Cluster analysis and persistence in college majors

Abstract

Purpose

The Bryant University Mathematics Department has been collecting math placement scores and admissions data for all incoming freshmen for many years. In the past, the authors have used these data mainly for placement in first-year classes and more recently to invite the most mathematically talented students to become mathematics majors. The purpose of this paper is to use the same data source to predict persistence in declared majors for all incoming students.

Design/methodology/approach

In order to categorize the students, the authors use cluster analysis, one of the tools of data mining, to see if students in particular majors share similar strengths based on the available data. The authors follow up this analysis by running a multivariate analysis of variance (MANOVA) to confirm that the means of the clusters are significantly different.

Findings

The cluster analysis resulted in five distinct clusters, which were confirmed by the results of the MANOVA. The authors also found how many students in each cluster persisted in their chosen major.

Originality/value

These results will help to improve counseling and proper placement of incoming freshmen. They will also be helpful in long-range planning of upper-level courses. Retention of students in their majors is an important concern for colleges and universities as it relates to planning issues, such as scheduling classes, particularly for upper classmen. This could also affect departmental requirements, such as the size of the faculty.

Citation

Quinn, J.T., Olinsky, A.D., Schumacher, P.A. and Smith, R.M. (2015), "Cluster analysis and persistence in college majors", Journal of Applied Research in Higher Education, Vol. 7 No. 2, pp. 275-291. https://doi.org/10.1108/JARHE-05-2014-0058