Prevention of substance use

Abstract

Subgroup analyses are often performed in prevention studies to assess whether intervention effects vary across subpopulations. The simple question “what works” has been replaced more and more by the advanced “what works for whom,” with the consequence of paying more attention to methodological and statistical approaches that can be used to answer this question. In the last two decades, advanced statistical models and their associated software tools have become widely available and, together with increasing computing power, have been frequently applied in prevention research. This chapter attempts to give an introductory overview on this topic in a nontechnical way. The description of statistical methods for the analysis of subgroup data can be skipped by readers who may not be interested in the application of these techniques. In a final section, strategies for dealing with the risks and limitations of subgroup analysis are discussed and some agreed-upon recommendations for reporting of results are provided.