Don't shoot the messengers: A response to Ellis and Hoskin's (2018) commentary on the prenatal androgen hypothesis

Abstract

In our two recently published meta-analyses of the literature assessing the “prenatal androgen hypothesis,” we failed to find supportive evidence in favor of the 2D:4D digit ratio—a rather crude proxy for exposure to fetal testosterone—being predictive of various forms of criminal, antisocial, and analogous behaviors (see Pratt, Turanovic, & Cullen, 2016; Turanovic, Pratt, & Piquero, 2017). Ellis and Hoskin (2018) recently published their own assessment of the hypothesis, which stood in part as a commentary on our previous work. Unlike the results from our meta-analyses, these authors argued that the weak to nonexistent effects of the digit ratio that we found in our studies were an artifact of certain methodological decisions that were made, and that our interpretation of the effect sizes that we calculated was unnecessarily pessimistic. As a result, Ellis and Hoskin (2018) maintained that evolutionary neuroandrogenic (ENA) theory—a perspective that relies heavily on the use of the 2D:4D digit ratio as a measure of its key construct—remains well supported in the face of negative evidence. We would like to take a little time to set the record straight about our methods and interpretations, and to do that we need to begin by explaining why we conducted these meta-analyses in the first place.