Purpose: Although many studies have reported findings on risk and protective factors associated with adolescent substance use, a systematic analysis of the magnitude of effects across studies has not been reported. Further, the quality of research on adolescent substance use risk and protective factors has not been systematically evaluated. The purpose of this study is two fold: a) to determine the magnitude of effects across studies for risk and protective factors for adolescent substance use, and; b) to evaluate the level of evidence for studies reporting on adolescent substance use risk and protective factors. For this analysis the term substance refers to any licit or illicit drug including; alcohol, marijuana, amphetamines, inhalants, hallucinogens, cocaine, crack, heroin, cigarettes, smokeless tobacco, steroids, tranquilizers, MDMA (ecstasy) or other club drug and LSD. Adolescent substance use risk and protective factors are organized into overarching domains of intrapersonal (e.g., psychopathological, personality, biological), interpersonal (e.g., peer, family, school) and cultural (e.g., SES, neighborhood) factors.

Methods: Computerized and manual methods were used to identify studies for this analysis. The computer search was conducted for the period 1992-2004 using MEDLINE, PsycINFO and Science Direct. The search terms used were adolescent, adolescence, drug or substance abuse, alcohol, protective factors, risk factors, ethnicity, SES, ADHD, and psychiatric comorbidity. The references of selected published articles were also searched for other relevant sources. Level of evidence evaluation was based on the APA task force guidelines (Chambless, Sanderson, Shoham, Johnson, Pope, Crits-Cristoph et al., 1996) and a summative scores method was used (Bangert-Downs, Wells-Parker & Chevillard, 1997).

Results: The final sample consists of 72 articles evaluated for magnitude of effect of adolescent substance use risk and protective factors and study level of evidence. Risk factors at level I evidence are psychopathology, in both domains of disruptive and mood disorders (d = .25 to 1.25), positive alcohol expectancies (d = .1 to .98), peer domains (d = .15 to .71), parental and/or sibling substance use (d =.26 to .5), school dropout and poor school attachment (d = -.23 to -.27), SES (d = .07 to .28) and living in a rural setting (d = .12 to .14). Risk factors at level II are personality (d =.1 to .30) and age of onset (d = .41 to .48). The protective factors at level I are parental monitoring (d = .21 to .77) and positive school attachment (d = .19 to .25), while family support (d = .11 to .17) is at level II evidence. Although space did not permit here, magnitude of effects and level of evidence are also presented by race/ethnic group and underserved populations.

Implications for practice and policy: The findings underscore the relative lack of attention on protective factors. Based on these results there appears to be a need for refocusing prevention strategies to address intrapersonal, interpersonal and macro level protective factors. The impact of family level factors on protection against adolescent substance use suggests the need for universal, selected, and indicated prevention programming that includes family system intervention.