Examining Spillover Effects of Job Skills Training on Substance Misuse Among Americans Facing Economic Disadvantages: A Longitudinal Cohort Study Using Propensity Score Weighting

Background/Purpose: Americans experiencing economic disadvantages are at increased risk of substance misuse, and substance-related problems are barriers to economic self-sufficiency. Evidence suggests that services to improve employment and earnings may have positive spillover effects on behavioral health outcomes such as substance misuse by addressing adverse life conditions and increasing treatment access. Evidence also suggests that programs focused on developing job skills, rather than short-term services (e.g., job search), produce better employment outcomes. To increase understanding of whether gains from job skills training extend beyond labor market outcomes to substance use behaviors, a longitudinal study would be useful.

Methods: Data came from the National Longitudinal Survey of Youth 1997 (NLSY97), a population-based longitudinal study of Americans born between 1980 and 1984. Information about respondents’ education/job training history, labor market outcomes, and behavioral health was obtained from 1997 until the latest survey conducted in 2015. Of 8,984 respondents, 581 reported participating in government-sponsored employment programs. To determine whether participation in these programs had spillover effects on substance misuse and whether these effects differed by service type, the 56.4% of participants who received job skills training services (on-the-job training, classroom job skills training, and work experiences [treatment group]) were compared with the 43.6% who received basic services only (e.g., adult basic education, GED preparation, and job readiness training [comparison group]). Three binary measures of substance misuse (past month binge drinking, past-year marijuana use, and past-year other illicit drug use [e.g., cocaine/crack, heroin]) were retrieved from multiple waves at approximately two-year intervals from 2000 to 2015. Statistical analyses were conducted in two phases. First, inverse propensity score weights were constructed to address potential selection into treatment status (i.e., job skills training vs. basic services) and achieve balance in the distributions of 28 baseline individual, family, and school characteristics across groups. Then, inverse-propensity score weighted generalized growth curve models for the three substance misuse behaviors were estimated while including the treatment indicator as a regressor.

Results: Compared to the basic service group, the job skills training group had higher employment rates and annual earnings, especially since year 3 after program participation. Regardless of service type, binge drinking prevalence was estimated to decrease from 40% at program entry to 30% 16 years after. No significant changes or group differences were found in marijuana use as the rates remained between 12% and 18% over the study period. Though a higher proportion of the job skills training group used other illicit drugs at program entry, only this group showed continued reductions in use prevalence, from 5.7% at program entry to 1.5% in year 16.

Conclusions and Implications: Findings suggest that job skills training produced better outcomes with regard to illicit drugs use in addition to better labor market outcomes over the decade following program participation. Given the high overall prevalence of substance misuse, incorporating screening and treatment services for government-sponsored employment program participants may be helpful.