Synthesizing positive psychological interventions: Suggestions for conducting and interpreting meta-analyses

Abstract

Systematic reviews, such as meta-analyses, are highly valued within scientific, professional, and lay communities because they provide an easily digestible aggregate of a large body of work. A recently published meta-analysis of positive psychology interventions concluded that these interventions have small effects and argued for the use of these interventions in diverse populations (Bolier et al., 2013). We caution researchers against drawing conclusions from this study because of the unusual definition of what is (and is not) a positive psychological intervention. Bolier and colleagues (2013) define their area of inquiry as “pure positive psychology interventions” and limit their sample to studies conducted within the years following the formal founding of the positive psychology movement. This decision – while well intentioned, as it provides specificity to their criteria for inclusion – is, in our view, too narrow, excluding a host of studies that use the same intervention strategies and target the same outcomes but do not explicitly reference “positive psychology”. The inclusion criteria of a systematic review directly impact its findings and conclusions. Using the criterion of papers that explicitly reference positive psychology creates an arbitrary boundary that reflects neither the research nor practice of the field; the best practitioners prioritize effectiveness and efficiency over explicit ties to “positive psychology”. Arbitrary boundaries hinder science and impair the ability of researchers, clinicians, and the general public to draw accurate conclusions from the findings. It also limits the meta-analyst’s ability to conduct moderation analysis that can help drive the field forward by answering research questions that are difficult to address in a single study. Positive psychology and psychology more generally would benefit from definitions of terms that are conceptually-based and thus meta-analyses that are theoretically sound.