Date of Award

2016

Document Type

Distinction Paper

Department

Business, Accounting, & Economics

Advisor

Michael A. Levin, Ph.D.

First Committee Member

Sonja Martin Poole, Ph.D.

Second Committee Member

Meredith Meyer, Ph.D

Keywords

college rankings, higher education, factor analysis, cluster analysis

Subject Categories

Business Administration, Management, and Operations | Higher Education | Higher Education Administration | Marketing

Abstract

Purpose – The objective of this paper is to propose a new method for classifying higher education institutions that differs from traditional rankings. The proposed model segments colleges and universities based on like attributes and describes the groups according to their differentiating features.

Research Method – Data from several rankings publications was collected and simplified for analysis. Several observations (i.e. schools) were eliminated because they did not fit the sampling frame and/or did not include a sufficient amount of information. Duplicate variables (i.e. attributes) were also eliminated. First, exploratory factor analysis was applied to reduce the number of variables being examined. Second, cluster analysis was employed to segment the observations.

Findings – Exploratory factor analysis and cluster analysis revealed five clusters of schools based on seven main underlying factors. Several post-hoc analyses determined that the EFA and cluster models were stable. These analyses confirmed that the constructs measured are in fact distinct, as are the five generated clusters.

Practical Implications – The five-cluster model generated from this study has practical and beneficial applications for both higher education managers and prospective college students and their parents. The insights gained from the model can help colleges and universities more effectively target students in their marketing efforts, and prospective students and their parents can make better-informed decisions about college plans.

Originality/Value – This paper is the first to propose the use of EFA and cluster analysis to segment higher education institutions. The validity of traditional college rankings has been questioned in recent years, and the model proposed in this paper solves many of the inherent problems with conventional rankings.

Download

Share

COinS