Effects of After School Program Participation On Mental Health and Substance Use Among Adolescents From 18 Public High Schools in Boston

Background and Purpose: Although After School Programs (ASP) provide diverse services to participants (e.g., academic, social, cultural, and recreational activities) to promote healthy development among adolescents, most studies on ASPs have overlooked how participation in those programs affect participants’ behavioral problems, mostly focusing on their academic benefits. However, considering that mental health problems in childhood and adolescence tend to persist in adulthood and that most cases of substance abuse start in adolescence, it is also important to determine whether ASPs are effective in reducing those behavioral problems. Thus, we examined the effects of participation in ASPs on mental health and substance use among adolescents from 18 public high schools in Boston. Also, we examined whether the effects of ASPs vary by type of program (school vs. community). To address the issue of selection bias, we used diverse empirical strategies, including school-fixed effects models and propensity score matching models.

Methods: We used data from the 2006 Boston Youth Survey, constructed by the Harvard Youth Violence Prevention Center. Our final sample was composed of 1,185 students who responded 1) yes or no to ASPs participation and 2) to at least one of the outcome measures. Five outcome measures were constructed: depression, aggression, use of alcohol, use of tobacco, and use of drugs. The independent variable was a binary indicator with a value of 1 if a student attended ASPs and 0 otherwise (290 students (25%) participated in ASPs). To address missing data in covariates, we conducted multiple imputation by using the ICE and MICOMBINE commands in Stata. To address the selection bias, we specified five models: Model1, only including ASP participation; Model2 added student characteristics to Model1; Model3 added school and family characteristics to Model2; Model4 was identical to Model3 but further controlled for school fixed effects to reduce bias from unobserved heterogeneity across schools; and Model5, the final one, was identical to Model4 but estimated regression-adjusted differences by using weights generated from the radius matching method within a caliper (0.001) with a common support option.

Results: We found that ASP participation was associated with reduced depression and less use of tobacco and drugs. Specifically, compared to non-participants, ASP participants showed a 0.21SD decrease in their total depression scores and their probability of using tobacco and drugs was reduced by 69% and 47%, respectively. We also found that the effects differed by type of program. In common with the main results, school-based ASPs were associated with reduced depression (a 0.19SD decrease) and less use of tobacco (a 63% reduction) and drugs (a 79% reduction), but community-based ASPs were only associated with less use of tobacco (a 68% reduction).

Conclusions and Implications: Our findings show that ASPs participation reduces students’ behavioral problems and that school-based ASPs are more effective than community-based ASPs. Based on these findings, we suggest that 1) school-based ASPs need to be enhanced to provide better services to more students and 2) policy supports are needed to improve the quality of community-based ASPs.