Tobacco and Marijuana Use Among US College and Noncollege Young Adults, 2002–2016

Data were from the 2002–2016 National Survey on Drug Use and Health (NSDUH), an annual national household-based survey of the civilian, noninstitutionalized US population ≥12 years of age. An independent, multistage area probability sample was used to produce nationally representative estimates. NSDUH participants include residents of households and noninstitutional group quarters (eg, college dormitories) and civilians living on military bases. Selected participants completed the survey in their homes through audio, computer-assisted, self-interview methods, which increase privacy and improve self-report of sensitive behaviors, such as drug and tobacco use. Overall response rates ranged from 61.2% (2014) to 71.9% (2002). Data before 2002 were not included because of differences in survey methodology.

The analyses in this report were restricted to persons aged 18 to 22 years for both the primary study population (college students) and the comparison group (noncollege young adults). College enrollment status was self-reported. College students were defined as those who reported being enrolled in school at the college level (full-time or part-time). Noncollege young adults were defined as those not enrolled in school at the college level, including those still in high school or other grades, high school graduates, school dropouts, or those enrolled in some other type of institution. Sample sizes of college students by year ranged from 3707 (2016) to 6605 (2008). Sampling for any given year of the NSDUH was independent from any other waves of the survey (ie, repeated survey rather than repeated longitudinal observations).

Three specific products were assessed: cigarettes, cigars (big cigars, cigarillos, and/or little cigars), and marijuana (in any form). These product types were collapsed into 2 broad product classes: (1) smoked tobacco (cigarettes and/or cigars) and (2) marijuana (regardless of form used). Cigarettes and cigars were the only tobacco products assessed because other tobacco products were either not captured in the NSDUH (eg, hookah and electronic cigarettes) or had no information on use behavior extending beyond the past 30 days (eg, pipes, smokeless tobacco, and roll-your-own tobacco). The most recent data from National Adult Tobacco Survey show that, among young adults aged 18 to 22 years, current cigarette and/or cigar smokers represented 91% of current smokers of any combustible tobacco.²⁶  Hence, although cigarette or cigar smoking underestimates any smoked tobacco use, this underestimation is minimal.

Participants were asked about use of smoked tobacco and marijuana with separate questions: “How long has it been since you last (smoked a cigarette, smoked any type of cigar, and/or used marijuana or hashish)?” We used responses to these questions to create measures of past-30-day use and past-12-month use for both smoked tobacco and marijuana. We assessed these 2 separate periods to increase sensitivity in capturing occasional users (ie, those whose frequency of use was less than monthly).

We examined 3 patterns of use behavior: (1) overall smoked tobacco or overall marijuana use (use of 1 product regardless of use of the other), (2) exclusive smoked tobacco or exclusive marijuana use (use of 1 product without use of the other), and (3) concurrent (dual) use of smoked tobacco and marijuana (use of both products).

Assessed sociodemographic characteristics included sex (male or female) and race and/or ethnicity (non-Hispanic white, non-Hispanic African American, Hispanic, and non-Hispanic other [including Asian American, multiracial, American Indian and/or Alaskan native, and Native Hawaiian and/or other Pacific Islander]).

Data were weighted to provide nationally representative estimates and to adjust for differences in the probabilities of selection and nonresponse. Prevalence was computed for overall use, exclusive use, and dual use of smoked tobacco and marijuana among the study population overall and by college enrollment status. Among college and noncollege young adults, prevalence was stratified by sex, race and/or ethnicity, and academic characteristics (full-time or part-time status for college students and academic attainment for noncollege individuals). Statistical differences between college and noncollege individuals were tested with χ² tests at P < .05. Estimates with relative SEs ≥30% were not reported. Trends during 2002 to 2016 were examined by using relative percentage change (RPC) and binary logistic regression via orthogonal polynomials. Orthogonal linear coefficients were constructed to model trends while controlling for sex and race and/or ethnicity.

Within these analyses, we did not account for the use of blunts because information on blunts was not fully available throughout the study period. Given that some blunt users might not consider or not know that blunts contain tobacco,²⁷  this could lead to underestimates of smoked tobacco use. However, most blunt users in our study population (79.9% and 82.3% for college students and noncollege individuals, respectively) reported using both smoked tobacco and marijuana within the past 12 months; the extent of misclassification was therefore deemed to be minimal. It should be noted that assessing the full extent of misclassification of exclusive tobacco use and marijuana use is difficult because of the overlap of different tobacco use behaviors in the current landscape. For example, there were no data in the NSDUH on whether respondents used marijuana within hookahs or electronic cigarettes. Therefore, the definition of exclusive use should be interpreted within the context of the assessed constructs in this study.

Among adults aged 18 to 22 years overall, the proportion of college students ranged from 44.5% (2002) to 51.2% (2012). During 2016, most college students were female (53.9%), white (53.6%), and enrolled full-time (80.5%), whereas the majority of noncollege young adults were male (56.1%) and white (54.2%), and 44.1% of them reported high school as their highest education attainment. Distributions were similar across study years.

Among college students, the prevalence of overall smoked tobacco use (regardless of marijuana use) in 2002 was 37.9% for past-30-day use and 50.5% for past-12-month use (Tables 1 and 2). In 2016, the corresponding prevalence was 19.1% (past 30 days) and 32.2% (past 12 months). Overall marijuana use (regardless of smoked tobacco use) in 2002 was 18.0% for past-30-day use and 32.7% for past-12-month use. In 2016, the corresponding prevalence was 20.2% (past 30 days) and 34.6% (past 12 months).

Among noncollege young adults compared with college students, significantly higher prevalence was reported for overall smoked tobacco use, as well as overall use of smoked tobacco and/or marijuana for both past-30-day and past-12-month use, in 2002 and 2016 (Tables 1 and 2).

Among college students, exclusive tobacco smoking declined during 2002 to 2016, from 23.5% to 10.4% (RPC = −55.7) for past-30-day use and from 23.2% to 12.2% (RPC = −47.4) for past-12-month use (Tables 3 and 4, Fig 1, Supplemental Fig 2). Declines were significant within all population subgroups assessed. During 2016, prevalence was significantly higher among male than among female college students for both past-30-day use (13.5% vs 7.8%) and past-12-month use (15.8% vs 9.1%). By race and/or ethnicity, prevalence was highest among white college students for both past-30-day (12.5%) and past-12-month (14.8%) use (all P < .05).

Among noncollege young adults during 2002 to 2016, exclusive tobacco smoking declined from 33.6% to 17.7% (RPC = −47.3) for past-30-day use and from 30.7% to 17.4% (RPC = −43.2) for past-12-month use (Tables 3 and 4, Fig 1, Supplemental Fig 2). Declines were significant within all noncollege population subgroups. During 2016, exclusive tobacco smoking prevalence was highest among non-Hispanic whites for both past-30-day (21.8%) and past-12-month (21.4%) usage. By education, prevalence was highest among high school graduates for past-30-day (21.3%) and past-12-month (20.1%) use. In 2016, compared with college students, noncollege young adults had a significantly higher prevalence of exclusive tobacco smoking for both past-30-day (17.7% vs 10.4%) and past-12-month (17.4% vs 12.2%) use (all P < .05).

Among college students, exclusive marijuana use increased during 2002 to 2016 from 3.6% to 11.5% (RPC = 219.7) for past-30-day use and 5.5% to 14.6% (RPC = 166.6) for past-12-month use (Tables 3 and 4, Fig 1, Supplemental Fig 2). Increases were significant within all population subgroups assessed. During 2016, past-30-day exclusive marijuana use did not differ by demographics, whereas past-12-month use was significantly higher among females than among males (16.9% vs 11.9%) and among non-Hispanic African Americans (21.6%) than among all other racial and/or ethnic groups (all P < .05).

Among noncollege young adults during 2002 to 2016, exclusive marijuana use increased from 3.2% to 8.6% (RPC = 167.2) for past-30-day use and from 4.6% to 10.8% (RPC = 133.7) for past-12-month use (Tables 3 and 4, Fig 1, Supplemental Fig 2). Increases were significant within all subgroups assessed. During 2016, past-30-day exclusive marijuana use did not differ by demographics, whereas past-12-month use was highest among non-Hispanic African Americans (14.3%). Compared with college students during 2016, exclusive marijuana use was significantly lower among noncollege young adults both for past-30-day (8.6% vs 11.5%) and past-12-month (10.8% vs 14.6%) use (all P < .05).

Among college students, dual use decreased during 2002 to 2016, from 14.4% to 8.7% (RPC = −39.4) for past-30-day use and from 27.2% to 20.0% (RPC = −26.4) for past-12-month use. Declines were significant among all population subgroups assessed except non-Hispanic African Americans and Hispanics (Tables 3 and 4). In 2016, prevalence was significantly higher among males than females for both past-30-day (11.6% vs 6.3%) and past-12-month (24.0% vs 16.5%) use. By race and/or ethnicity, prevalence was highest among white college students for both past-30-day (11.0%) and past-12-month (24.3%) use (all P < .05).

Among noncollege young adults, dual use decreased during 2002 to 2016, from 15.7% to 13.0% (RPC = −16.9) for past-30-day use and from 27.1% to 23.4% (RPC = −13.6) for past-12-month use (Tables 3 and 4, Fig 1, Supplemental Fig 2). Within subgroup analyses, declines were significant among males and non-Hispanic whites (for both past-30-day and past-12-month use) and those with less than a high school education (for past-12-month use); in contrast, prevalence increased among Hispanics (for both past-30-day and past-12-month use). During 2016, the prevalence of dual use was higher among males than among females for both past-30-day (15.8% vs 9.4%) and past-12-month use (28.1% vs 17.4%). By race and/or ethnicity, prevalence was highest among whites for both past-30-day (14.9%) and past-12-month use (27.0%). By education, prevalence was highest among high school graduates for past-12-month use (25.0%). Compared with college students during 2016, noncollege young adults had a significantly higher prevalence of dual use for both past-30-day (13.0% vs 8.7%) and past-12-month (23.4% vs 20.0%) use.

Among all college-aged young adults during 2002 to 2016, exclusive use of marijuana increased, whereas exclusive use of smoked tobacco products decreased, and rates of these changes were greater among college students. The contrasting trends in tobacco versus marijuana could reflect differences in risk perception and social norms regarding the use of these products. Specifically, the recent increase in legalization of marijuana in various states across the country could be an indication of changing social norms toward marijuana use.^23,28,29 

Our findings revealed different behavioral patterns between college students and noncollege individuals. During 2016, whereas the prevalence of overall marijuana use was not significantly different between college students and noncollege young adults for both past-30-day and past-12-month usage, a larger segment of noncollege marijuana users concurrently engaged in smoking tobacco. During 2002, however, the prevalence of dual use did not vary between college students and noncollege individuals. Furthermore, among noncollege young adults during 2002 to 2016, the prevalence of past-12-month exclusive tobacco smoking was lower than that of past-30-day exclusive tobacco smoking, suggesting that many individuals who recently smoked tobacco had engaged in concurrent use of marijuana earlier in the year. Tobacco products are still widely available for college-aged young adults, and marijuana sale continues to evolve across states. Given this landscape, our findings reinforce the importance of continued surveillance of the use, perceptions, and social acceptability of these products to understand usage behaviors among college-aged adults overall and in demographic subgroups coupled with research on the health effects of dual use of these products.

Use of marijuana in the past 12 months, but not in the past 30 days, could suggest experimental or occasional use rather than frequent use. Comparison of past-30-day and past-12-month prevalence estimates in this study indicate a high prevalence of such intermittent marijuana use for certain population subgroups. For example, in 2016, although no significant differences existed by race and/or ethnicity in past-30-day exclusive use of marijuana, significant differences were found for past-12-month exclusive marijuana use. The largest differences were noted for non-Hispanic African Americans, both among college students (9.8% vs 17.8% [past 30 days versus past 12 months]) and noncollege young adults (10.7% vs 14.3% [past 30 days versus past 12 months]). This pattern was not observed for exclusive smoked tobacco use. In light of these findings, further studies could be beneficial to better understand intermittent marijuana use behavior among college-aged young adults. The relatively high prevalence of intermittent marijuana use also reinforces the importance of continued surveillance of not only recent usage behaviors of marijuana and tobacco but also those extending over longer periods.

Issues of college-aged young adults are of importance for youth. Pediatric practitioners can play a key role in addressing some of the social underpinnings of marijuana use and tobacco smoking among young people, including addressing youth substance use behaviors during their visits. Furthermore, given that college-aged young adults are a reproductive population, the implications of use of or exposure to tobacco and marijuana smoke can extend beyond the immediate harms to the smoker and may also impact pregnancy outcomes.^30,31  When counseling youth, pediatric practitioners can reinforce this point for preparing youth for life as early adults.

Strengths of this study include the use of nationally representative data over a continuous 15-year period to assess trends in marijuana and tobacco use behaviors among college students. The use of a comparison group (noncollege young adults) further strengthens the validity of the findings. Nonetheless, the findings in this study are subject to at least 4 limitations. First, data were self-reported, which might have resulted in recall and social desirability biases. Specifically, we were unable to examine whether decriminalization and legalization of adult marijuana use affected self-reporting bias; that is, respondents might have felt more comfortable reporting marijuana use as it became legal in more states. Second, the definition of tobacco smoking in this study did not include all forms of tobacco products available in the United States because cigarettes and cigars were the only products for which both past-30-day and past-12-month use were assessed. Third, the survey did not assess the mode by which marijuana was consumed (eg, smoking, vaping, or eating), which prevented nuanced analyses of marijuana consumption patterns and their relationships with tobacco use. Finally, because the survey does not include institutionalized populations and persons in the military in its sample, the results might not be generalizable to those populations.

Continued surveillance of tobacco and marijuana use will be important to monitor the impact that legal adult use of marijuana has on usage patterns among young people in the United States.

RPC

relative percentage change

NSDUH

National Survey on Drug Use and Health