Exploring connections between moderators and mediators: commentary on subgroup analyses in intervention research.

Although subgroup analyses, and moderator analyses more generally, in intervention research are fraught with analytic and conceptual challenges, they provide invaluable insights into the effectiveness of intervention strategies and the theoretical models upon which they rest. The papers in this Special Issue engage these challenges and offer investigators a set of strategies that will enhance how they think about, measure, test, and report evidence of moderated intervention effects. In this commentary, I echo the call for more thoughtful work on moderated intervention effects and, in particular, urge investigators to pursue opportunities to integrate questions regarding mediation as they work to specify how intervention strategies operate across populations, settings, and behavioral domains.