Meta-Analytic and Empirical Estimates of the Resource Depletion Effect Size

Abstract

Published empirical studies of self-control in humans have provided evidence suggesting that the ability to exert self-control relies on a limited resource. Recent failed replications of the resource depletion effect, in addition to conflicting meta-analytic evidence, have called the robustness of the resource depletion effect into question. This dissertation aims to obtain a more accurate estimate of the depletion effect size using both an empirical replication and a novel meta-analytic method, p-curve. Monte Carlo simulations testing the accuracy of p-curve effect size estimates in the presence of publication bias and questionable research practices are also reported. Simulation results show that p-curve effect size estimates are unaffected by publication bias but fluctuate wildly when questionable research practices are simulated. Results from the empirical replication and meta-analysis suggest that the resource depletion effect is not as robust as previously thought. Further work is necessary to reliably differentiate resource depletion effects from type I error.