In this study, the National Educational Longitudinal Study (ELS:2002) dataset was used and a predictive data mining technique, decision tree analysis, was implemented in order to examine which factors, in conjunction to computer use, can be used to predict high or low probability of success in high school mathematics. Specifically, this study aimed to determine whether the frequency of certain computer use types, and student, and school factors could be used to classify students into high or low probability of success groups. Results from the decision tree analyses suggested that a lower frequency of computer usage for lower level thinking skills, such as reviewing math work or practicing math drills, may have a more positive impact on math achievement for high school students than extensive use. (Contains 4 tables and 2 figures.)