Best‐practice recommendations for estimating interaction effects using meta‐analysis

Abstract

SummaryOne of the key advantages of meta-analysis (i.e., a quantitative literature review) over a narrative literature review is that it allows for formal tests of interaction effects-namely, whether the relationship between two variables is contingent upon the value of another (moderator) variable. Interaction effects play a central role in organizational science research because they highlight boundary conditions of a theory: Conditions under which relationships change in strength and/or direction. This article describes procedures for estimating interaction effects using meta-analysis, distills the technical literature for a general readership of organizational science researchers, and includes specific best-practice recommendations regarding actions researchers can take before and after data collection to improve the accuracy of substantive conclusions regarding interaction effects investigated meta-analytically.

Examining the “Pros” and “Cons” of TeamConflict: A Team-Level Meta-Analysis of Task, Relationship, and Process Conflict

“…Hunter & Schmidt, 2004;Steel & Kammeyer-Mueller, 2002), and we argue that the current moderators form discrete categories and could be analyzed as such using Hunter and Schmidt's categorical approach (cf. Aguinis, Gottfredson, & Wright, 2011). Indeed, in post hoc analyses, for which data were not reported, de Wit et al noted that "when testing the moderating effect of group task type in isolation (using subgroup analyses), we found a small positive correlation among studies on decision-making tasks" (p. 372).…”

“…This involved conducting separate meta-analyses in a priori subgroups that were of potential theoretical interest. Moreover, it seemed to be the optimal strategy given that all of our moderators were clearly discrete categorical variables, and WLS regression, employed by de Wit et al (2012), is intended for continuous moderators (Aguinis et al, 2011;Hedges & Olkin, 1985;Steel & Kammeyer-Mueller, 2002). Downloaded by [University Of Pittsburgh] at 12:10 12 October 2014…”

Corporate Social Responsibility, Employee Organizational Identification, and Creative Effort

“…Therefore, both convergent and discriminant validities of our measures are achieved. We applied hierarchical moderated regression (Aguinis, Gottfredson, & Wright, 2011;Aiken & West, 1991) to test our hypotheses. We standardized all continuous independent variables to facilitate the generation of interactions (product terms).…”

The influence of entrepreneurs’ social referents on start-up size

“…Random effects analysis allows regression coefficients and intercepts to vary across provinces (Aguinis et al 2011). In studies with more than one level of analysis, researchers have agreed that lower-level entities (e.g., individuals) are nested within higher-level ones (e.g., provinces) (Aguinis et al 2013).…”