Predicting College Enrollment from Student Interaction with an Intelligent Tutoring System in Middle School
Research shows that middle school is an important juncture for a student where he or she starts to be conscious about academic achievement and thinks about college attendance. It is already known that access to financial resources, family background, career aspirations and academic ability are indicative of a student’s choice to attend college; though these variables are interesting, they do not necessarily give sufficient actionable information to instructors or guidance counselors to intervene for individual students. However, increasing numbers of students are using educational software at this phase of their education, and detectors of specific aspects of student learning and engagement have been developed for these types of learning environments. If these types of models can be used to predict college attendance, it may provide more actionable information than the previous generation of predictive models. In this paper, we predict college attendance from these types of detectors, in the context of 3,747 students using the ASSISTment system in New England, producing detection that is both successful and potentially more actionable than previous approaches; we can distinguish between a student who will attend college and a student who will not attend college 68.6% of the time.
Predicting STEM Career Choice... presentation at the 6th International Conference on Educational Data Mining, Memphis, TN, June 2013.