A fair selection for university?

Christina Wikström of Umeå University finds that better tests and more trained teachers are the key to more successful university admissions.

Application formWhen there are more eligible applicants for higher education than there are places, selection must take place. But just how this selection should be carried out has been debated for a long time. Beliefs about fairness, efficiency and consequences are often important for this decision, as is the weight of tradition.

A very common approach is the meritocratic one, by which applicants are ranked by their former merits, measured by test scores or previous grades. Sometimes there are influences from a utilitarian approach, where a ‘greater good’ involving societal consequences is considered. Or universities may aim for a selection that makes the student body more representative of society as a whole, or in some other way more balanced. However, this approach is seldom applied. It may be regarded as unfair to give certain groups advantages, even if they are under-represented in higher education. To avoid this problem, a lottery – where all eligible applicants have the same chance of getting admitted – can be applied. This is carried out in the Netherlands and in a few other places. Although it may be an easy way of selecting a certain number of students, while also solving some of the problems of an unbalanced selection, this approach is not common as it violates the common belief in rewarding the best applicants. Instead, most observers regard a meritocratic selection as both fair and efficient; the applicant with the highest merits ‘wins’ and the universities receive high-performing students.

Different instruments are used for this purpose. In some systems, grade averages and other types of school information are used. In other systems there are admissions tests. Sometimes both measures are used, either in combination (e.g. Israel and the US) or as two separate instruments, as in Sweden.

A good selection instrument is expected to take three main aspects into consideration: fairness, efficiency and consequences. However, most research on selection instruments focuses on how well the instrument can predict academic success, and on group differences (Wolming & Wikström, 2010). The results show that it is very difficult to find an instrument that meets and considers all three requirements, while also being a good predictor of academic success. The international research seems coherent; although the content and format of the instruments will differ between systems, predictive strength is moderate for both admissions tests and school grades, but is slightly higher for grades.

One hypothesis is that grades also reflect personality aspects such as motivation and industriousness, while the test can only measure what a student knows and can do. However, personality traits are not accepted as valid measures in the competition for university admission. Taking them into account may be perceived as unfair and lacking in transparency, unless it is for a special reason. Examples might be selection for vocational programmes educating future doctors, teachers, social workers or police officers. In these subjects, reviews or personal inventories are often used when selecting among the top applicants from the ranking produced by traditional selection instruments. 

Grades and tests often rank students similarly, and are sensitive to group differences, but they do exhibit some variation. Female students often perform higher in terms of grades, while male students perform higher on tests. Students from a higher socio-economic and non-immigrant background often perform higher than other students on both instruments. The gaps are often larger with admissions tests than with school grades. In addition, the academic performance of students with a non-traditional academic background is sometimes over-predicted (Zwick, 2003). It is not clear why this happens. Studies in higher education may be more favourable to certain groups, and non-traditional students tend to have higher drop-out rates.

Since grades are closer to the criterion – i.e. they predict future grades or other outcomes in higher education better – they are often regarded as more relevant selection instruments. Admissions tests are often perceived as fair since they are scored objectively, although they are also criticised for measuring a narrower domain than grades and for being unfair to students who underperform on tests because of anxiety or other causes.

However, grades have other validity problems too. They are often the results of assessments that vary between teachers, schools, and over time. Also, in many systems a grade average (GPA), or equivalent, can consist of grades from different courses, and courses on different levels, meaning that two GPAs are not necessarily comparable. This is not fair to the applicants. An inflated grade system will lead to problems for students who do not continue to higher education directly after upper secondary school because they were unsuccessful in their first application.

This may also apply to the situation in the UK, where A-levels are used to ensure eligibility but also to rank-order students. Students may apply with different combinations of A-levels, and with A-levels from different schools and years. Given the possible effects of suspected grade inflation, it is very likely that students will apply with different credentials, which makes the requirements mentioned above difficult to meet.

In countries where tests and grades are used, separately or in combination, there are ambitions to make the grades more objective – perhaps with better criteria, teachers educated in assessment, more or better tests, or psychometric methods for grade calibration – and to make admissions tests more aligned with school curricula. This will most likely improve reliability, and hence fairness. But whether validity and the predictive strength of the instruments will improve is yet to be seen.

Christina Wikström is senior lecturer in applied educational science at Umeå University, Sweden.

References: 
  1. Wolming, S., & Wikström, C. (2010). The concept of validity in theory and practice. Assessment in Education: Principles, Policy & Practice, 17(2), 117-132.
  2. Zwick, R. (2002). Fair Game? The use of standardized admissions tests in higher education. Routledge.

 

Share this page