Disa Hansson: Inferring sexual preferences from egocentric network data on sexual outcomes

Abstract: In this talk, I will present models that will allow us to estimate the relative frequency of the occurrence of unobservable action disposition among individuals in a population. The interactions constitute the direct outcome of the unobservable disposition under study. These models will be applied on two different types of dyadic activities of high relevance for the spread of sexually transmitted infections (STI). One of the most striking differences of sexual risk-taking compared to other health-related risk behaviours is that the very act in itself consists of an interaction between two individuals. To fully understand the mechanism of the two risk behaviour determinants we must try to understand the individual unobserved dispositions, preferences, and not only the observed outcome of the pair. In a heterosexual population, the sexual behaviour occurs between two individuals of the opposite sex. Therefore, the models will be extended to let men and women have different preferences. Additionally, the models will be extended to test if the preferences differ between relationship types (e.g. casual or main partner).