The application of probit,logit, and tobit in marketing: A review

Recent work on the analysis of qualitative dependent variables having considerable potential in marketing research are reviewed and illustrated. In particular, three related models, probit, logit, and tobit, are discussed in a nontechnical manner. In the last few years, there has been a resurgence of interest among biologists, economists, and other social scientists in developing and generalizing these three models. Yet, despite their apparent utility in marketing research, there have been only one or two rather limited applications to date. The object of this review is to enable researchers to evaluate these models and to relate them to their existing statistical procedures. The models are defined and related to the multivariate methods commonly used in marketing research. Second, the theoretical problems associated with using regression analysis and similar procedures on qualitative data are noted and the assumptions needed to resolve such problems are presented. Third, marketing applications are reviewed and an illustration of a general form of the probit model is evaluated and compared.