What do college ranking data tell us about student retention: Causal discovery in action

Authors:

Marek J. Druzdzel
University of Pittsburgh
School of Information Sciences
and Intelligent Systems Program
e-mail: marek@sis.pitt.edu

Clark Glymour
Carnegie Mellon University
Department of Philosophy
e-mail: cg09+@andrew.cmu.edu

Abstract:

We describe an application of the TETRAD II causal discovery program to the problem of search for causes of low student retention in U.S. universities. TETRAD II discovers a class of possible causal structures of a system from a data set containing measurements of the system variables. The significance of learning the causal structure of a system is that it allows for predicting the effect of interventions into the system, crucial in policy making.

Our data sets contained information on 204 U.S. national universities, collected by the US News and World Report magazine for the purpose of college ranking in 1992 and 1993. One apparently robust finding of our study is that student retention is directly related to the average standardized test scores of the incoming freshmen. When test scores of incoming students are controlled for, factors such as student faculty ratio, faculty salary, and university's educational expenses per student are all independent of graduation rates, and, therefore, do not seem to directly influence student retention. As the test scores are indicators of the overall quality of the incoming students, we predict that one of the most effective ways of improving student retention in an individual university is increasing the university's selectivity.

Keywords:

TETRAD II, causal discovery, knowledge discovery in databases