This paper examines social learning regarding HIV infection, using HIV test results and sibling death data from Malawi. In the analysis, we compare hypotheses on social learning, selection. and common factors. Empirical results show that young women are less likely to be HIV-infected if they observed prime-age deaths among their siblings, whereas HIV infection is found to be positively related to prime-age sibling deaths among older women. This supports the social-learning hypothesis. Notably, schooling reinforces the social-learning effect of sibling deaths on HIV infection in women regardless of age. The above findings are robust to age (cohort) effects and unobserved location factors.
Evidence from Malawi
International Food Policy Research Institute (IFPRI)