In this paper, we applied a number of clustering algorithms on pretest data collected from 264 high-school students. Students took the pre-test at the beginning of a 5-week experiment in which they interacted with an intelligent tutoring system. The primary goal of this work is to identify clusters of students exhibiting similar knowledge patterns. In particular, we show that the DP-means clustering algorithm yields very good results using binary response data. Other clustering algorithms such as k-modes have demonstrated better results when using categorical response data. [For the full proceedings, see ED599096.]