Meta-analytical issues for prevention intervention research.

Lack of systematic methods for comparing diversified programs has limited the use of research results. Drug prevention intervention research has a history of mixed or marginal results, a situation that leads to the supposition that programs do not work. Meta-analytical methods have successfully resolved problems of conflicting results and are a cost-effective method for building a knowledge base. Using both qualitative and quantitative methods, meta-analysis applies all of the scientific rigor of primary research to the integration of this research. Quantitative synthesis is accomplished by computing an effect size, which, unlike significance test, allows comparisons across studies having varied sample sizes. One advantage for drug prevention intervention research, which seldom shows statistically significant results, is the powerful findings produced when small positive effect sizes are consistent across many studies. Generalizability is possible through meta-analytic aggregation, as a large body of studies contain all the exigencies of real-world research. Troublesome areas that can distort conclusions are presented to alert readers of literature reviews so they are able to interpret meta-analytic reviews accurately. Specific problematic issues are introduced, such as preexisting differences, combining efficacy and implementation studies, and the use of the weighted effect size with a group of studies that has a large range in sample sizes. Meta-analytic procedures are illustrated by comparing the results of 114 experimental and quasi-experimental school-based adolescent drug prevention programs with a selected subset of 56 higher quality experimentally evaluated programs.