Date of Graduation


Document Type



Despite a robust, concordant literature demonstrating that lesbian, gay, and bisexual (i.e., sexual minority) populations smoke more than their heterosexual peers, the factors contributing to this disparity are not well understood. Consequently, the main goal of this project and its three component studies was to enhance current understanding of risk factors that contribute to smoking disparities among sexual minority persons. Study 1 details a systematic review of the literature, synthesizing empirical investigations that tested associations of risk factors and correlates of smoking among sexual minority samples. Of the twenty-six studies included in the analysis, most used sexual identity to operationalize sexual orientation and few used probability-based samples. Results showed risk factors that were both unique to sexual minority populations (i.e., internalized homophobia) and common among the general population (i.e., lower education). Yet, there was considerable discordance among the minority stress variables. Future studies are needed to examine factors that are associated with uptake of smoking, which necessitates samples of young individuals as adolescence is the most prevalent period of smoking initiation. Study 2 involved a secondary analysis of the Fall 2008 and Spring 2009 National College Health Association data sets to examine associations of discrimination and violence victimization with cigarette smoking within sexual orientation groups. The data were truncated to respondents 18-24 years of age (n=92,470). Since heterosexuals comprised over 90% of respondents, a random 5% subsample of heterosexuals was drawn, creating a total analytic sample of 11,046. Results showed that sexual minorities indicated more experiences of violence victimization and discrimination when compared with their heterosexual counterparts and had nearly twice the current smoking (i.e., smoked cigarettes in the last 30 days) prevalence of heterosexuals. After adjusting for age and race, lesbians/gays who were in physical fights or were physically assaulted had higher proportional odds of being Current Smokers when compared with their lesbian/gay counterparts who did not experience those stressors. When possible, lesbian/gay and bisexual groups should be analyzed separately, as analyses revealed that bisexuals had a higher risk profile than lesbians/gays. Further research is needed with more nuanced measures of smoking (e.g., intensity), as well as examining if victimization may interact with smoking cessation. Finally, Study 3 used a web-based sample of sexual minority emerging adults to examine the association of minority stressors with smoking behaviors. Sexual minority individuals ages 18-24 (n=482) were recruited using two strategies of network sampling and advertising on the social networking website Facebook. Among the entire sample, experiencing some internalized homophobia was associated with decreased odds of smoking. In gender-stratified models, gay-related stress was associated with increased odds of smoking among males but not among females. The stratified models also revealed that females were the principle drivers of the protective association between internalized homophobia and smoking. Internalized homophobia has been shown to decrease the likelihood of integrating with or participating in gay and lesbian communities – populations known to have high smoking rates – thus, it may also, by proxy, decrease a person's probability of socializing with smokers or being subjected to targeted marketing from the tobacco industry. Together, the three studies suggest that the etiology of smoking disparities in sexual minority communities is complex, involving both common correlates of smoking and factors that may be unique to sexual minority populations. There is a clear need for probability-based data collection methods in order to produce generalizable results and to better gage the breadth of this public health problem. Cessation research and clinical trials with sexual minority populations are alarmingly scant in the face of demonstrated disparity.