Barrington Lecture 2006/07: Association Rule Analysis of CAO Data

What the paper says

This lecture is delivered under the auspices of the Barrington Trust (founded by the bequest of John Barrington, Esq.) with the collaboration of the Journal of the Statistical and Social Inquiry Society of Ireland. 1. INTRODUCTION 1.1 Background On Monday 21st August 2006 the Irish Times carried an article entitled 'Points race' may be over as CAO requirements tumble. The author, Sean Flynn, explained that the points race may be coming to an end; a claim supported in the article by John McGinnity, deputy registrar at NUI Maynooth (NUIM), who was quoted as stating that year has seen a rebalancing between the supply and demand for places. The term 'points race' is used to describe the struggle to attain points; the metric by which school leavers have come to judge their Leaving Certificates. In past generations the vital statistic in assessing the strength of one's Leaving Certificate was the number of passes and honours attained. Today however, the assessment is usually given in cold hard points, so much so that the attainment of points has become, for many, an end as well as a means. Points are awarded by the Central Applications Office (CAO) on the basis of grades attained at the Leaving Certificate examination. The intended purpose of these points is to allow colleges to decide which students should get into which courses. In this sense, it is understandable that students should want to get as many points as possible; the more points attained, the less likely a student is to be pipped at the finishing line and lose out on their preferred college course. However, it would be a disturbing prospect if students were choosing their courses based on points value rather than on topic; if the status of taking a college course with high points outweighed the wisdom of taking one that the student might enjoy or even excel at. This would be a points race in a different sense: points for points-sake. Considering that the points requirement for a course is dictated by the demand for entry to the course versus the amount of places available, it is not necessarily true that courses requiring higher points are better than those requiring lower points. Furthermore, in a situation where this type of points race existed, our prospects of succeeding as a knowledge-based society in the future would have to viewed as questionable. 1.2 Literature Tuohy (1998)(1) completed the first research paper for the Commission on the Points System, which was established in 1997 by then Minister for Education and Science, Micheal Martin. Amongst his conclusions, the author found that gender may have been an issue in course choice. Moreover, he maintained that the answer to the question over whether or not the points system directly affected students' choices was unclear. Lynch et al. (1999) completed the Points Commission's fourth research paper. Amongst their conclusions, the authors stated that while there is a relationship between LCGPA [Leaving Certificate grade point average] and performance, students with identical Leaving Certificate grades display very different performance outcomes in higher education. The final report of the commission, including recommendations (Hyland, 1999)(2) found that the Leaving Certificate was a valid method of predicting higher education performance, pointing to Lynch et al. (1999) for substantiation. In fact, the report largely vindicated the Leaving Certificate together with the CAO system as a college entrance system. The report highlighted application for medicine courses as an exceptional case and recommended further investigation by the relevant state and healthcare bodies. Ideas around allocation of bonus points for subjects especially relevant to third-level course of interest, for an applicant's preferred courses, or for consistency of choice in completing the CAO form were considered but none were thought favourable. …