A comparison is made between 2 different approaches to the linear logistic regression analysis of retrospective study data: the prospective model wherein the dependent variable is a dichotomous indicator of case/control status and the retrospective model wherein the dependent variable is a binary or polychotomous classification of exposure. The 2 models yield increasingly similar estimates of relative risk with increasing degrees of covariate adjustment. When covariate effects are saturated with parameters, relative risk estimates are identical. The prospective model is generally to be preferred for studies involving multiple quantitative risk factors. Analysis of retropective study data is oriented toward estimating the relativde risk (RR), i.e., the ratio of disease incidence among persons exposed vs. those not exposed to risk factor(s) under investigation (Mantel and Haenszel 1959). Adjustments are usually made for confounding factors, related to both disease and exposure, whose effects might otherwise bias results. An important feature of the analysis is to identify and quantify effects of factors which modify the relative risk (Miettinen 1974).