; Ward & Winstanley, 2006). In this respect, disclosure has been described as a double-edged sword (Day & Schoenrade, 1997). Thus, openness may be both risky and emotionally costly (Wax et al., 2018), whilst hiding or “staying in the closet” may restrict opportunities to socially integrate, thereby limiting access to valuable information, which may potentially negatively impact career progression (Griffith & Hebl, 2002). Concealing sexual orientation can, in its own right, place a psychological strain on individuals leading to stress-related illness (Meyer, 2003; Ragins, 2008). Therefore, any potential personal gains emerging from disclosure must be considered against likely adverse reactions.

Returning to our theoretical line of reasoning, openness about non-heterosexuality would make sexual orientation as a social category more salient, thus increasing the opportunity for bringing it into conflict with the in-group’s ideal type (in most cases heterosexuality). One would envisage that this would increase the chance of becoming a target of intimidating and exclusionary acts and responses, even where the risk is deemed to be acceptable for disclosure still to occur.

Hypothesis 3 (H3): The more open LGBs are about their sexual orientation the greater the risk of exposure to bullying.


Sample and procedure

To achieve a statistically viable sample we aimed to recruit 500 LGB employees and an equivalent number of heterosexual employees working in British workplaces or who had been in employment within the last six months.

We adopted a face-to-face structured interview approach using CAPI (Computer-Assisted Personal Interviewing) to interview workers at their home residences, replicating the approach taken by studies into sensitive workplace issues such as harassment (Grainger & Fitzner, 2007; Fevre et al., 2009) and workplace ill treatment (Fevre, Lewis & Jones, 2012). CAPI systems prevent the researcher from seeing or accessing the respondent’s answers to the sexual orientation questions provided on the screen thus ensuring privacy for LGB participants. To obtain our sample we used an Omnibus Survey and a quota sampling strategy. Interviewees were selected from a representative sample of around 4,000 adults per week (two waves of 2,000 respondents). To reach the target of 500 LGB employees and 500 heterosexual respondents, the fieldwork was conducted across 44 waves, thus taking approximately six months to achieve the LGB sample. Key screening criteria for participation were current employment or had been employed within the last six months.

The final sample included 712 heterosexuals (353 men and 369 women) and 500 non-heterosexuals: 147 gay men, 122 lesbians, 151 bisexuals, of whom 40 were men and 111 women. A total of 56 respondents labelled themselves as “Unsure” (31 men and 25 women) and as “Other sexuality/sexual orientation” (9 men and 15 women). The categories “Unsure” and “Other” were excluded from the analysis as we were unable to relate respondents within these categories to some of the questions regarding disclosure and openness about sexuality.

The age distribution was as follows: 16.5% were between 16–24 years of age, 34.5% between 25–39, 33.2% between 40–54, 13.2% between 55–64 and finally 2.5% was 65 years or older. The distribution of social grade/class was as follows: 24.1% belonged to social grade AB (upper class and middle class), 34.1% with C1 (lower-middle class), 22.7% with C2 (skilled working class) and finally 19.2% with DE (semi-skilled and unskilled manual workers). Approximately 90% was white and 4.8% reported some form of disability.


To measure bullying, we combined behavioural and self-labelling methods (see Leon-Perez et al., 2014): by presenting respondents with a common definition of bullying and measuring behavioural experiences using the shortened version (9 items) of the Negative Acts Questionnaire (S-NAQ) (Notelaers, Van der Heijden, Hoel & Einarsen, 2019) and a further four items emerging from a review of the literature (e.g., ACAS, 2007; Colgan et al., 2008; Griffith & Hebl, 2002; Minton et al., 2008; Ragins & Wiethoff, 2003; Stonewall, 2007; Williams & Tregidga, 2014), a total of 13 items. The additional items were: “Being confronted with unwanted jokes or remarks which have a sexual undertone”; “Receiving unwelcome comments about the way you dress”; “Experiencing unwanted physical contact, e.g., touching, grabbing, groping”. The following response scale was applied: “never”; “occasionally”; “monthly”; “weekly”; and “daily”.

Disclosure was measured with a single question: “How open are/were you about your sexuality in your current/most recent job?” Response categories were “I give the impression that I am heterosexual”, “I am not open at all”, “I only reveal my sexuality/sexual orientation if asked”, “I avoid drawing attention to my sexuality/sexual orientation”, “I make no secret about my sexuality/sexual orientation” and “I am totally open”. Note that questions about openness about sexual orientation were only answered by non-heterosexual respondents, that is everyone who did not identify as heterosexual or straight.

Ethical approval

Full ethical approval of the study, including approval of the research instrument (questionnaire) and strategy for participant recruitment was obtained by the University’s Ethics Committee.

Analytic strategy

The analytic strategy here is rather complex because we first have to explain why we use latent class modelling as well as explain how this statistical technique works. Next, we have to make clear how we test whether the bullying experience is different for LGBs and finally we need to outline how we test whether the risk of LGB, older employees, disability and “disclosing” is higher or lower.

a) Why latent class modelling

Scholars suggest that the complex and dynamic nature of bullying makes the case for the use of latent class modelling (LC) (Nielsen et al., 2020; Notelaers & van der Heijden, 2021). In this way not only the complex nature of the concept but also the strong violations of the distributional assumption, for example, normality, and the fact that these measure are in reality employing a categorical response set (see: Hershcovis & Reich, 2013), are being addressed (Notelaers & van der Heijden, 2021).

LC is a statistical method that classifies respondents into mutually exclusive groups with respect to a not directly observed (latent) trait (e.g., bullying) (Notelaers et al., 2006). The LC analysis starts with the assumption that there is only one group, and subsequently estimates two (e.g., not bullied/bullied), three, four … and finally n different classes, until an LC model is found that statistically fits the data best (Magidson & Vermunt, 2004). An important difference from traditional cluster methods (such as K-means clustering) is that LC analysis is based on a statistical model that can be tested (Magidson & Vermunt, 2002). As a consequence, determining the number of latent classes is less arbitrary than when using traditional cluster methods. Hence, this method allows for empirically testing whether different target groups exist, based on the responses to an inventory measuring exposure to different kinds of bullying behaviours (Notelaers et al., 2006). The metric of a single latent variable is typically nominal. The Bayesian Information Criterion (BIC) is most often used for model selection (Magidson & Vermunt, 2002; 2004) that is to determine the number of latent class clusters. McCutcheon (1987) and Hagenaars (1990) suggested accepting the model with the lowest BIC because the models are non-nested. Next to this test, we also report descriptive fit measures. We assess how well the clusters are separated by inspecting the total rate of classification errors due to adjacent erroneous classification. Finally, we also inspect local fit, that is how well the model described the initial association between the 13 indicators by comparing the total amount of bivariate residuals of the 1-profile model with that of the final model. In general, the bivariate residuals (BVRs) should be lower than, or equal to, 3.84 (Vermunt, 2013). For readers who are not familiar with LC modelling, these residuals are comparable to the residuals or associations that remain after having modelled a factor structure which is meant to account for the bivariate correlations between the indicators. The latter, 3.84, may be relaxed because the L2 that follows a χ2 distribution is quite sensitive to large sample size (Paas, 2014). Taking into account previous applications to workplace bullying, the reduction in BVR should be at least 85% (Einarsen et al., 2009; Leon-Perez et al., 2014; Notelaers et al., 2011).

b) Testing for equivalence of measurement invariance

Before testing hypothesis 1 (H1) that entails comparing heterosexuals’ and LGB employees’ exposure to bullying in terms of risk, it is critical to establish whether their experience of bullying is similar. Hence, to discern hypothesis 2 (H2) stating that the behavioural nature of the bullying experience of LGB individuals will differ from non-LGB individuals, the level of equivalence, that is, the extent that the measurement is similar (invariant) for both groups must be discerned. Earlier, Clogg and Goodman (1985, 1986) used a multiple-group analysis to inspect whether the measurement instrument differs across groups. The backward elimination of differences may be considered as a conservative modelling strategy as it starts from a complete heterogeneous model and eliminates the differences or fixes parameters step by step (Eid & Diener, 2001). The forward inclusion of differences strategy starts with the complete homogeneous model (pooled dataset). This model assumes no impact upon the measurement model from the grouping variable (heterosexual/LGB). This means that the measurement of bullying at work is assumed to be the same across groups. In subsequent steps these assumptions are relaxed (parameters are freed). We employ the forward inclusion of differences strategy (Chegeni et al., 2021; Hagenaars, 1990; Kankaras et al., 2010). Before introducing the group variable, a model is estimated irrespective of the grouping variable (complete homogeneity; cf. model 0) (see Kankaras, et al., 2010). Next, the group variable is introduced with the assumption that there are no direct relationships between the group variable and indicators of the measurement model. This means that the latent trait (bullying) fully mediates the relationship between the groups (heterosexual/LGB) and negative behaviours (indicators) that are meant to measure bullying (structural homogeneity; cf. models 1–5). In the following, the remaining relationships between the grouping variable and the indicators are inspected. When these are higher than 3.84 this may indicate that there is a direct association between the grouping variable and indicator. This means that the relationship between the latent variable and the indicators can differ across groups (partial homogeneity; cf. models 6–9). Next, any interactions between latent variable and grouping variable on the indicators are estimated (no homogeneity or heterogeneity; cf. model 10). Finally, we estimate the full heterogeneous model (model 11) by performing a LCC analysis in both the sample of heterosexuals and the sample of LGBs. When these models are nested, the difference in L2 and respective degrees of freedom are employed to tax the extent of homogeneity of heterogeneity; otherwise the BIC is employed. The latter penalises for the number of parameters used. As a consequence, BIC prefers more parsimonious models.

Previous research (Zapf et al., 2011) has shown significant relationships between socio-demographical variables and workplace bullying such as age, gender and occupational status. Therefore, we controlled for gender, age, and social grade/class. In addition, disability was taken into account to prevent possible differences between heterosexuals and LGB employees with regard to the experience of workplace bullying attributed to them.

c) Risk groups, 3-step LCA with covariates

To test the hypothesis 1 (H1), whether LGB employees have a higher risk of being bullied, we conducted a STEP3 latent class analysis. We examined whether co-variates such as disability, age, socio-economic class (income), heterosexual/LGB (sexual orientation), gender, occupational position, educational level and nature of contract (full-time/part-time) were related to the 3 latent classes. Because we have covariates in the STEP3 model we used ML estimation. For the ease of interpretation, we used dummy coding using the first category as the reference category. This means that for the sometimes bullied and target of bullying clusters, the not bullied cluster operates as a reference category. The logits and the relative risk ratios Exp(b) portray relationships between categories of the co-variate and LC clusters. They are similar to a multiple comparison procedure in a traditional analysis of variance.

To test the hypothesis relating to disclosure and bullying (H3), we introduced the disclosure item as a co-variate of the STEP 3 LC model in addition to the socio demographic variables.


The results of the analysis in Latent Gold 4.5 (Vermunt & Magidson, 2009) are summarised in the fit statistics that are portrayed in Table 1.

Table 1

Fit statistics.


Complete homogeneity 0 3 11062.1 6225.39 1029 0.0462

Structural homogeneity 1 1 13520.5 9549.139 1073 0

2 2 11348.41 7215.462 1050 0.0233

3 3 11062.2 6767.661 1027 0.0466

4 4 11094.22 6638.093 1004 0.0678

5 5 11122.05 6504.343 981 0.0716

Partial homogeneity 6 3 11064.72 6763.163 1026 0.0468

7 3 11064.36 6762.795 1026 0.0467

8 3 11065.36 6763.796 1026 0.0468

9 3 11067.9 6766.334 1026 0.0466

Heterogeneity 10 3 11422.0 7366.19 1021 0.0477

Full heterogeneity 11 3 6515 948

Legend: BIC: Bayesian Information Criterion; Model 1–5: Traditional LC models (structural homogeneity); Models 6–9: Testing invariance or heterogeneity across heterosexuals and LGBs; Model 6: h/LGBLGB with dress, Model 7: h/LGBLGB ignored; Model 8: h/LGB rem sex and Model 9: h/LGB unwanted contact. Model 10: Heterogeneous model (interaction term between LC and covariate). Model 11: Full homogeneous model (based on the merged data). Model 12: Full heterogeneous model (based on separate analysis of heterosexuals and LGBs).

Before testing the level of heterogeneity, we note that BIC of both the full homogeneous (model 0), the structural homogenous model (models 1–5) and the full heterogeneous model (11) was the lowest when extracting 3 latent classes. Hence, the statistical fit is best when extracting 3 classes – to illustrate the iterative procedure we only portray the fit of the different LC models for the structural homogeneous model (model 1–5). Following the forward inclusion of difference, we see that the homogeneous model (0) fits better than the structural homogeneous model (3) as L2 does not decrease significantly. But the BIC of both is almost the same. Because both models are not nested, we must rely on the BIC and conclude that both are equally well fitting. Given the main focus of the article, we prefer model 3 because it accounts for sexual orientation. Allowing for more heterogeneity by adding direct relationship between sexual orientation and items for which the BVR > 3.84 and BVR were larger than 15% did not lead to improvement of fit because L2 did not decrease significantly. The negative behaviours: “Receiving unwelcome comments about the way you dress”, “Being confronted with unwanted jokes or remarks which have a sexual undertone”, “Experiencing unwanted physical contact, e.g., touching, grabbing, groping” and “Being ignored by people at work”, of which the three first relate to sexual orientation, seemed, therefore, to not function differently for heterosexuals and LGB people. If the L2 and BIC of these models had been significantly lower than that of the homogeneous models, this would have suggested that the experience of bullying was strictly speaking not comparable between both groups (partial homogeneity). As said, the models six to nine led to a worse fit. Note also that the BIC statistic yielded that allowing such a co-variation was associated with a deterioration of fit because BIC increased compared to model 3. Thus, from a statistical point of view, the differences between LGB’s and non-LGBs with respect to these items was not meaningful as they did not significantly improve fit. Note also that assuming more heterogeneity did not improve fit. In conclusion, we must reject the second hypothesis stated that the experience of bullying would be qualitatively different for LGB and heterosexual employees.

Next, we can describe the latent class clusters. The three different latent class clusters are portrayed in Figure 1.

Profile of workplace bullying. Average score on negative behaviours for each latent class. Dress: “Receiving unwelcome comments about the way you dress”. Remsex: “Being confronted with unwanted jokes or remarks which have a sexual undertone”. Unwphcon: “Experiencing unwanted physical contact, e.g. touching, grabbing, groping”. Gossip: “Spreading gossip and rumours about you”. Shouted at: Being shouted at“. Insulted: “Being insulted or having offensive remarks made about you (i.e. about habits and background, attitude or private life, etc)”. Withinfo: “Someone withholding information which affects your performance”. Remerror: “Receiving repeated reminders of your errors or mistakes”. Creffort: “Persistent criticism of your work or performance”.Ignored: “Being ignored by people at work”. Approach: “Facing a hostile reaction when you approach others. Jokes: Being the subject of unwanted practical jokes”. Self-label: “Using the definition above, have you been bullied at work over the last six months?”
Figure 1 

Profile of workplace bullying. Average score on negative behaviours for each latent class. Dress: “Receiving unwelcome comments about the way you dress”. Remsex: “Being confronted with unwanted jokes or remarks which have a sexual undertone”. Unwphcon: “Experiencing unwanted physical contact, e.g. touching, grabbing, groping”. Gossip: “Spreading gossip and rumours about you”. Shouted at: Being shouted at“. Insulted: “Being insulted or having offensive remarks made about you (i.e. about habits and background, attitude or private life, etc)”. Withinfo: “Someone withholding information which affects your performance”. Remerror: “Receiving repeated reminders of your errors or mistakes”. Creffort: “Persistent criticism of your work or performance”.Ignored: “Being ignored by people at work”. Approach: “Facing a hostile reaction when you approach others. Jokes: Being the subject of unwanted practical jokes”. Self-label: “Using the definition above, have you been bullied at work over the last six months?”.

Figure 1 envisages the profile plot in Latent Gold 4.5. On the Y-axis the conditional average scores – these are the average of each item given the latent class cluster – are given, while on the X-axis the negative behaviours are printed. This plot is commonly used to portray the relationship between indicators and the latent variable (LC clusters). The plot shows three distinct lines that portray the different clusters. At the bottom of the figure, the respondents show on average 1.02 which corresponds with the “never” response category. Some 69.7% of the respondents are classified in the “not bullied” latent cluster. The line just above “not bullied” has an average score that varies between 1.15 and 1.55, thus situated between “never” and “sometimes”. Accordingly, we label this latent class cluster as “occasional bullying”. Approximately 24.1% of the respondents are, according to the latent class model, occasionally a target of bullying. The final cluster at the top of the figure represents the targets of workplace bullying who have the greatest probability for the highest average response. Approximately 6.1% of the respondents in our sample are, according to the latent class model, targets of workplace bullying.

Sexuality and risk of workplace bullying

To test the first hypothesis, we relied upon the Wald statistic, which expresses whether there is a significant relationship between the co-variates and the LC clusters. The Wald statistic showed that neither gender, social grade/class, occupational position, educational level nor type of contract (full-time/part-time) were significantly related to three bullying classes (See Table 2). However, disability, age, and sexual orientation were related. The relative risk for LBGs of 3.32 indicates, that compared to heterosexuals, LGBs were 3.32 more likely to be a target of bullying as compared to not being a target of bullying, which is in line with our first hypothesis (H1). We further note, that LGBs are also more likely to be occasionally bullied (1.27 times more likely than heterosexuals). The relative risk ratios found for disability and age were all noteworthy. Compared to respondents without a disability, disabled employees are 4.54 times more likely to be a target of bullying than not being a target of bullying. As far as age is concerned, compared to respondents under the age of 25, respondents between the age of 40 and 64 are far less likely (5 to 10 times) to be a target of bullying, thereby indicating that younger respondents were especially vulnerable.

Table 2

Risk groups of workplace bullying.



Yes/No 0.75** 2.12 1.514*** 4.54 26.87***


25–39/16–24 0.122 1.13 –0.614 0.54 26.32***

40–54/16–24 –0.53** 0.59 –1.513** 0.22

55–64/16–24 –0.5* 0.61 –2.291* 0.1

65+/16–24 –0.76 0.47 –0.614 0.54


C1/AB –0.53* 0.59 0.403 1.5 11.12

C2/AB –0.28 0.76 1.133 3.11

DE/AB –0.64 0.53 0.780 2.18

Female/Male 0.054 1.06 0.551 1.74 2.752

LGB/Heterosexual 0.236** 1.27 1.199** 3.32 13.06**

Intermediate occupations/Managerial and professional occupations –0.75* 0.47 –0.274 0.76 10.05

Routine and manual occupations/Managerial and professional occupations –0.55 0.58 –0.369 0.69

Not stated-not classifiable/Managerial and professional occupations –0.63 0.53 –1.016 0.36

Undergrad/Higher or postgrad 0.189 1.21 –0.445 0.64 9.033

HE quals/Higher or postgrad –0.28 0.76 0.061 1.06

A/AS Levels/Higher or postgrad 0.114 1.12 –0.371 0.69

GCSE (A-C)/Higher or postgrad 0.032 1.03 0.281 1.32

GCSE (D-G)/Higher or postgrad –0.01 0.99 –0.91 0.4

Other/Higher or postgrad 0.238 1.27 –0.778 0.46

None/Higher or postgrad 0.398 1.49 –1.63 0.2

Part time/Full time 0.006 1.01 –0.35 0.7 1.007

Legend *: 0.01< = p < 0.5; **: 0,001 < = p < 0,01 and ***: p < 0,001.

Disclosure and bullying

To test the hypothesis relating to disclosure and bullying, we inspected the Wald statistic. The Wald statistic of 9,095 was not significant. Hence, the degree of openness was not related to the risk of being bullied. We therefore reject the third hypothesis (H3). However, when inspecting further the multiple group comparisons, we note one significant difference, that is, between being totally open and “I reveal my sexuality/sexual orientation only when asked”. The latter were 1.7 times more likely to be occasionally exposed to negative behaviours than not to be exposed, compared to the LGSs indicating that they were “totally open” about their sexual orientation.


This study explores the rarely investigated influence of sexual orientation on the experience of workplace bullying by means of a robust, albeit ethically sensitive methodology of sampling, data-collection, sample size and application of valid measures. It is the first study that investigates whether the nature of negative behaviour experienced, and the form that bullying takes, are similar, or indeed different, for LGB employees than heterosexuals.

As the first study of its kind, it makes several contributions to the burgeoning literature on workplace bullying. As expected, in terms of the nature of bullying we found that LGB people were particularly prone to experiencing sexualized, intrusive behaviour, such as “Being confronted with unwanted jokes or remarks which have a sexual undertone” and “Experiencing unwanted physical contact, including touching, grabbing and groping”. The presence of such sexualized behaviour could demonstrate homophobia and hostility to homosexuality (Herek & McLemore, 2013). The results also suggest LGB people are being over-sexualized, stereotyped (Hogg & Terry, 2000) and defined in terms of their sexual orientation alone, implying some employees think it is more “acceptable” to display sexualized behaviour towards LGB people.

However, being particularly vulnerable to exposure to sexualized negative acts does not in itself necessarily imply that the experience of bullying is qualitatively different for LGBs than for heterosexual employees. In fact, based on invariance testing in a latent class modelling framework, we did not find a separate latent cluster which consists of such sexualized intrusive behaviours. Instead, we found three clusters that appear to have the same meaning for heterosexuals and LGB employees. This suggests that heterosexuals who are the targets of bullying may also face sexualized behaviours, perhaps women in particular, and that LGB targets similarly face negative behaviours other than sexualized acts. This finding also corresponds with the theoretical notion of the process of workplace bullying, which results in targets facing the highest frequency of exposure to the greatest number of negative acts (Björkqvist, Österman, & Hjelt-Bäck, 1994; Einarsen et al., 2011). Hence, our study rejects hypothesis 2, which predicted that the bullying phenomenon would look different for LGBs and heterosexual employees.

Following application of a latent class cluster analysis (Nielsen et al, 2020), we concluded that 6.1% of respondents were targets of workplace bullying, with a further 24.1% being exposed to bullying acts occasionally, both figures being largely in line with previous studies. A bullying prevalence rate of 6.1% is somewhat higher than that found in previous large UK samples applying a latent cluster (LC) approach (Einarsen, Hoel, & Notelaers, 2009). However, the latent profile plot (cf. Figure 1) shows that in comparison with previous studies, applying an LC approach to distinguish different types of target groups with respect to workplace bullying, the target group in this study reports less frequent exposure to negative acts. Indeed, the conditional probabilities to respond “weekly” and “daily” in terms of frequency of exposure appear to be somewhat lower in this sample. Yet, targets had undisputedly the highest probability of reporting negative behaviours and labelling themselves as victims given the definition of bullying applied. Several methodological reasons may explain this finding including Leymann’s (1992) recommendation to measure bullying by means of a structured face-to-face interview design, a strategy adopted in this study. This may suggest that respondents respond more conservatively than when left to complete a survey in private. It is also possible that targets, given the emotional turmoil they face, could be more likely to decline to participate when asked to participate in a study about their workplace experiences. Furthermore, in terms of risk of exposure to bullying, our study confirms previous findings, suggesting that sexual orientation does indeed influence bullying, with LGBs 3.32 times more likely to be a target of bullying than heterosexuals. Noteworthy as well, LGBs were 1.27 times more likely to be occasionally bullied than heterosexuals, thereby altogether confirming hypothesis 1.

The reported findings pointing to an elevated risk of bullying for those identifying as LGBs and is in line with self–categorization theory (Ashfort & Mael, 1989) where members of outgroups are considered deviant or inferior to the ingroup (Hogg & Terry, 2000). In this case non-heterosexuals experienced enhanced risk of exposure to negative behaviour. Equally, by directing sexualised behavior at LGBs, irrespective of any personal sexual interest, members of the in-group, most likely to be heterosexuals, could be seen to denigrate non-heterosexual orientations per se, thus reinforcing the perceived superiority of heterosexuality (Herek & McLemore, 2013) and assumed inferiority of non-heterosexuals (Lewis, Glambek & Hoel, 2020).

Our hypothesis predicting that those most open about their non-heterosexual orientation would be of greatest risk of bullying was rejected. This suggests that actively disclosing non-heterosexuality at work in most cases does not carry any additional risk of bullying. This also implies that LGBs generally are able to accurately assess whether disclosing is safe in terms of any negative repercussions and their own abilities of dealing with such responses. Nevertheless, the fact that those who “only reveal their sexual orientation when being asked” were significantly more likely to be exposed to negative behaviours than those who were “totally open”, the group we hypothesised would be most at risk, is intriguing and worthy of comment.

As decisions about disclosing a non-heterosexual orientation imply cost/benefit considerations (e.g., Clair et al., 2005), this seems to suggest that those who only reveal their sexual orientation when confronted appear to take a dimmer view of the situation and/or their personal ability or self-efficacy (Lidderdale et al., 2007) in responding to potential negative outcomes. Given the premise of heteronormativity (Jackson, 2006), with heterosexuality taken as given if otherwise not stated, the fact that these respondents are being asked about their sexual orientation seems to question the widespread assumption that homosexuality is by definition invisible (e.g., Ragins, 2008). As argued earlier, it is possible that despite efforts by individuals to keep their sexual orientation secret, and notwithstanding their possible enhanced levels of social perceptiveness (Tilcsik et al., 2015), colleagues may still assume non-heterosexuality and continue to guess, query and potentially pester LGB employees about their sexual orientation. Returning to theories of social categorisation and group prototypicality, perceived deviant appearance or behaviour could easily make expectations about sexual orientation and gender role salient categories, igniting negative feelings, thoughts and exclusionary behaviour from other in-group members against the perceived deviant group member, thus undermining the integrity of the group and how it sees itself, leading to stigmatization and rejection (Di Marco, Hoel & Lewis, 2021), although not necessarily taken to the stage of bullying.

Whilst not contradicting the fact that in many cases, gay men, lesbians and bisexuals can hide their sexual orientation and disclose at will, for those LGB employees who fit stereotypical assumptions about appearance, dress or gestures, the management of their sexual orientation might be compromised and not entirely be a matter of their own choice (Einarsdóttir et al., 2016) as our findings indicate. Theoretically, our findings indirectly therefore question a key premise in identity management literature, where disclosure is seen as a conscious process under control of the discloser (e.g., Croteau et al., 2008; Lidderdale et al., 2007).

Practical implications

Researchers of workplace bullying can draw confidence from the fact that frequently applied research instrument measures such as the Negative Acts Questionnaire, here applied in its short version (SNAQ) (Notelaers & Einarsen, 2008), appears to effectively discriminate between targets and non-targets of bullying among minority populations, in this case LGB employees. However, in order to be able to detect sexualized bullying where LGB employees are the focus of enquiry, one should, in the future, include additional behavioural items measuring sexualized negative acts to which LGB people, and indeed women, appear to be exposed more frequently than others.

The fact that more than 6% of employees emerged as targets of severe or regular bullying suggests that policies on bullying common among large and medium-sized UK organizations (CIPD, 2005) are not as effective as intended. Moreover, with LGB people seemingly facing bullying at work to a considerably greater extent (11%) than heterosexual employees (3%), organisations need to pay specific attention to sexual orientation and the needs of LGB employees as an equal opportunity/diversity priority. Addressing this in policies and within diversity awareness training, the unacceptability of sexualized behaviour directed at LGB employees, or indeed heterosexual employees is demonstrated as important. The fact that disabled and young employees report much elevated levels of exposure to bullying also suggests a need for further investigation and indeed attention with respect to practical intervention at the level of the organization.

Whilst many lesbians, gay men and bisexuals feel they can disclose their sexual orientation at work, others feel they need to hide it. As pointed out above, respect for sexual minorities should also mean that for those who, for whatever reason, do not want to reveal their sexual orientation, their privacy should be respected, and those reluctant to draw attention to their sexuality in the workplace should subsequently not be exposed to unwanted questioning or interrogation. Hence, organisations should adopt a more diverse and inclusive approach, stimulating the creation of a culture where people feel safe to disclose without repercussions, and where respect for minorities of all status, including LGB people, is embedded. To that end, organisations could foster high-profile LGB role models, challenging normative heterosexuality (heteronormativity) (e.g., Jacksen, 2006) and signalling that heterosexuality is not a precondition for a successful career trajectory.

Directions for future research

For the purposes of this article, we treated LGBs as a unified group. However, recent research into discrimination (e.g., McLaughlin, Hatzenbuehler & Keyes, 2010) has drawn attention to issues of intersectionality or inter-relationship of multiple identities in the work context (McBride, Hebson & Holgate, 2015), and we would argue that it makes sense to examine bullying in the same way – namely, to examine how bullying may affect the intersection of gender and sexual orientation and explore to what extent the experience of bullying differs between lesbians, gay men and for bisexual women and men respectively. Future studies need to account for such potential differences, bearing in mind that in terms of LGB employees, such potential differences are unlikely to be limited to the experience of bullying, with lesbians, gay men and bisexual men and women, facing potentially obvious and even subtly different experiences within the workplace altogether (Einarsdottír et al., 2015).


Notwithstanding the low baseline of LGB people in this sample, our sample appears to be among some of the largest studies of LGB employees carried out using a sound methodology anywhere. A recent study on the issue of power and sample size show that with less than 100 observations an entropy R2 over .80 is needed to speak of a stable result (Gudicha, Tekle & Vermunt, in press) or, differentiating latent class clusters is not a problem with small samples if entropy R2 is above .80. In our case entropy was .85. Still, for a latent class cluster (LC) approach a larger sample might have allowed us to model for cross-categorization of sexual orientation and other socio-demographical variables, which may have resulted in more latent clusters thereby possibly distinguishing additional target groups of bullying. An enlarged sample size among the LGB sample, that is, over 100 observations and preferably 200 observations per type of sexual orientation would also have allowed us to assess the measurement variance in a more refined fashion. This, however, has significant cost implications as this study required 44 waves to capture a sample of n = 500 which took six months to achieve. Although, the current LC analysis did not give indications in this direction, it may be plausible to assume that, with a larger sample, a specific lesbian cluster may emerge where sexualized behaviours are more dominant than for others. It is also plausible that a specific LGB cluster corresponding with a less escalated form of bullying that starts with targeting infrequently the stereotypical or salient distinctive characteristics of some LGB people (Einarsdóttir et al., 2016) by using sexualized negative behaviours. Notwithstanding, there is no conceivable reason to assume that the latent class solution was not stable; power issues emerged when studying the relationship between sexuality and bullying in more depth only among LGB people. The LCA framework offered us a way to overcome this by including heterosexual employees when studying disclosure by recoding them from “not applicable” to “hetero” and treating disclosure issues as nominal response scales. Unfortunately this comes at a price, namely a lack of statistical power. Therefore, we adapted the type I error from 5% to 10%. In measuring disclosure, we asked LGB employees to assess how open they considered themselves to be about their sexual orientation, which may or may not correspond with their heterosexual colleagues’ perceptions of their openness. Moreover, these were single item operationalisations, which in the case of formative measures, should not pose many problems, although future research may want to operationalise these differently.

Readers may have question why the study is limited to LGB employees, leaving out transgender (T). This reflects the fact that at the time of designing the questionnaire we sought advice from Stonewall, the leading British LGBT rights charity and advocacy group. At that time, they recommended not to include transgender (T), as T was considered associated with gender rather than sexuality/sexual orientation, a position Stonewall has since changed. Accordingly, to capture the entire spectrum of sexuality or sexual orientation future studies should also include transgender and non-binary people expressed by the umbrella term LGBTQ+.

Readers should also note that the term “disability” measured several forms of disability; “physical disability”, “emotional/psychological disability” and “learning difficulties”, which for the purpose of this study were collapsed into one category. A more detailed investigation here would have required a larger sample to be meaningful.


As the first dedicated study to explore the impact of sexual orientation on the experience of bullying, this article shows that being LGB is associated with elevated risks of bullying despite, in the case of the UK, LGBs benefitting from protected group status in line with British and European Union anti-discrimination regulation. By applying a Latent Cluster Analysis, we demonstrate that overall, the nature of the experience of LGBs is very similar to that of heterosexuals. Nevertheless, LGB employees are, to a greater extent, exposed to intrusive, sexualized behaviours, reflecting continuing and potentially deep-rooted negative attitudes and prejudices towards LGB people, whether this is expressed as outright hostility or through disrespectful behaviour (Einarsdottir, Hoel & Lewis, 2015).

Our study reveals that, contrary to expectation, openness about being LGB is not associated with any elevated risk of becoming a target of bullying, indicating that disclosing non-heterosexuality in the context of UK workplaces is in most cases safe and without repercussions. However, many LGBs still choose not to disclose their sexual orientation, which in light of the greater perceived risk for LGBs of becoming a target of bullying is understandable. It also seems the case that decisions not to disclose is not consistently respected by colleagues, implying that decisions about disclosure is not always fully under the control of LGBs. Altogether, this warrants a need for organizational responses, which encourages inclusiveness, actively challenges homophobia and expressions of disrespect for LGBs, and where privacy is respected for those who may not wish to draw attention to their sexual orientation.

Funding Information

This work was funded by Economics and Social Research Council (RES-062-23-2412).

Competing Interests

The authors have no competing interests to declare.


  1. Acas. (2007). The experiences of sexual orientation and religion or belief discrimination employment tribunal claimants. London: Acas. 

  2. Ashforth, B. E., & Mael, F. (1989). Social identity theory and the organization. The Academy of Management Review, 14, 20–39. DOI: 

  3. Björkqvist, K., Österman, K., & Hjelt-Bäck, M. (1994). Aggression among university employees. Aggressive Behavior, 20, 173–184. DOI:<173::AID-AB2480200304>3.0.CO;2-D 

  4. Brewer, M. B. (1999). The psychology of prejudice: Ingroup love and outgroup hate? Journal of Social Issues, 55, 429–444. DOI: 

  5. Brewer, M. B., & Miller, N. (1984). Beyond the contact hypothesis: Theoretical perspectives on desegregation. In N. Miller & M. Brewer (Eds.), Groups in Contact: The Psychology of Desegregation (pp. 281–302). New York: Academic Press. DOI: 

  6. Bowring, J. B., & Brewis, J. (2009). Managing lesbian and gay identity in the Canadian workplace. Equal Opportunities International, 28, 361–377. DOI: 

  7. Chegeni, R., Notelaers, G., Pallesen, S., & Sagoe, D. (2021). Aggression and psychological distress in male and female anabolic-androgenic steroid users: A multigroup latent class analysis. Frontiers in psychiatry, 12, 629428. DOI: 

  8. CIPD. (2005). Managing conflict at work A survey of the UK and Ireland. London: Chartered Institute of Personnel Development. 

  9. Clair, J. A., Beatty, J. E., & MacLean, T. L. (2005). Out of sight but not out of mind: Managing invisible social identities in the workplace. Academy of Management Review, 30, 78–95. DOI: 

  10. Clogg, C. C., & Goodman, L. A. (1985). Simultaneous latent structure analysis in several groups. In N. B. Tuma (Ed.), Sociological Methodology (pp. 18–110). San Francisco, CA: Jossey-Bass. DOI: 

  11. Clogg, C. C., & Goodman, L. A. (1986). On scaling models applied to data from several groups. Psychometrika, 51, 123–136. DOI: 

  12. Colgan, F., Creegan, C., McKearney, A., & Wright, T. (2008). Lesbian workers. Journal of Lesbian Studies, 12, 31–45. DOI: 

  13. Colgan, F., Wright, T., Creegan, C., & McKearney, A. (2009). Equality and diversity in the public services: Moving forward on lesbian, gay and bisexual equality? Human Resource Management Journal, 19, 280–301. DOI: 

  14. Cowie, H., Jennifer, D., Neto, C., Angulo, C., Pereira, B., Del Barrio, C., & Ananiadou, K. (2000). Comparing the nature of workplace bullying in two European countries: Portugal and UK. In M. Sheehan, S. Ramsay, & J. Patrick (Eds.), Transcending boundaries: Integrating people, processes and systems’ Conference proceedings (pp. 120–125). Queensland, Australia; Griffith University. 

  15. Croteau, J. M. (1996). Research on the work experiences of lesbian, gay and bisexual people: An integrative review of methodology and findings. Journal of Vocational Behavior, 48, 195–209. DOI: 

  16. Croteau, J. M., Anderson, M. Z., & VanderWal, B. L. (2008). Models of workplace sexual identity disclosure and management: Reviewing and extending concepts. Group & Organization Management, 33, 532–565. DOI: 

  17. Day, N., & Schoenrade, P. (1997). Staying in the closet versus coming out: Relationships between communication about sexual orientation and work attitudes. Personnel Psychology, 50, 147–63. DOI: 

  18. D’Cruz, P. (2012). Workplace bullying in India. New Delhi: Routledge. 

  19. Di Marco, D., Hoel, H., & Lewis, D. (2021). Discrimination and exclusion on grounds of sexual and gender identity: Are LGBT people’s voices heard at the workplace? The Spanish Journal of Psychology, 24, 1–6. DOI: 

  20. Di Martino, V., Hoel, H., & Cooper, C. L. (2003). Preventing violence and harassment in the workplace. European Foundation for the Improvement of Living and Working Conditions. Luxemburg: Office for Official Publications of the European Communities. 

  21. Eid, M., & Diener, E. (2001). Norms for experiencing emotions in different cultures: Inter- and intranational differences. Journal of Personality and Social Psychology, 81, 869–885. DOI: 

  22. Einarsdottír, A., Hoel, H., & Lewis, D. (2016). Fitting the bill: Disclosure and embodiment. Work, Employment and Society, 30, 489–505 DOI: 

  23. Einarsdottír, A., Hoel, H., & Lewis, D. (2015). ‘It’s nothing person’: Anti-homosexuality in the British workplace. Sociology, 49, 1183–1199. DOI: 

  24. Einarsen, S., Hoel, H., & Notelaers, G. (2009). Measuring exposure to bullying and harassment at work: Validity, factor structure and psychometric properties of the Negative Acts Questionnaire-Revised. Work & Stress, 23, 24–44. DOI: 

  25. Einarsen, S., Hoel, H., Zapf, D., & Cooper, C. L. (2020). The concept of bullying at work: The European tradition. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Theory, Research and Practice. Third Edition (pp. 3–53). London/New York: CRC Press - Taylor and Francis Group. DOI: 

  26. Ellison, G., & Gunstone, B. (2009). Sexual orientation explored: A study of identity, attraction, behavior and attitudes in 2009. Manchester: Equality and Humans Rights Commission. 

  27. Emdad, R., Alipour, A., Hagberg, J., & Jensen, I. B. (2013). The impact of bystanding to workplace bullying on symptoms of depression among women and men in industry in Sweden: An empirical and theoretical longitudinal study. International Archives Occupational Environ Health, 86, 709–16. DOI: 

  28. Fevre, R., Lewis, D., Robinson, A., & Jones, T. (2012). Trouble at work. London: Bloomsbury Academic Press. DOI: 

  29. Fevre, R., Nichols, T., Prior, G., & Rutherford, I. (2009). Fair treatment at work report: Findings from the 2008 survey. Employment Relations Research Series, No. 103. London: Department for Business, Innovation and Skills. DOI: 

  30. Fevre, R., Robinson, A., Lewis, D., & Jones, T. (2013). The ill-treatment of employees with disabilities in British workplaces. Work, Employment and Society, 27, 288–307. DOI: 

  31. Fox, S., & Stallworth, L. (2005). Racial/ethnic bullying; Exploring links between bullying and racism in the US workplace. Journal of Vocational Behavior, 66, 438–456. DOI: 

  32. Froyum, C. M. (2007). ‘At least I’m not gay’: Heterosexual identity making among poor black teens. Sexualities, 10, 603–22. DOI: 

  33. Giga, S. I., Hoel, H., & Lewis, D. (2008). A review of Black and minority ethnic (BME) employee experiences of workplace bullying. London: Unite the union/Department for Business, Enterprise and Regulatory Reform. 

  34. Grainger, H., & Fitzner, G. (2007). The first fair treatment at work survey (2007). Employment Relations Research Series, No. 63, London; DTI. 

  35. Griffith, K., & Hebl, M. (2002). The disclosure dilemma for gay men and lesbians: ‘Coming out’ at work. Journal of Applied Psychology, 87, 1191–1199. DOI: 

  36. Gudicha, G. W., Tekle, F. B., & Vermunt, J. K. (2016). Power and sample size computation for Wald tests in latent class models. Journal of Classification, 33, 30–51. 

  37. Hagenaars, J. A. (1990). Categorical longitudinal data–loglinear analysis of panel, trend and cohort data. Newbury Park: CA: Sage. 

  38. Herek, G. M., & McLemore, K. A. (2013). Sexual prejudice. Annual Review of Psychology, 64(1), 309–333. DOI: 

  39. Hershcovis, M. S., & Reich, T. C. (2013). Integrating workplace aggression research: Relational, contextual, and method considerations. Journal of Organizational Behavior, 34(1), 26–42. DOI: 

  40. Hoel, H., Cooper, C. L., & Einarsen, S. (2020). Organisational effect of workplace bullying. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Theory, Research and Practice. Third Edition (pp. 209–234). London/New York: CRC Press – Taylor and Francis Group. DOI: 

  41. Hoel, H., Cooper, C. L., & Faragher, B. (2001). The experience of bullying in Great Britain: The impact of organizational status. European Journal of Work and Organizational Psychology, 10, 443–465. DOI: 

  42. Hogg, M. A., & Terry, D. I.. (2000). Social identity and self-categorization processes in organizational contexts. The Academy of Management Review, 25, 121–140. DOI: 

  43. Jackson, S. (2006). Gender, sexuality and heterosexuality: The complexity (and limits) of heteronormativity. Feminist Theory, 7, 105–121. DOI: 

  44. Jones, T., Robinson, A., Fevre, R., & Lewis, D. (2011). Workplace assaults in Britain: Understanding the influence of individual and workplace characteristics. British Journal of Criminology, 51, 159–178. DOI: 

  45. Kankaras, M., Moors, G., & Vermunt, J. K. (2010). Testing for measurement invariance with latent class analysis. In: E. Davidov, P. Schmidt, & J. Billiet (eds.), Cross-Cultural Analysis: Methods and Applications (pp. 359–384). New York: Routledge. 

  46. Leon-Perez, J. M, Notelaers, G., Arenas, A., Munduate, L., & Medina, F. J. (2014). Identifying victims of workplace bullying by integrating traditional estimates approaches into a latent class cluster model. Journal of Interpersonal Violence, 29, 1155–1177. DOI: 

  47. Lewis, D., Giga, H., & Hoel, H. (2011). Bullying and workplace discrimination. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Development in Theory, Research and Practice. Second Edition, 267–281. Oxford: Taylor and Francis. 

  48. Lewis, D., Glambek, M., & Hoel, H. (2020). The role of discrimination in workplace bullying. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Theory, Research and Practice. Third Edition (pp. 363–383). London/New York: CRC Press – Taylor and Francis Group. DOI: 

  49. Lewis, D., & Gunn, R. W. (2007). Workplace bullying in the public sector: Understanding the racial dimension. Public Administration and International Quarterly, 83, 641–665. DOI: 

  50. Lewis, D., Hoel, H., & Einarsdóttir, A. (2013). Hard to research and hard to reach: Methodological challenges in exploring bulling, harassment and discrimination with lesbians, gay, and bisexual employees. 13th European Academy of Management (Euram) Conference, Istanbul, 26–20 June. 

  51. Leymann, H. (1992). Vuxenmobbning på svenska arbeidsplatser. Delrapport 1 on frekvenser. [Adult bullying in Swedish workplace: Report 1 about prevalence]. Stockholm: Arbetarskyddstyrelsen. 

  52. Lidderdale, M. A., Croteau, J. M., Anderson, M. Z., Tovar-Murray, D., & Davis, J. M. (2007). Building LGB vocational psychology: A theoretical model of workplace sexual identity management. In R. Bieschke & K. DeBord (Eds.), Handbook of Counseling and Psychotherapy with Lesbian, Gay and Bisexual Clients. Second Edition (pp. 245–270). Washington, DC: American Psychology Association. DOI: 

  53. Lutgen-Sandvik, P., Tracy, S. J., & Alberts, J. K. (2007). Burned by bullying in the American workplace: Prevalence, perception, degree, and impact. Journal of Management Studies, 44, 835–860. DOI: 

  54. Magidson, J., & Vermunt, J. K. (2002). Latent class models for clustering: A comparison with K-means. Canadian Journal of Marketing Research, 20, 37–44. 

  55. Magidson, J., & Vermunt, J. K. (2004). Latent class models. In D. Kaplan (Ed.), The Sage Handbook for Quantitative Methodology (pp. 175–198). Thousand Oakes, CA: Sage Publications. 

  56. Martin, J. L., & Konx, J. (2000). Methodological and ethical issues in research on lesbians and gay men. Social Work Research, 24, 51–59. DOI: 

  57. McBride, A., Hebson, G., & Holgate, J. (2015). Intersectionality: Are we taking enough notice in the field of work and employment relations? Work, Employment and Society, 29, 331–341. DOI: 

  58. McCutcheon, A. L. (1987). Latent class analysis. Newbury Park, CA: Sage. DOI: 

  59. McLaughlin, K. A., Hatzenbuehler, M. L., & Keyes, K. M. (2010). Responses to discrimination and psychiatric disorders among Black, Hispanic, female, and lesbian, gay and bisexual individuals. American Journal of Public Health, 100(8), 1477–1484. DOI: 

  60. Meyer, I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129, 674–697. DOI: 

  61. Minton, S. J., Dahl, T., O’Moore, A. M., & Tuck, D. (2008). An exploratory survey of the experiences of homophobic bullying among lesbian, gay, bisexual and transgendered young people in Ireland. Irish Educational Studies, 27, 177–191. DOI: 

  62. Ng, E. S., & Rumens, N. (2017). Diversity and inclusion for LGBT workers: Current issues and new horizons for research. Canadian Journal of Administrative Sciences, 34, 109–120. DOI: 

  63. Niedhammer, I., Chastang, J.-F., & David, S. (2007). Importance of psychosocial work factors on general health outcomes in the national French SUMER survey. Occupational Medicine, 58, 15–24. DOI: 

  64. Nielsen, M. B., & Einarsen, S. (2012). Outcomes of workplace bullying: A meta-analytical review. Work and Stress, 26(4), 309–332. DOI: 

  65. Nielsen, M. B., Notelaers, G., & Einarsen, S. (2020). Methodological issues in the measurement of workplace bullying. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Theory, Research and Practice. Third Edition (pp. 235–265). London/New York: CRC Press – Taylor and Francis Group. DOI: 

  66. Nielsen, M. B., Skogstad, A., Matthiesen, S. B., Glasø, L, Aasland, M. S., Notelaers, G., & Einarsen, S. (2009). Prevalence of workplace bullying in Norway: Comparison across time and estimation methods. European Journal of Work and Organizational Psychology, 18, 81–101. DOI: 

  67. Notelaers, G., Baillien, E., Vermunt, J. K., De Witte, H., & Einarsen, S. (2011). Exposure to workplace bullying: Exploring risk groups with categorical data. Industrial Health, 49, 73–88. DOI: 

  68. Notelaers, G., Einarsen, S., De Witte, H., & Vermunt, J. K. (2006). Measuring exposure to bullying at work: The validity and advantages of the latent class cluster approach. Work and Stress, 20(4), 288–301. DOI: 

  69. Notelaers, G., Van der Heijden, B., Hoel, H., & Einarsen, S. (2019). Measuring bullying at work with the short-negative acts questionnaire: Identification of targets and criterion validity. Work & Stress, 33, 58–75. DOI: 

  70. Notelaers, G., & Van der Heijden, B. I. (2021). Construct validity in workplace bullying and harassment research. In D’Cruz, P., Noronha, E., & Notelaers, G. (Eds.), Concepts, Approaches and Methods. Handbooks of Workplace Bullying, Emotional Abuse and Harassment, 369–424. DOI: 

  71. Paas, L. (2014). Comments on: Latent Markov models: A review of a general framework for the analysis of longitudinal data with covariates. Test, 23(3), 473–477. DOI: 

  72. Pinkos Cobb, E. (2017). Workplace bullying and harassment: New development in international law. Routledge, Taylor & Francis. DOI: 

  73. Price, E. (2011). LGBT sexualities in social care research. London: London School of Economics and Political Science. 

  74. Ragins, B. (2004). Sexual orientation in the workplace: The unique work and career experiences of gay, lesbian and bisexual workers. Research in Personnel and Human Resource Management, 23, 35–120. DOI: 

  75. Ragins, B. (2008). Disclosure disconnects: Antecedents and consequences of disclosing invisible stigmas across life domains. Academy of Management Review, 33, 194–215. DOI: 

  76. Ragins, B. L., & Wiethoff, C. (2004). Understanding heterosexism at work: The straight problem. In R. L. Dipboye & A. Colella (Eds.), Discrimination at Work: The psychological and Organizational Bases (pp. 177–201). Mahwah, NJ: Lawrence Erlbaum Associates. 

  77. Ramsey, S., Troth, A., & Branch, S. (2011). Work-place bullying: A group processes framework. Journal of Occupational and Organizational Psychology, 84, 799–816. DOI: 

  78. Röndahl, G., Innala, S., & Carlsson, M. (2007). ‘To hide and not to hide, that is the question’: Lesbians and gay men describe experiences from nursing work environment. Journal of Homosexuality, 52, 211–233. DOI: 

  79. Rosander, M., Salin, D., Viita, L., & Blomberg, S. (2020). Gender matters: Workplace bullying, gender, and mental health. Frontiers in Psychology, Oct 6. DOI: 

  80. Rumens, N., & Broomfield, J. (2012). Gay men in the police: Identity disclosure and management issues. Human Resource Management Journal, 22, 283–98. DOI: 

  81. Salin, D. (2001). Prevalence and forms of bullying among business professionals: A comparison of two different strategies for measuring bullying. European Journal of Work and Organizational Psychology, 10, 425–441. DOI: 

  82. Salin, D., & Hoel, H. (2013). Workplace bullying as a gendered phenomenon. Journal of Managerial Psychology, 28, 235–51. DOI: 

  83. Simpson, R., & Cohen, C. (2004). Dangerous work: The gendered nature of bullying in the context of higher education. Gender, Work and Organization, 11, 163–186. DOI: 

  84. Stonewall. (2007). Bullying: Preventing the bullying and harassment of gay employees. London, UK: Stonewall. 

  85. Tajfel, H. (1972). Social categorization. English manuscript of ‘La catégorisation sociale.’ In S. Moscovici (Ed.) Introduction á la Psychologie Soiale, 1, (pp. 272–302). Paris: Larousse. 

  86. Tilcsik, A., Anteby, M., & Knight, C. R. (2015). Concealable stigma and occupational segregation: Towards a theory of gay and lesbian occupations. Administrative Science Quarterly, 60, 446–481. DOI: 

  87. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Whetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford, UK: Blackwell. 

  88. Vermunt, J. K. (2013). Markov tutorial #2: Latent GOLD longitudinal analysis of life satisfaction. Belmunt, MA: Statistical Innovations Inc 

  89. Vermunt, J. K., & Magidson, J. (2009). Latent Gold 4.5 user’s guide. Belmont, MA: Statistical Innovations Inc. 

  90. Ward, J., & Winstanley, D. (2003). The absent presence: Negative space within discourse and the construction of minority sexual identity in the workplace. Human Relations, 56, 1255–1280. DOI: 

  91. Ward, J., & Winstanley, D. (2006). Walking the watch: The UK fire service and its impact on sexual minorities in the workplace. Gender, Work and organization, 13, 193–219. DOI: 

  92. Wax, A., Coletti, K. K., & Ogaz, J. W. (2018) The benefit of full disclosure: A meta-analysis of the implications of coming out at work. Organizational Psychology Review, 8, 3–30. DOI: 

  93. Wenzel, M., Mummendey, A., & Waldzus, S. (2007). Superordinate identities and intergroup conflict: The ingroup projection model. European Review of Social Psychology, 18, 331–372. DOI: 

  94. Williams, M., & Tregidga, J. (2014). Hate crime victimisation in Wales: psychological and physical impacts across seven hate crime victim types. British Journal of Criminology, 54, 946–967. DOI: 

  95. Zapf, D., Escarín, J., Van der Heijden, B., Einarsen, S., Hoel, H., & Vartia, M. (2020). Empirical findings on prevalence and risk groups of bullying in the workplace. In S. Einarsen, H. Hoel, D. Zapf & C. L. Cooper (Eds.), Bullying and Harassment in the Workplace: Theory, Research and Practice. Third Edition. (pp. 105–162). London/New York: CRC Press – Taylor and Francis Group. DOI: 

comments powered by Disqus