Likelihood-based inference of odds ratios in logistic regression models is problematic for small samples. For example, maximum-likelihood estimators may be seriously biased or even non-existent due to separation. Firth proposed a penalized likelihood approach which avoids these problems. However, his approach is based on a prospective sampling design and its application to case-control data has...

Epidemiologists benefit greatly from having case-control study designs in their research armamentarium. Case-control studies can yield important scientific findings with relatively little time, money, and effort compared with other study designs. This seemingly quick road to research results entices many newly trained epidemiologists. Indeed, investigators implement case-control studies more fr...

Case-control association studies use genetic markers as putative etiologic risk factors. The approach is controversial and has tended to produce associations in neuropsychiatry that do not stand the test of time. We studied the processes that can bias the outcomes away from a true representation of the relationship between a genetic marker and a neuropsychiatric disorder. If conducted with care...

We consider likelihood based inference in a class of logistic models for case-control studies with a partially observed covariate. The likelihood is a combination of a nonparametric mixture, a parametric likelihood and an empirical likelihood. We prove the asymptotic normality of the maximum likelihood estimator for the regression slope, the asymptotic chisquared distribution of the likelihood ...

Power calculations are derived for matched case-control studies in terms of the probability po of exposure among the control patients, the correlation coefficient phi for exposure between matched case and control patients, and the odds ratio psi for exposure in case and control patients. For given Type I and Type II error probabilities alpha and beta, the odds ratio that can be detected with a ...

A stratified version of nested case-control sampling which we call "countermatching" is presented. This design uses data available for all cohort members to obtain a sample for collecting additional information in a case-control substudy. Hitherto the only stratified sampling design for such studies has involved matching of controls to cases with respect to confounding variables. However, in so...

The difference between the disease rates of two groups, one exposed and one not exposed to a suspected aetiological factor, is a useful measure of the risk associated with the factor. This measure is of most interest when the association is considered causal, and has been termed the attributable risk (MacMahon and Pugh, 1970). Expressed as a percentage of the total disease rate in a group (the ...