Logistic Regression Analysis in Experimental Bronchial Carcinogenesis

The technique of logistic regression is shown to be a usable method for testing the significance of associations in the two-sample problem of comparing rates of progression through an ordered series of states as a function of another ordered variable (such as time).The technique is illustrated using an example comparing the development of preneoplastic and neoplastic changes in two groups of mice exposed to a carcinogen and subsequently receiving or not receiving a potential modifier of carcinogenesis.