University rankings from the revealed preferences of the applicants

TL;DR

This paper introduces a novel method to rank universities based on students' application preferences, leveraging a unique Hungarian admission system feature to derive partial preferences and produce robust, theoretically sound rankings.

Contribution

It presents a new ranking methodology that uses students' application lists and the Hungarian system's admission rules, avoiding multicollinearity and ad hoc parameters.

Findings

The ranking method is theoretically sound and robust.

It performs well in practical applications.

The approach offers a viable alternative to traditional composite indices.

Abstract

A methodology is presented to rank universities on the basis of the lists of programmes the students applied for. We exploit a crucial feature of the centralised assignment system to higher education in Hungary: a student is admitted to the first programme where the score limit is achieved. This makes it possible to derive a partial preference order of each applicant. Our approach integrates the information from all students participating in the system, is free of multicollinearity among the indicators, and contains few ad hoc parameters. The procedure is implemented to rank faculties in the Hungarian higher education between 2001 and 2016. We demonstrate that the ranking given by the least squares method has favourable theoretical properties, is robust with respect to the aggregation of preferences, and performs well in practice. The suggested ranking is worth considering as a reasonable alternative to the standard composite indices.