Using interactions of area dose and individual exposure to estimate effects of population health interventions
The (mis)specification of discrete time duration models with unobserved heterogenity: a Monte Carlo study