A meta-analysis of 52 studies found no link between testosterone levels and risk-taking. In general, only studies where participants completed specific lottery-based economic tasks showed a modest association between testosterone levels and risk-taking, while other types of behavioral studies did not. Furthermore, the lack of association did not depend on participants’ sex. The paper was published in Neuroscience and Biobehavioral Reviews.
Testosterone is a hormone that belongs to a group of hormones called androgens. It is present in both males and females, but it is usually found in much higher levels in males. In males, testosterone is produced mainly in the testes; in females, smaller amounts are produced in the ovaries and adrenal glands.
Testosterone plays an important role in the development of male reproductive organs before birth and during puberty. During puberty, it contributes to changes such as a deeper voice, facial and body hair growth, increased muscle mass, and the growth of the penis and testes. In adults, testosterone helps regulate sexual desire, sperm production, bone density, red blood cell production, and muscle strength. Testosterone levels naturally vary by age, time of day, health status, sleep, stress, body fat, and the use of certain medications.
Study author Irene Sánchez Rodríguez and her colleagues note that, on average, men tend to be more prone to taking risks than women. Various hypotheses have been proposed to explain this gender gap. One prominent biological theory states that the gap is produced by differences in testosterone levels. Another theory, the “dual-hormone hypothesis,” suggests that the behavioral effects of testosterone actually depend on concurrent levels of cortisol, the body’s main stress hormone.
However, studies have not clearly supported the link between testosterone and risk-taking. While some studies have reported that individuals with higher testosterone levels were somewhat more prone to taking financial or physical risks, other studies have found absolutely no association.
The authors of this study conducted a meta-analysis aimed at synthesizing the existing findings to clarify the association between testosterone levels and risk-taking. They searched scientific databases—Google Scholar, PubMed, and Scopus—using “risk seeking,” “risk attitude,” and “risk aversion” as search terms. They looked for studies conducted on humans and sought to include studies examining testosterone alone as well as those testing the dual-hormone hypothesis.
To be included, studies needed to report a statistical association between testosterone and risk preference, to either measure or administer testosterone, and to use at least one behavioral or self-report measure of risk preference. Studies also needed to be written in English, Spanish, or Italian, and provide sufficient data to allow the researchers to calculate an “effect size” (the strength of the link between testosterone levels and risk-taking). In the end, 52 studies comprising 17,340 participants were included in the analyses.
These studies measured risk in various ways, such as gambling games, balloon-popping tasks, or self-reported questionnaires. They also measured testosterone in different ways: some studies used direct blood or saliva tests, some administered experimental doses of the hormone, and some relied on “morphological proxies” (like the ratio of a person’s index and ring fingers, which is theoretically linked to testosterone exposure in the womb).
When the researchers aggregated the data, the results showed that the overall association between testosterone levels and risk-taking across all 52 studies was practically zero. In other words, having high or low testosterone did not reliably predict whether a person would take a risk.
While the overall effect was zero, the data reported by individual studies were very heterogeneous—some studies reported a positive association (higher testosterone – higher risk-taking), while others reported a negative association (higher testosterone – lower risk-taking).
Further analyses revealed why the results varied so wildly. The researchers found that the method of measurement heavily influenced the outcome. For example, only studies that used lottery-based economic tasks to measure risk-taking showed a modest positive association, while studies measuring risk-taking via other methods (like impulsive games or self-reporting) did not. Similarly, the researchers noted that studies relying on indirect finger measurements sometimes hinted at a link, while highly rigorous studies using direct hormone measurements or administration did not.
Importantly, the lack of an association between testosterone levels and risk-taking did not depend on sex, meaning the relationship (or lack thereof) was no different in males than in females.
“Overall, the evidence challenges the notion that testosterone provides a general hormonal basis for human risk preferences,” the study authors concluded. “Instead, findings support a biopsychosocial framework in which ‘risk taking’ reflects the interaction of task demands, cognitive–affective processes, and situational context, with endocrine effects appearing narrow, context-dependent, and method-specific.”
The study contributes to the scientific knowledge surrounding the behavioral effects of testosterone, suggesting that societal and psychological factors likely play a much larger role in risk-taking than single hormones. However, the study authors note that their search did not yield a sufficient number of appropriate studies to reliably examine the dual-hormone hypothesis (whether a specific interaction between cortisol and testosterone might predict risk-taking).
The paper, “No relationship between testosterone and risk aversion: A meta-analytic review,” was authored by Irene Sánchez Rodríguez, Luca Bailo, Folco Panizza, Emiliano Ricciardi, and Francesco Bossi.