Online consumers have come to expect recommendations from their favorite shopping websites. Amazon recommends books, movies, and all manner of items you might like. Facebook and LinkedIn recommends people you might know. Netflix recommends films. These companies all have complex algorithms that assess your tastes based on your purchasing history and your reviews, and with varying degrees of accuracy they're able to gauge what you like.
Now, one school is testing whether the same sort of algorithm can be applied to choosing college courses. Can we look at a student's academic history and determine what subjects he or she will enjoy and do well in?
Students already get advice from staff and faculty advisers as to which classes they should take. But much of that advice is restricted to fulfilling degree requirements, and it's easy to imagine how advisers' own preferences or their lack of knowledge of a students' might skew the recommendations.
The students at Austin Peay State University in Tennessee have a new program that allows them to zero in on their most likely preferred class, in addition to having access to face-to-face advising sessions. Tristan Denley, Austin Peay's provost and a former math professor, has designed a new course recommendation engine for the college. It uses data based on students' majors, class history, grades, as well as other similar student performance to help students decide on courses.
Denley says he's tested outcomes for students who take the software-recommended classes, and he's found their GPAs were half a point higher than those who chose courses not suggested by the program, according to an article in The Chronicle of Higher Education.