The relationship between lifestyle behaviors and the onset of neoplasm development has been researched extensively. This study took a multivariate approach in attempting to identify lifestyle variables which could predict group membership among subjects diagnosed as having cancer and those subjects who have not been diagnosed as having cancer. Eighty adults without cancer and 36 adults with cancer completed the Northeast Louisiana University Cancer Study Questionnaire by providing patient identification, next of kin data, patient data, patient biographical and historical data, medical data, home ecological data, hobbies and recreational data, nutritional data, and eating outside the home data. A total of 19 lifestyle behaviors were measured. Discriminant analysis with a stepwise entry of variables was used to treat the data and resulted in a statistically significant discriminant function with three variables in the prediction equation. Variables measuring Vitamin C dose, educational background, and agricultural occupation were found to contribute significantly to prediction of group membership, with Vitamin C dose and education level being inversely related to cancer incidence and agricultural occupation being positively related to cancer incidence. The findings suggest that lifestyle behaviors can be used to successfully differentiate between cancer and noncancer groups. (Author/NB)