Decades of Trivers-Willard research on humans: what conclusions can be drawn?

The Trivers-Willard hypothesis predicts that parents in good condition are positively biased towards sons, while parents in poor condition are positively biased towards daughters. An extensive literature testing this hypothesis has accumulated in the last five decades. We take stock of results concerning humans in a systematic review, which yielded 87 articles, reporting a total of 821 hypothesis tests. A p-curving analysis did not reveal a pattern of p-values consistent with p-hacking. Effects are consistent with the Trivers-Willard hypothesis overall. We then went on to check whether there was a difference between sex ratio and post-birth investment. Theoretical work suggests that, the conditions under which the Trivers-Willard hypothesis is verified should be more restrictive in the case of post-birth investment than for sex ratio. We explored this question in two ways and obtained mixed results. We put forward recommendations for future studies that aim to further assess the validity of the Trivers-Willard hypothesis or mechanisms subtending it, and we discuss the implications of different ways of measuring parental status and investment.