Basal cortisol's relation to testosterone changes may not be driven by social challenges

Multiple studies show a negative correlation between basal cortisol and testosterone changes in the presence of competition and social-evaluative stressors. These negative associations are proposed to be derived from psychological responses to competition and social-evaluative stress. However, we argue that the association between basal cortisol and testosterone change may instead be a statistical consequence of positively associated variables. In this paper, we present a mathematical rationale for this alternative explanation and examples from two studies that are consistent with this alternative explanation. Both studies show that the associations between basal cortisol and testosterone change have covariance patterns consistent with this alternative possibility. We conclude that the often-found positive association between basal cortisol and basal testosterone opens the door for alternative explanations of the basal cortisol-testosterone change association rooted in the patterns of associations between hormones measured over time. We also suggest future research directions and methods for testing alternative explanations.