Are meta-analytic estimates of team performance relationships robust? Exploring the impact of publication bias and outliers

Purpose This study aims to explore whether meta-analytic estimates of relations between team performance and several team-level variables may be distorted by the influence of publication bias and outliers. The concern is that such distortion may negatively impact primary research and practitioner perceptions of associated relations. Design/methodology/approach Seven existing meta-analyses, comprising nine meta-analytic estimates, were examined using current techniques to assess the degree to which publication bias and outliers may have impacted these estimates. The adjusted estimates were used to index the robustness of the original meta-analytic estimates. Findings Results indicated that eight of nine meta-analytic estimates may not be robust, requiring that they be adjusted by at least 20% of their original value. Research limitations/implications For researchers, the authors describe how these adjusted estimates, and the general pattern of their results, might be considered in the design of future primary studies on team performance and inform practices for conducting meta-analyses in team research. Practical implications For practitioners, this paper provides advice on how to evaluate whether publication bias has been adequately examined in a meta-analysis and suggests how an estimate might be adjusted if this is not the case. Originality/value This study contributes to an emerging movement analyzing the robustness of meta-analytic estimates in the organizational sciences by focusing on the robustness of meta-analytic estimates of the relations between team performance and several team-level predictors.