Effect size benchmarks: Time for a causal renaissance

Effect size benchmarks guide theory, aid in interpreting practical significance, and help gauge scientific progress. However, effect size benchmarks derived from correlations typically violate the definition of an “effect” because they do not capture a singular causal relationship and instead represent an ambiguous amalgamation of additive, multiplicative, and interactive causes. Therefore, correlational benchmarks can be highly misleading to the point of threatening the very livelihood of society at large by misinforming policy and decision-making. To highlight these issues and demonstrate a more productive path forward, we begin by reviewing the four key challenges in creating effect size benchmarks and establishing evidence of causal inference strength. We then illustrate the limitations and opportunities in current practice through a systematic review of the leadership literature that highlights four themes related to causally identified effect sizes. We conclude this work with a blueprint that provides a meaningful redirection of the conversation so that future meta-analytic studies can provide accurate, specific, and unconfounded effect size benchmarks to achieve a more robust and cumulative science.