The development of new statistical methods of log-linear models by Leo Goodman has led to major advances in the statistical analysis of categorical data. Goodman's logit analysis, (the simplest form of log-linear models) can be applied in evaluation studies to estimate the "main effects" and "interaction effects" of categorical explanatory variables on a dichotomous dependent variable. Logit analysis is analogous to multiple regression analysis of a continuous dependent variable. It is a "modified regression approach" in place of the linear model of multiple regression, and the estimation procedure is maximum likelihood. The logit model was used with hypothetical data to analyze and predict classroom teachers' attitudes toward proficiency tests based on dichotomous explanatory variables of ethnicity, teaching grade level, and sex of the teachers. The partial regression coefficients were calculated and used to find main effects and interaction effects. Additive and saturated models were used to find the independent logit model in which the criterion analyzed reflects the expected odds of attitude as a function of the explanatory variables. An analysis of the indices of the magnitude of contribution is provided. (CM)