PREDICTION OF ENGINEERING STUDENT PROGRESSION FROM ENTRANCE DATA

Abstract

Most Western economies are suffering from problems in attracting students to study Science, Engineering and T echnology courses at university level. Concurrently, engineering programmes te nd to have higher dropout rates than courses in humanities, business or health sciences. They also typically have a higher unit cost per student. Many reasons have been suggested in the literature why the retention rates are lower in such programmes. However little data exists in the Irish context for what information is useful in identifying which factors are predictive of retention and to what degr ee they influence retention probability. What information is available, typically considers single factors (e.g. mathematical attainment, overall grades, English attainment etc), but does not consider any interaction effects. The work reported in this paper examines entrance data for approximately 22,000 students in Trinity College over a 10 year period. Those variables which are predictive of retention are identifie d – both for engineers and for students generally. It is shown that appropriate u se of this information provides significant extra discrimination (over either random sel ection or any single factor model) in identifying those students most likely to encounter progression difficulties.