What is the difference in the interpretation of the coefficient for the interaction term in a logistic regression model when data are derived from a cohort study vs a case-control study? You can't test for interaction using logistic regression in a case-control study because the alpha is not interpretable The coefficient for the interaction term is not interpretable in a case-control study The exponentiated beta only represents the effect of interaction in a case-control study when incidence density sampling is used There is no difference in the interpretation of this coefficient based on study design