This paper examines how predictive modeling can be used to study application behavior. A relatively new technique, artificial neural networks (ANNs), was applied to help predict which students were likely to get into a large Research I university. Data were obtained from a university in Iowa. Two cohorts were used, each containing approximately 20,000 records. The results of the new techniques were compared to those from the traditional analysis tool, logistic regression modeling. ANNs have several advantages over traditional statistical methods. ANNs do not require knowledge of the functional relationship between the independent variables and the dependent variables to estimate the model. Unlike traditional statistical techniques, ANNs learn from examples with a small number of prior assumptions about structural relationships. Other advantages and disadvantages are discussed. The addition of artificial intelligence models is an exciting new area that may be used to explore the complex processes in educational institutions. (Contains 3 figures, 7 tables, and 48 references.) (SLD)