Back in September
Dinkelman, Taryn, and David Lam. 2009. "A Model for Understanding Gender Discrepancies in Sexual Behavior Reports." PSC Research Report No. 09-669. January 2009.
Data from nine recent African Demographic and Health Surveys indicate that men report between 10 and 80% more sex partners than women do, and up to 3.6 times as many men report condom use at last sex than women. This paper formalizes the notion that in a closed, heterosexual population without misreporting, the number of sex partners reported by men and women should balance and, perhaps less intuitively, that condom use reports do not have to balance. The simple equilibrium equation that we develop provides a formal way for thinking about the adding-up constraints that must hold when multiple reports are given on the same event. We take plausible values for key parameters from the literature– the fraction of sex workers in the population, the number of sex worker clients and the fraction of recent sex-worker client contacts – to show that sampling bias can account for the range of gender gaps observed in the DHS. The simulation exercises highlight the role that an under-sampled sex worker population could play in explaining observed gaps in partnership and condom use reports and draw attention to the set of conditions under which we should expect reports to balance.