Romantic attraction and adolescent smoking trajectories
Highlights
► Six distinct developmental smoking trajectories are identified in adolescence. ► Non-heterosexual romantic attraction is a significant risk factor, but only for women. ► Change from hetero attraction more predictive than consistent non-hetero attraction. ► Suggests coping and socialization mechanisms link romantic attraction to smoking.
Introduction
Decades of research on sexual orientation and substance use has established that lesbian, gay, and bisexual (LGB) adolescents and young adults are more likely to smoke, drink alcohol, and use other substances than heterosexual youth (D'Augelli, 2004, DuRant et al., 1998, Garofalo et al., 1998, Stall et al., 2001). Reviews of the literature on sexual orientation and substance use conclude that sexual minority youth, especially bisexuals, are 2–5 times more likely to use drugs, alcohol, and cigarettes than heterosexuals (Marshal et al., 2008, Ryan et al., 2001). However, with a few notable exceptions, most earlier work was cross-sectional and based on convenience samples (Rosario et al., 1997, Rotheram-Borus et al., 1994) or area-based samples (DuRant et al., 1998, Faulkner and Cranston, 1998, Garofalo et al., 1998).
Much of the literature addressing the reasons for higher substance use among LGB youth focuses on the stresses associated with stigmatized identities (Bux, 1996, Hatzenbuehler, 2009). According to minority stress theory (Meyer, 2003), disparities in substance use may be due to LGB youth being more likely to experience depression, loneliness, discrimination, and victimization (Gonsiorek, 1988, Savin-Williams, 1994). Developmental models such as the “overload model” additionally suggest that risk behaviors can result from experiencing several developmental transitions in short succession (Schulenberg & Maggs, 2002). In addition to normative developmental tasks during the transition to adulthood, acknowledging and integrating a marginalized identity may contribute to further stress for LGB youth, for which substance use may serve as a coping strategy (Savin-Williams and Diamond, 2001, Schulenberg and Maggs, 2002). Minority stress theory and the overload model are neither mutually exclusive nor exhaustive, but both predict higher substance use among LGB youth as a coping mechanism. Higher use among LGB youth is also predicted by socialization-based explanations that posit greater use as a consequence of more frequent socializing in contexts supporting substance use (e.g., clubs, parties), and accompanying stronger use norms in LGB communities, rather than as a result of coping with stress (Heffernan, 1998, McKirnan and Peterson, 1989, Remafedi, 2007).
Only five longitudinal studies have been published that consider developmental smoking patterns among LGB adolescents. Tucker, Ellickson, and Klein (2008) examined a West Coast cohort of heterosexual and bisexual women over a ten year period (ages 14–23), with sexual orientation assessed once at age 23. Women who reported being bisexual were already more likely to be smokers at age 14 than heterosexual women, and while the smoking rate did not change over time for heterosexuals, it increased a further 50% among bisexuals. Talley, Sher and Littlefield (2010) used four years of data drawn from a public university in Missouri (n = 2854) to model latent trajectories of cigarette use. The measure of cigarette use was an 8-point scale measuring frequency of use over the past three months (0 = never/not in the past three months; 7 = more than 40 times). Results indicated that sexual minority self-identification (anything other than exclusively heterosexual) predicted greater initial and sustained use of cigarettes compared to their sexual majority counterparts, as did reports of same-sex attraction and same-sex sexual behavior. In contrast to much of the prior literature, Talley, Sher and Littlefield did not find evidence of different effects by gender. They also included longitudinal measures of sexual orientation, but did not identify an association between the timing of developing of a minority sexual identity and subsequent smoking behavior.
The other three studies used data from the National Longitudinal Study of Adolescent Health (Add Health), the first nationally representative study to include information about both sexual orientation and substance use. Each of these studies relies on measures of romantic attraction to assess sexual orientation. Russell, Driscoll, and Truong (2002) examined change in the quantity of past month smoking over a one-year period, finding that bisexual women were more likely to increase their smoking over time compared with lesbian women, but changes in smoking did not differ as a function of sexual orientation for men. Easton, Jackson, Mowery, Comeau, and Sell (2008) examined whether smoking initiation over a one-year period varied according to sexual orientation, finding that adolescents with both-sex attractions or relationships were significantly more likely to initiate smoking one year later compared to those with opposite or same-sex attractions or relationships. Differences by sexual orientation were more pronounced for girls than boys. Marshal, Friedman, Stall, and Thompson (2009) used latent growth curve models and all three measures of sexual orientation (attraction, behavior, and self-definition) to examine the growth in frequency of substance use (including smoking) among LGB youth compared to heterosexual youth over a six-year period. Somewhat surprising in light of previous research, smoking escalated over time at a faster pace for homosexual youth, but not bisexual youth, compared to heterosexuals. Results were consistent across the three measures of sexual orientation.
The studies by Talley et al. (2010) and Marshal et al. (2009) are the most sophisticated longitudinal examinations of LGB smoking to date, but there are methodological or conceptual limitations to each, some of which also hamper comparability with much of the prior literature. First, both studies considered gender as a covariate rather than examining gender differences in the association of sexual orientation with smoking. Several studies show that sexual orientation is more strongly associated with substance use for women than men (Easton et al., 2008, Hatzenbuehler et al., 2008), and that sexual identity is more fluid among adolescent girls than adolescent boys (Diamond 2003, Savin-Williams and Ream, 2007). Second, smoking was assessed in terms of frequency (days smoked, either actual number or an ordinal scale response), rather than using a more common measure that includes both the quantity and frequency of smoking. Third, the studies focused on smoking trajectories based on the overall mean smoking frequency for each sexual orientation group, ignoring the considerable heterogeneity in smoking patterns that exists within each of these groups. Marshal et al. (2009) modeled linear trajectories because only three waves of data were available, whereas Talley et al. (2010) were able to estimate non-linear trajectories. Finally, Marshal et al.'s (2009) sexual orientation measures were based on data from the six-year follow-up only and thus were insensitive to the development of sexual identity. The most prevalent model of sexual orientation posits that individuals possess an early predisposition to experience attractions, and sexual identity is the eventual culmination of acknowledging and accepting these attractions (Diamond, 2000, Savin-Williams, 1988) It may be that change in attraction and sexual orientation during adolescence is itself associated with patterns of substance use. Talley et al. (2010) is the only study to include longitudinal measures of sexual orientation but it did not identify a timing effect on days smoked (although they did note timing effects on frequency of binge drinking and cannabis use in the same study, with higher consumption patterns observed among individuals who endorsed a minority sexual identity earlier).
The present analysis addresses each of these limitations and furthers the examination of smoking behaviors across a key component of sexual orientation – types of romantic attraction – in several respects. First, we describe how heterosexual and same-sex/bi-sexual attractions are associated with distinct developmental trajectories of smoking rather than variations around a single trajectory. Second, we examine how changes in romantic attraction over time (e.g., transition from heterosexual attraction to LGB attraction) are linked to these trajectories. Third, we use a measure of smoking that is based on both the quantity and frequency. Finally, we examine these associations separately for men and women. Although preliminary analyses did not find gender differences in the overall number and type of smoking trajectories, we investigate whether the associations between romantic attraction (and changes in attraction) and smoking trajectory membership differ by gender.
We tested several hypotheses based on applying the theories of sexual orientation and substance use just reviewed specifically to romantic attraction. Premised on the idea that if smoking is a mechanism for coping with stress, then changes in stress levels should be followed by changes in smoking levels over time. While the various theories lead to an array of hypotheses, the primary aim here is not to test one theory against the other, but rather to guide description of differences between romantic attraction groups. First, consistent with prior literature we hypothesize that differences by romantic attraction will be more pronounced for women than men (H1). Second, based on the pervasive stress posited by minority stress theory, we hypothesize that consistently LGB-attracted youth (H2a) or youth who transition to LGB attraction (H2b) will be more likely to belong to any smoking trajectory group compared with consistently heterosexually-attracted youth. Third, the overload model suggests that as youth transition to adulthood, consistently LGB-attracted individuals will accumulate greater stress and will be more likely to increase their smoking over time. Thus, we hypothesize that LGB-attracted youth will be more likely than consistently heterosexually-attracted youth to belong to a trajectory group exhibiting either an early increase (H3a) or a delayed increase (H3b) in smoking. Finally, youth that transition to an LGB attraction face additional acute stress associated with the transition period itself, making them even more likely to belong to one of the increasing use trajectory groups (either early [H4a] or delayed [H4b]) than the consistently alternative orientation youth who have already transitioned prior to observation. It should be generally noted that while early heavy smoking may be an indicator of coping with stress, delayed escalation may be more indicative of socialization.
Section snippets
Data
The analyses are based on data drawn from Waves I through III of the National Longitudinal Study of Adolescent Health. Add Health is a nationally representative study of adolescents in grades 7 through 12 (roughly ages 12–18) in the United States in 1995 who have been followed with multiple interview waves into young adulthood. The sampling frame included all high schools in the United States. Over 90,000 participants from 145 schools were given a basic interview at school. Data from this
Romantic attraction by gender
Table 1 presents the unweighted sample frequencies for each romantic attraction group, by gender, in the final study sample. The table also provides detail on the specific types of attraction transitions that comprise each final category. Ninety percent of males and 86% of females reported a consistent romantic attraction type, with nearly all of these cases (97%) reporting consistently heterosexual attraction. Of the 300 males and 469 females whose responses indicated a transition in romantic
Discussion
Longitudinal descriptions of LGB smoking trajectories are important complements to literature showing that LGB individuals are at high risk of smoking. Although prior research has demonstrated this point, it is unclear whether higher levels of adolescent LGB smoking persist over time, or represent a relatively temporary or exploratory period during the transition to adulthood. Latent class growth analysis also illustrates that there are a variety of distinct developmental smoking trajectories
Role of funding sources
Funding for this study was provided by grant 16RT-0169 from the California Tobacco-Related Disease Research Program (TRDRP) of the University of California, awarded to Dr. Tucker as Principal Investigator. TRDRP had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.
Contributors
Dr. Pollard was primarily responsible for the design of the study, conducted the data analysis, and wrote the first draft of the manuscript. Dr. Tucker assisted with the design of the study and literature review, and provided feedback on drafts of the manuscript. Drs. Green, Kennedy, and Go assisted with the design of the study and provided feedback on drafts of the manuscript.
Conflict of interest
All authors declare that they have no conflicts of interest.
Acknowledgments
This research was supported by funds from the California Tobacco-Related Disease Research Program of the University of California, Grant Number 16RT-0169. This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with Cooperative funding from 23 other federal agencies and foundations. Special
References (60)
The epidemiology of problem drinking in gay men and lesbians: A critical review
Clinical Psychology Review
(1996)- et al.
Victimization, use of violence, and drug use at school among male adolescents who engage in same-sex sexual behavior
Journal of Pediatrics
(1998) - et al.
Substance use behaviors among college students with same-sex and opposite-sex experience: Results from a national study
Addictive Behaviors
(2003) Mental health issues of gay and lesbian adolescents
Journal of Adolescent Health Care
(1988)- et al.
Dynamic measures of primary and secondary school characteristics: Implications for school effects research
Social Science Research
(2009) The nature and predictors of substance use among lesbians
Addictive Behaviors
(1998)- et al.
Substance use and abuse trajectories across adolescence: A latent trajectory analysis of a community-recruited sample of girls
Journal of Adolescence
(2010) - et al.
Psychosocial and cultural factors in alcohol and drug abuse: An analysis of a homosexual community
Addictive Behaviors
(1989) - et al.
Friendship networks and trajectories of adolescent tobacco use
Addictive Behaviors
(2010) - et al.
Smoking among lesbians, gays, and bisexuals: A review of the literature
American Journal of Preventative Medicine
(2001)