Using linked meta-analysis to build policy models.

There are many readily identifiable applications of meta-analysis to the area of drug abuse prevention and related topics. Meta-analysis of preventive intervention research (e.g., Tobler 1986), for instance, can identify more and less effective approaches, as can an analogous meta-analysis of rehabilitative treatment research. Meta-analysis of the correlates of intervention-induced change can illuminate the psychological processes involved in the response to intervention. Meta-analysis of the predictive relationships of risk variables with subsequent abuse can indicate which types of variables are most strongly related to the target behavior and chart the developmental course of drug abuse problems (cf. Loeber and Dishion (1987) on antisocial behavior). Meta-analyses of the relationships among risk factors might better identify their structure and the independent clusters they represent. Meta-analysis of the consequences and correlates of drug abuse can trace the patterns of dysfunction in which abuse is embedded. While each of these individual applications may have considerable merit, the intent here is to look ahead to the prospects of linking a number of such meta-analyses into an integrated whole that covers multiple aspects of problem behavior in a coordinated manner.