A Meta-Analysis of Reinforcement Sensitivity Theory: On Performance Parameters in Reinforcement Tasks

Abstract

J. A. Gray's Reinforcement Sensitivity Theory (RST) has produced a wealth of quasi-experimental studies in more than 35 years of research on personality and reinforcement sensitivity. The present meta-analysis builds on this literature by investigating RST in conflict and nonconflict reinforcement tasks in humans. Based on random-effects meta-analysis, we confirmed RST predictions of performance parameters (e.g., number of responses, reaction time) in reinforcement tasks for impulsivity- and anxiety-related traits. In studies on anxiety-related traits, the effect size variance was smaller for conflict tasks than for nonconflict tasks. A larger mean effect size and a larger variability of effect sizes were found for conflict compared to nonconflict tasks in studies on impulsivity-related traits. Our results suggest that problems with RST confirmation in reinforcement tasks are at least partly caused by insufficient statistical power of primary studies, and thus, encourage future research on RST.