Variations in stigma by sexual orientation and substance use: An investigation of double stigma

Abstract Stigma can have a debilitating effect on mental health. Both LGBTQ+ individuals and substance users experience stigma, and concurrent membership of both groups can result in double stigma. This study aimed to investigate differences in stigmatized attitudes based on substance use behaviors and sexual identity using data from 102 participants (83% female). Findings suggest the presence of a hierarchy of stigma where heterosexual individuals were typically viewed more positively than LGBTQ+ individuals, regardless of the substance using status of the LGBTQ+ individual. However illicit drug use by heterosexual individuals resulted in the most negative attitudes. Implications are discussed.

with a long documented history of stigmatization and victimization, however, a full exploration of negative effects of these experiences remains scarcely documented (Sattler, Escande, Racine, & G€ oritz, 2017). Importantly, as a result of stigma and victimization, the research that does exist indicates that many individuals report experiences of bullying, being made to feel ashamed and being fearful of "coming out" (i.e., disclosing sexuality to friends, family and the public) (Pistella, Salvati, Ioverno, Laghi, & Baiocco, 2016;Walch, Ngamake, Francisco, Stitt, & Shingler, 2012).
Research also consistently identifies substance users as a stigmatized group, and one often judged more negatively than individuals with mental health disorders; for example, Phillips (2011) reports that people who are alcohol dependent are viewed as more dangerous, more troubled, and more likely to be violent than those with schizophrenia or depression. Those with substance use disorders are also perceived to be more responsible and blameworthy for their illness compared to individuals who experience other disorders relative to a mental or physical condition (Schomerus, Matschinger, & Angermeyer, 2014). Attribution of responsibility and blame are particularly notable in relation to stigma and accessing support, as individuals who are viewed as responsible for their situation are less likely to be supported when accessing the relevant help (Corrigan, Kuwabara, & O'Shaughnessy, 2009). Additionally, there is also evidence of considerable variation in stigmatization dependent on the type of substance that is being used (Link, Phelan, Bresnahan, Stueve, & Pescosolido, 1999, Luty, Fekadu, Umoh, & Gallagher, 2006, with alcohol and tobacco consumption being more socially acceptable and less stigmatized than illicit drugs. While it is established that members of these individual minority groups are subjected to stigma, there is also evidence to support that experiences of stigma are often conflated by concurrent multiple group membership and behaviors (Mizock & Mueser, 2014). Consequently, being identifiable as a member of more than one minority group can mean that the individual may experience double or multiple stigmatization.
Double or multiple stigma is the process whereby an individual experiences heightened stigma due to concurrent membership of multiple groups which are negatively perceived. Double or multiple stigma is consequently associated with poorer wellbeing (Braithwaite, Taylor, & Treadwell, 2009). Individuals from LGBTQþ communities are known to be vulnerable to double or multiple stigmatization. This is particularly documented by research in relation to stigmatized attitudes toward HIV status or sexual risk taking behaviors (Braithwaite et al., 2009), as well as with substance use behaviors.
There are established associations between LGBTQþ and substance using populations (Flentje, Heck, & Sorensen, 2015), empirically supported by connections between the "gay scene" and substance use (Blosnich, Lee, & Horn, 2013;Lea, Reynolds, & de Wit, 2013). The gay scene, within the research literature, is a collective term for clubs, bars and other social venues that specifically cater to LGBTQþ individuals, and it has been linked with increased levels of substance use (Valentine & Skelton, 2003). In addition to Valentine and Skelton (2003) study, there is a consensus within the research literature of elevated drug and alcohol use within this community (Case et al., 2004;Corliss et al., 2010;Hatzenbuehler, McLaughlin, & Slopen, 2013;Hughes & Eliason, 2002;Kelly, Parsons, & Wells, 2006;Marshal, Friedman, Stall, & Thompson, 2009;Meyer, 2003;Talley, Hughes, Aranda, Birkett, & Marshal, 2014;Wong, Kipke, & Weiss, 2008). For example, young men who have sex with men are consistently found to engage with higher rates of alcohol, marijuana, and other illicit drug use compared with their heterosexual peers (Corliss et al., 2010;Kelly et al., 2006;Marshal et al., 2009;Talley et al., 2014).
Given that the experience of stigma can be heightened by the concurrent identification as a member of two or more minority groups-the LGBTQþ community and substance use, in this case-there is a need to clarify the nature of such stigma by exploring the attitudes of members of society. This is particularly important given that substance using behavior which may be invoked as a coping response due to stigma exposure, may serve to further the stigmatization of those individuals as substance users. In order to support these minority groups, there is a need to develop better interventions to target stigma in the general public, and to do so it is critical to know how the public view those individuals who are members of these minority groups, and members of multiple minority groups.
The current study aims to investigate variations in attitudes held toward different sexual orientation and substance using groups, and particularly to identify if substance using LGBTQþ individuals are doubly stigmatized. Based on past research, it is expected that individuals from the LGBTQþ community will be rated more negatively than heterosexual individuals regardless of substance using status. It is also expected that substance users will be rated more negatively than both heterosexual and LGBTQþ abstainers. Furthermore, it is anticipated that there will be differences in stigma toward licit substance use (alcohol) and illicit substance use (drugs).

Participants
A convenience sample was recruited through an online survey to which 107 individuals responded. Data were recorded in Google Forms, and then exported to SPSS for analysis. Data were screened for missing data and duplication of entries and five responses were excluded from the analysis due to incomplete data, leaving a final sample size of 102. The majority of the sample were female (83%) and heterosexual (77%). The sample ranged in age from 18 to 60 with a mean age of 26.6 years (sd ¼ 9.14).

Materials
Participants were asked to report their age as a free text response, to select a gender from a range of pre-determined responses (male, female, transgender, other, and an option to complete as free text), and sexual orientation (response options straight/heterosexual, gay, lesbian, bisexual, queer, prefer not to say, other and an additional free text option).
Participants were then presented with stimulus material consisting of six vignettes to read, and their attitudes toward the individual described in each vignette was recorded immediately following the vignette.
The six vignettes described a named individual identified as either heterosexual (n ¼ 3) or LGBTQþ (n ¼ 3), and engaged in a licit (alcohol, n ¼ 2) or illicit (drugs, n ¼ 2) substance use behavior or non-substance use behavior (n ¼ 2). The information related to the independent variables (i.e., sexual orientation and substance use behavior) was embed into a description of an activity undertaken by the named individual. The follow example is the vignette involving a person who identified as LGBTQþ and engaged in illicit substance use: Jack and his boyfriend Peter go out to their local gay bar every weekend after work. They take several lines of cocaine each whilst out at the bar. They stay out all night and have to walk home early in the morning. They remain in bed the following day.
Stigma toward the minority group individual was assessed using an adaptation of the Attitudes to Mental Illness Questionnaire (AMIQ; Luty et al., 2006). The AMIQ is a five-item measure, with item responses ranging from strongly disagree to strongly agree, and scored on a À2 to þ2 scale with positive scores indicating positive attitudes, and negative scores indicating negative attitudes depending on question phrasing. Neutral and don't know responses are coded as 0. All items are summed, to create a total score ranging from À10 to þ10, with lower scores indicative of greater stigma. Items on the AMIQ include (I would be comfortable if named person was my colleague at work, and Do you think this would damage named person's career? The AMIQ demonstrates good psychometric properties, including stability, test-retest reliability, alternative test reliability, and face, construct and criterion validity (Luty et al., 2006). While conceptualized as a measure of attitudes toward mental illness, the original validation of the AMIQ included vignettes portraying substance use, specifically alcohol consumption and heroin use (Luty et al., 2006). Adaptations of the AMIQ in this study did not alter the wording of four out of five items, beyond changes to the name of the individual each question was being asked about in the specific vignette. One item on the likelihood of the named person's wife leaving him was edited as appropriate to the gender mix, orientation and relationship status reported in the vignette (i.e., husband/wife/partner, leaving him/her/them. The AMIQ showed good to excellent reliability when applied to each vignette description, which was performed separately to assess the reliability for each vignette combination, and the appropriateness of the tool for use with LGBTQþ vignette, ranging from .60 to .81 (data available on request from authors).

Procedure
Ethics approval for the current study was obtained from Glasgow Caledonian University, Psychology Ethics Committee. Participants were recruited online via social media advertisement (i.e., Facebook and Twitter) in 2017. Participants were directed from a recruitment poster and text via a link to an information sheet, and subsequently to consent items, and an online survey on Google Forms. Participants were eligible to participate if they were over aged 18, and no other exclusion criteria were applied. Participation in the study was not incentivised. After completing the demographic items, participants were presented with the six vignettes in turn, and immediately after each vignette they were asked to complete the AMIQ with the individual described in the vignette in mind. All data were imported into SPSS v23 for analysis, and a mixed ANOVA was performed to compare responses across the six vignettes and differences by gender and sexual orientation of the participant. Post-hoc paired samples t-tests with a Bonferroni correction of p < 0.03 where then performed to further investigate a significant main effect of the within-subjects vignette condition. Due to the nature of the online survey host it was not possible to randomize the presentation of the vignettes, therefore they were presented in the following order consistently for each participant. Participants were presented with a range of support materials at the end of the survey as part of the debrief material.

Results
Table 1 provides data on the sexual orientation of the sample. As can be seen in Table 2, there were differences in mean attitude scores across the vignettes. There was a general trend for heterosexual individuals to be rated more favorably than LGBTQþ individuals, with the exception of heterosexual drug users who were rated the most negatively. All LGBTQþ individuals were rated more favorable than heterosexual drug users, but less so than heterosexual drinkers and abstainers. For both LGBTQ individuals and heterosexual individuals there was a clear trend for non-users to be rated the most favorably, followed by drinkers, and the most negative attitudes were reported to drug users.
A mixed ANOVA was performed to explore the main effect of vignette characteristics (orientation and substance use), the main effects of participant sex, main effects of participant sexual orientation, and the interaction effects between vignette characteristics with both participant sex and sexual orientation separately. This indicated a significant main effect of vignette characteristic (F (5,425) ¼ 82.334, p < 0.001), but non-significant main effects of participant sex (F (1,85) ¼ 2.524, p ¼ 0.116), and for participant sexual orientation (F (1,85) ¼ 0.137, p ¼ 0.712). Additionally there was no significant interaction effect of vignette x participant sex (F (5,425) ¼ 0.812, p ¼ 0.496), or of vignette x participant sexual orientation (F (5,425) ¼ 0.374, p ¼ 0.787).
To further explore the within-subjects differences in attitudes across the vignettes, post-hoc paired samples t-tests were performed. Bonferroni corrected at p < 0.03, these analyses revealed several significant pairings. Heterosexual individuals without substance use behavior were rated the most positive, and significantly more so than heterosexual alcohol users, LGBTQ alcohol users and LGBTQ non-users. Heterosexual alcohol users were significantly more positively rated than LGBT non-users and alcohol users.
LGBTQ non-users were rated significantly more positively than LGBTQ alcohol users, and LGBTQ alcohol users significantly more positively than heterosexual drug users. Finally, LGBTQ drug users were significantly more negatively perceived than LGBT alcohol users, but more positively perceived than heterosexual drug users.

Discussion
Overall attitudes toward individuals within the vignettes in this study indicate that individuals from the LGBTQþ community are perceived more negatively than heterosexual individuals, and that substance users are perceived more negatively than abstainers are viewed. Those engaging with illicit drug use are the most stigmatized followed by alcohol users. When considering the dual minority group status of the individuals described there was an interesting trend whereby attitudes were generally less negative toward heterosexual individuals whether an abstainer or alcohol user than all of the LGBTQþ individuals, except where the heterosexual individual was a drug user. All LGBTQþ individuals described were rated more positively than the heterosexual drug user but were more stigmatized than either the heterosexual non-user or alcohol user. These findings support consistent findings in the literature of stigma experienced by LGBTQþ individuals (Darwich, Hymel, & Waterhouse, 2012;Weber, 2008), and the growing evidence of a double stigma particularly for LGBTQþ individuals involved in substance use (Luty et al., 2006;Phillips, 2011). The evidence of negative attitudes toward the LGBTQþ community within the context of alcohol and illicit drug-related settings may reflect a perception that LGBTQþ individuals are more likely to engage in these behaviors, along with a judgment of that engagement. For instance, these negative attitudes could reflect stereotypical perceptions of the gay scene and association of LGBTQþ individuals with this scene, such as extreme representations of chemsex (sexual activity undertaken by LGBTQþ individuals under the influence of illicit substances) in the media (Heritage & Baker, 2021). Alongside the elevated levels of stigma for non-substance using LGBTQþ individuals this points to a clear stigmatization of LGBTQþ individuals which is further confounded by engagement in a substance using behavior. The double stigmatization of substance using LGBTQþ individuals is of concern, given the link between the experience of stigma and substance use behaviors as posited within the minority stress hypothesis (Blosnich et al., 2013), and indeed there is evidence of elevated substance use behaviors in the LGBTQþ population (Espelage, Aragon, Birkett, & Koenig, 2008;McCabe, Bostwick, Hughes, West, & Boyd, 2010).
Of additional interest is the finding that heterosexual drug users were the only heterosexual individuals to be stigmatized at a higher level than any of the LGBTQþ individuals represented in the vignettes. There is a lack of research evidence investigating public attitudes toward heterosexual substance users, with most studies either focusing on the general public without consideration of sexual orientation, or on specific minority groups such as LGBTQþ individuals. However, one possible explanation for this finding may be that this greater stigmatization reflects an expectation that heterosexual individuals should not be engaging in such behaviors, while it may be more normatively accepted and expected of LGBTQ individuals, and hence it is more negatively judged. The findings of this study suggest a need for further research that investigates public perceptions of both LGBTQþ substance users, and heterosexual substance users, to better understand how these individuals are stigmatized.
It is important to note that this study did not account for the participants' own substance use behaviors. As noted by Weber (2008) participant's own alcohol or drug use can have an influential role in attitudes toward others use, and is therefore an important consideration that is not accounted for in this analysis. Future research should address this limitation to explore the role that this may have played in the attribution of negative attitudes.
Furthermore, the work of Link et al. (1999) sheds light on the possible effect of drug type in influencing attitude scores. Their results suggest that drugs overall are negatively perceived, similar to the current study, but that the classification of the drug and specific drug itself can play a role in attitudes toward use. While in this research attempts were made to ensure parity among the illicit substances described (ecstasy and cocaine, both class A drugs in the UK), it may be that perceptions of those substances, and perhaps those that use those substances, differ, and future research is warranted to account for this.

Limitations
A number of limitations to the current study should be noted. The study utilized a convenience sampling method and was an online study. While the online methodology has benefits in reducing potential response bias, and can be conducive to more open disclosure of attitudes and behaviors due to anonymity, this anonymity can also allow for participants to misrepresent themselves. Furthermore, an online convenience sample was adopted as a relatively inexpensive, and efficient way to recruit participants to the current study, however, the resulting sample has a sizeable gender bias, with more females than males participating. This may be particularly important due to gender differences in attitudes toward sexual orientation (Horn, 2009) and in engagement with substance using behaviors (e.g., Chen, Strain, Crum, & Mojtabai, 2013). The impact of the gender bias has been controlled statistically to attempt to reduce the impact of that bias. Limited demographic information was collected as part of this study, and it is important for future research to consider characteristics of participants providing stigma ratings as they may have an impact on perceptions of minority groups and substance users (Sattler et al., 2017). Additionally, it was not possible to randomize the ordering of the vignettes and therefore, they were presented in the same order to every participant, which may have introduced some bias in the results.

Conclusion
The results from the current study, however, also adds to the body of research that suggests that individuals who have multiple minority group status are subject to greater stigma, and in particular that attitudes toward substance using LGBTQþ individuals are compounded by stigma on the basis of both sexual orientation and substance use. In order to address the experience of stigma toward these individuals and deliver to them the necessary and appropriate support, it would appear necessary to address both on public attitudes around sexual orientation and on substance use.

Informed consent
All participants provided fully informed consent for the research, and for use of their data in publications

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

Data availability statement
Anonymised data files are available on request from the corresponding author.