Testosterone-cortisol interactions and risk-taking: A reply to Hayes et al.


We thank Hayes and colleagues for their comments on the first studies to examine interactive effects of cortisol and testosterone on risk-taking (Mehta et al., 2015). We look forward to experimental research testing the causal relationships we propose. In our analyses, we controlled for time of day, and further, we now note that the time does not interact with our reported dual-hormone interactions (ps > .112). Although the wider timespan of data collection may increase error variance, this would more likely obscure the interactive effects we report rather than contribute to their occurrence. In Study 2, due to informed consent, participants knew in advance they would be completing a computerized game (the Balloon Analog Risk Task), although they were not informed of the name of the measure or that it was a measure of risk-taking. However, in Study 1, the possibility of hormones reflecting an anticipatory rise is less likely, as the study measured baseline hormone concentrations and risk-taking personality on different days. Hayes et al. suggested that testosterone differences in gender and race/ethnicity may have influenced our results. However, as reported in the paper (see Section 1.2.3 as well as footnote 3), the pattern of results was similar in men and women (Study 1; Mehta et al., 2015). For race, the work Hayes and colleagues cite (Litman et al., 2006) does not report any significant differences in serum or bioavailable testosterone differences among Blacks, Hispanics, and Caucasians using a large sample size (N = 1881). If racial differences in testosterone exist but are undetected in such a large sample size, the differences are likely slight and of negligible influence on our data. Because of our predominantly Caucasian samples, any effects of racial differences in testosterone would have to be extremely large to systematically influence our results. Nevertheless, we did examine if differences between Caucasians and non-Caucasians existed for cortisol, testosterone, and our risk-taking measures in each study, but none were found (ps ≥ .380). Furthermore, controlling

DOI: 10.1016/j.psyneuen.2015.05.012

Cite this paper

@article{Welker2016TestosteronecortisolIA, title={Testosterone-cortisol interactions and risk-taking: A reply to Hayes et al.}, author={Keith M. Welker and Samuele Zilioli and Justin M. Carr{\'e} and Pranjal H. Mehta}, journal={Psychoneuroendocrinology}, year={2016}, volume={63}, pages={381-2} }