Mortality perceptions in the context of the Trivers-Willard hypothesis

The Trivers-Willard hypothesis (TWH) posits that parents who can easily provide for their children are expected to have more sons and invest more in sons while parents who have difficulty providing for their children are expected to have and invest more in daughters. In the present study, we test plausible proxies of parental condition largely ignored by prior investigations: worry about family member mortality and perceptions of community mortality risk. We administered an in-person survey to 302 married women living in Matlab, Bangladesh, and captured both mortality and financial-related proxies of parental condition and three parental bias variables (hypothetical investment in sons vs. daughters, preferences for sons vs. daughters, and the sex ratio of respondents' children). Using Bayesian regression, we found only two of nine of our a-priori predictions were supported: there is a 61 % probability that perceptions of community mortality risk are positively associated with preferring daughters over sons and a 66 % probability that financial standing is positively associated with preferring more sons than daughters. Other tests did not reveal clear effects or revealed an effect opposite to what was expected. However, in line with our TWH-inspired hypotheses, exploratory tests revealed that early life exposure to mortality is positively associated with having more daughters than sons and early life financial standing is positively associated with having more sons than daughters. We conclude that tests of the TWH remain sensitive to how parental condition and bias are operationalized, and that assessments of parental condition earlier in life may be particularly predictive of parental bias patterns in accordance with the Trivers-Willard hypothesis.