No Evidence for a Causal Effect of Exogenous Testosterone on Risky Decision-Making in Women: An Experiment and Meta-Analysis

While strong claims have been made that testosterone increases risk-taking, the existing literature is inconclusive. Thus, our experiment aimed at addressing some shortcomings of previous work. First, risk-taking was assessed using the Columbia Card Task, which allows to decompose overt risk-taking into three task factors-gain amount, loss amount, and the probability of losing-in both a dynamic, more affective ("hot") and a static, more deliberative ("cold") decision-making context. Second, we conducted a testosterone administration study in 80 females using a triple-blind (i.e., blinded participants, experimenters, and data-analysts), placebo-controlled, randomized, between-subjects design to investigate the causal effect of exogenous testosterone on risk-taking. Reviewers were also blind to the treatment conditions during the reviews. Third, we preregistered our analyses. We investigated (a) the main effect of testosterone, (b) the influence of gain amount, loss amount, and the probability of losing on risky decision-making, each in a more affective and more deliberative decision-making context, and (c) whether testosterone moderated any of those effects. Although we replicated previous studies showing that risk-taking was affected by gains, losses, probabilities, and decision-making context, we found no evidence for a main or interaction effect involving testosterone. This finding was further supported by our meta-analysis, which suggests that the effect of administered testosterone on risk-taking in women is smaller than a small effect size. In conclusion, these results provide no evidence for an effect of exogenous testosterone on decision-making under risk, raising some doubts about the commonly suggested direct link between the two.