Better Serving Students by Refining Admissions Criteria with Analytics

Admission criteria vary between open and exclusive. Why? One motivation of being more selective is to better allocate scarce resources such a residence space, faculty, and classrooms to the applicants most likely to graduate. Online education has done away with residence halls and classroom constraints. Faculty may still be limiting, but with a national employment market from which to recruit, that is less of a limitation than in the past.

Setting aside other motivations, such as being known as “exclusive,” the obligation an institution has is to admit students it believes will be served by its model and to exclude those will not. In the middle, the murky middle some have called it, are those applicants who are not strongly predicted to do well, and yet not also strongly predicted to do poorly. Being more restrictive does help minimize the risk to applicants and the institution, and if the pool is large enough, then a happy balance can be struck.

An important question is whether the criteria that are used are predictive enough or simply convenient for ranking? Does high school GPA tell us anything about future performance? If it does, high widely does it apply? Test scores? Same questions.

We don’t have to be in the dark on these questions. With access to data, such as the transcript, the GPA, transfer credit, demographics, test scores, extra curricular activities and other variables, we may apply analytics to build a predictive function, apply it experimentally to a set of applicants, and test whether the predictive function does a better job of predicting success, say in the first term, than our admissions criteria.

Testing first let’s us see the consequences with no risk to the institution and its applicants. Important questions can be tested, such as how does the prediction deferentially treat under-served categories of applicants? Do the strongest predictors give us any insights that allow us to adjust our practices for at-risk students proactively?

You may also be surprised to find what matters and what does not in predicting student success. It’s not as simple as high-school GPA, particularly for non-traditionals.

Please subscribe to our blog.