New meta-analysis:
In tasks from the fields of behavioral economics, neuroendocrinology, and social neuroscience, the correlation between participant testosterone levels and their risk preferences came out to...
0.
There was no relationship, and no publication bias!