A large amount of research has been done to improve Intelligent Tutoring Systems (I.T.S.). Though implementing the latest Artificial Intelligence (A.I.) techniques, I.T.S.'s fail to approach the adaptability of human teachers. New methods are needed for observing computer users, analyzing and interpreting their behavior, and effectively offering online instruction.

This project, called Mentor, is aimed at finding new methods for identifying observable behavior and what can be learned from those observations. Time stamped keystrokes were collected from a small group of Introduction to Computer Programming students for various assignments over the course of one semester. Specific instances of behavior (e.g., run , arrow keys, deletes, etc.) were then graphed in order to look at the student's behavior as he or she progressed through the semester as well as to compare the students to each other. The goals of this study are (1) to find behavioral patterns in computer use, and (2 ) to set the groundwork for a method of observation and analysis that can be generalized for various applications.