p. 1 Online Persistence in Higher Education Web-supported Courses Arnon Hershkovitz, Rafi Nachmias Knowledge Technology Lab, School of Education, Tel Aviv University, Israel {arnonher, nachmias}@post.tau.ac.il Abstract This research consists of an empirical study of online persistence in Web-supported courses in higher education, using Data Mining techniques. Log files of 58 Moodle websites accompanying Tel Aviv University courses were drawn, recording the activity of 1189 students in 1897 course enrollments during the academic year 2008/9, and were analyzed with statistical procedures and the Decision Tree algorithm. This yielded five groups of students whose behavior throughout the semester was described: Low-extent Users, Late Users, Online Quitters, Accelerating Users, and Decelerating Users. Results suggest that 46% of the students either decelerated their online activity or totally quitted it; on the other hand, 42% either accelerated their activity or utilized the course website only towards the end of the semester. Additional state-or-trait analysis showed that type of persistence of online activity might be explained by both personal and course characteristics. We discuss these results. Keywords: Educational Data Mining, Learning Management Systems, Online Activity, Persistence.