Comment on "Social contagion, adolescent sexual behavior, and pregnancy: a nonlinear dynamic EMOSA model"

The epidemic model of the onset of social activities (EMOSA model) for adolescent sexuality by J. L. Rodgers, D. C. Rowe, and M. Buster (1998) is examined. Strengths of the model include the nonlinear dynamic specification that seems reasonably realistic and the potential when used as a theoretical tool to generate new hypotheses for further testing at the individual level. Limitations include the lack of a well-developed statistical framework, the inability to include a variety of predictors to capture both social contagion and individual differences in the same model, and the focus on aggregate-level data. The EMOSA model for sexuality can be reparameterized to a hazard rate or an event history model that eliminates these shortcomings. Growth curve analysis represents another analytic alternative for the EMOSA model that also does not have these limitations.