• Open Access

Cross-Sectional Evidence for a Stress-Negative Affect Pathway to Substance Use among Sexual Minority Girls


  • Michael P. Marshal Ph.D.,

    Corresponding author
    1. Department of Pediatrics, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    • Department of Psychiatry, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • Chad M. Burton Ph.D.,

    1. Department of Psychiatry, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • Deena J. Chisolm Ph.D.,

    1. Department of Pediatrics, College of Medicine, Ohio State University, Columbus, Ohio, USA
    Search for more papers by this author
  • Gina S. Sucato M.D.,

    1. Department of Pediatrics, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author
  • Mark S. Friedman Ph.D.

    1. University of Pittsburgh, Department of Behavioral and Community Health Sciences, Graduate School of Public Health, Pittsburgh, Pennsylvania, USA
    Search for more papers by this author

Correspondence: MP Marshal (marshalmp@upmc.edu)


Sexual minority girls (SMGs) are four times more likely to engage in substance use than are heterosexual girls. A better understanding of the explanatory mechanisms of this disparity is needed to inform prevention and intervention programs. The goal of this study was to conduct a preliminary test of a “stress-negative affect” pathway by examining gay-related victimization and depression as mediators of substance use among SMGs. Adolescent girls (N = 156, 41% SMGs) were recruited from two urban adolescent medicine clinics to participate in an NIH-funded study of adolescent substance use. The average age was 17.0 years old and 57% were nonwhite. Mediation analyses were conducted in a multiple regression framework using SPSS and a mediation macro utilizing bias-corrected bootstrapping. Four models were estimated to test mediated pathways from sexual orientation to gay-related victimization (Mediator 1), to depression symptoms (Mediator 2), and then to each of four substance use variables: cigarettes, marijuana, alcohol, and heavy alcohol use. Significant mediated pathways (mediation tests with 95% CIs) were found for cigarette, alcohol and heavy alcohol use outcome variables. Results provide preliminary support for the minority stress hypothesis and the stress-negative affect pathway, and may inform the development of future prevention and intervention programs.


Sexual minority girls (SMGs; girls who report same-sex sexual orientation) are over 400% more likely to report substance use than are heterosexual girls.[1] The minority stress hypothesis[2] suggests that SMG disparities may be due to a pervasive homophobic culture that promotes discrimination and violence against members of the lesbian, gay, and bisexual (LGB) community, causing stress and long-term negative health consequences of stress. A recent meta-analysis showed that LGB youth do indeed suffer from higher rates of violence and victimization compared with heterosexual youth.[3] Other recent cross-sectional studies have shown that stress mediated depression outcomes among LGB youth[4] and importantly that gay-specific victimization acted as a mediator.[5] The goal of the current cross-sectional study was to build upon these important findings by extending this mediated pathway to substance use outcomes, and determining whether the combination of gay-related victimization and depression symptoms explained substance use outcomes among SMGs, in what has become known as the stress-negative affect pathway of risk.[6] Given that there are currently no empirically supported substance use interventions available for the LGB youth population, identifying mediated pathways of risk for substance use among this high-risk group is an important first step in future efforts to translate psychosocial health research findings into effective interventions.



Participants were recruited from patients being seen by adolescent medicine specialists for routine or consultative medical care at two urban adolescent medicine clinics in Ohio and Pennsylvania. Inclusion criteria were: (1) age 14–19 years old and (2) ability to read and understand English at the sixth grade level. Neither the participants’ health status at the time of their clinical visit nor their presenting problem were considered as inclusion or exclusion criteria for recruitment. Participants were 156 adolescent females ranging in age from 14.1 to 19.7 (m = 17.0, SD = 1.33). Represented racial groups included 43% African-American, 44% White, 1% Asian-American. Twelve percent reported a multiracial heritage and 10% reported Hispanic heritage.


Study procedures were approved by the Institutional Review Boards at the University of Pittsburgh and Nationwide Children's Hospital. Parental consent was obtained for all participants 14–17 years old. Participants completed mental health and psychosocial health questionnaires including substance use on a computer in a private room.


Sexual orientation

Participants' sexual orientation was measured by one item that stated “Please choose the description that best fits how you think about yourself.” Response options were: 100% heterosexual (straight); mostly heterosexual (straight), but somewhat attracted to people of your own sex; bisexual—that is, attracted to men and women equally; mostly homosexual (gay), but somewhat attracted to people of the opposite sex; and 100% homosexual (gay). Girls who indicated any category other than 100% heterosexual (n = 64/156, 41%) were considered SMGs.

Gay-related victimization experiences

Victimization was measured by four items that assessed the frequency during the past 6 months of being teased/bullied, hit/beaten up, treated unfairly, or called bad names because someone thought the participant was gay/lesbian. Responses were scored as 0 = never, 1 = once or twice, 2 = a few times, and 3 = many times (Cronbach's alpha = 0.86).

Depression symptoms

The CESD[7] was administered to assess frequency of depression symptoms during the past week. A mean score of 20 common symptoms of depression scored on a scale of 0 (rarely or none of the time) to 3 (most or all of the time) was used.

Substance use

Frequency of cigarette, alcohol, heavy alcohol, and marijuana use in the past six months were measured. Response scale ranged from 0 (none) to 10 (several times a day).


Sixty-seven percent of SMGs and 51% of heterosexual girls endorsed cigarette, alcohol, and/or marijuana use in the previous six months. Mediation was tested in a regression framework using the PROCESS macro for SPSS[8] and followed recommended statistical methods.[9] Participants' age and race were included as covariates. Mediated effects were calculated by multiplying the unstandardized beta coefficients that represented each of the paths (Figure 1). Mediated effects were tested for significance by estimating bias-corrected 95% confidence intervals utilizing bootstrapping. Four models were estimated to test the pathways from sexual orientation to gay-related victimization (Mediator 1), to depression symptoms (Mediator 2), to each of the outcome variables. With the exception of marijuana use, all of the individual path coefficients were significant: sexual orientation to victimization, β = 0.21, p < 0.01; victimization to depression symptoms, β = 0.27, p < 0.01; and depression symptoms to cigarette use β = 1.16, p < 0.05, frequency of alcohol use, β = 0.64, p < 0.01, and heavy alcohol use, β = 0.61, p < 0.001. Results showed that the serial mediation test was significant for cigarette use (CI: 0.01–0.26), alcohol use (CI: 0.01–0.13) and heavy alcohol use (CI: 0.01–0.13). The relationship between depression symptoms and marijuana use, and the associated mediation test, were not significant.

Figure 1.

Gay-related victimization experiences and depression symptoms statistically mediate substance use disparities among sexual minority girls.


To date there are no empirically supported substance use interventions that are tailored for the unique needs of the LGB youth population. In order to design successful interventions researchers must focus on changing the mediators of substance use. Doing so requires the translation of important empirical results from the “other” basic sciences including epidemiology and psychology.[10] Statistical mediation analysis can be used to test the fundamental tenets of explanatory frameworks such as the minority stress hypothesis[2] and the stress-negative affect pathway.[6] Our results provide preliminary support for both of these models, showing that gay-related stress experiences and depression symptoms statistically mediated alcohol and cigarette use outcomes among SMGs. Of course mediator models are inherently causal models,[9] thus the cross-sectional design of our study must be noted as a limitation. Future research with larger samples and longitudinal data are needed to conduct more rigorous mediation tests that can ensure the temporal ordering of variables and rule out other potential confounding effects. Although translating our results directly into the design of an intervention study would be premature, these results are an important first step in translating self-reported health and wellness data to inform the design and implementation of much needed substance use prevention and intervention programs for SMGs. In combination with the other mediation papers recently published,[4, 5] these results suggest that the reduction of gay-related stress experiences and associated depression symptoms may reduce substance use among SMGs.


This study was supported by the National Institute of Drug Abuse through an R01 awarded to Drs. Michael P. Marshal and Deena J. Chisolm (DA026312). The authors would like to thank Dr. Elizabeth Miller, M.D., for her role as the visionary, leader, and organizer of the Pittsburgh Adolescent Health and Translational Research Symposium.