ABSTRACT
Case-control studies examine the association between exposures or other lifestyle choices and the risk of disease. When the disease of interest is rare, the case-control study allows us to evaluate these associations using fewer subjects than a follow-up study by over-sampling cases and under-sampling controls from the population. This chapter describes logistic regression analysis of unmatched case-control data, and Chapter describes logistic regression analysis of matched and frequency-matched case-control data. I begin by reviewing several measures of disease occurrence and the association of exposures to disease in the population, showing how different association measures can and cannot be estimated solely from data collected in a case-control study. Sections containing practical advice for the use of logistic regression to analyze unmatched case-control data follow.
