Challenging the challenge hypothesis on testosterone in fathers: Limited meta-analytic support.

ABSTRACT

In fathers testosterone levels are suggested to decrease in the context of caregiving, but results seem inconsistent. In a meta-analysis including 50 study outcomes with N = 7,080 male participants we distinguished three domains of research, relating testosterone levels to parental status (Hedges' g = 0.22, 95% CI 0.09 to 0.35; N = 4,150), parenting quality (Hedges' g = 0.14, 95% CI 0.03 to 0.24; N = 2,164), and reactivity after exposure to child stimuli (Hedges' g = 0.19, 95% CI -0.03 to 0.42; N = 766). The sets of study outcomes on reactivity and on parenting quality were both homogeneous. Parental status and (higher) parenting quality were related to lower levels of testosterone, but according to conventional criteria combined effect sizes were small. Moderators did not significantly modify combined effect sizes. Results suggest that publication bias might have inflated the meta-analytic results, and the large effects of pioneering but small and underpowered studies in the domains of males' parental status and parenting quality have not been consistently replicated. Large studies with sufficient statistical power to detect small testosterone effects and, in particular, the moderating effects of the interplay with other endocrine systems and with contextual determinants are required.