The availability of education at a universal scale is paramount when it comes to secure equality of opportunities in a society. In as much as this is desirable, the success at university of students with a low socio-economic status (SES) should be pursued.
Affirmative-action policies will be more effective if an increased presence of these students in higher education does not impact average grades. Determining whether this is the case seems to be a well-motivated research goal.
The Australian government has declared its ambition of reaching a minimum of 20% enrollment of low SES students in the country’s universities. In the heat of this public interest, Craft (2018) sets herself to examine how SES relates to student success within an Australian rural campus. Her choice is justified based on its larger presence of students with limited economic possibilities. The study relies on a cohort of 309 first-year students with data on their age, gender, mother-tongue, parents’ educational background, socio-economic status and previous school achievements.
After going out of her way to discuss the relationship between university entry test scores and first-year university performance, the author moves into the analysis of the evolution of this performance in comparison with the above demographic variables. Several interesting conclusions are drawn. First, older students seem to obtain better grades than their younger counterparts. Second, female pupils also do better than males. As to the main point of this paper, the analysis seems to indicate that SES has no impact on grades. Unfortunately, the robustness of this welcome conclusion is dubious.
Comment: The above results can be questioned on several grounds. It seems strange that the author merely bases her empirical work on an examination of correlations. Regression analysis is a low-hanging fruit which should have been executed. Isolated correlations may be misleading due to the potential impact of other variables. For example, the group of SES students may have a larger presence of females or older students. This would bias the correlation upwards. Also, grades are truncated variables in as much as they can only take values between 1 and 7. Room for improvement or worsening decreases as we reach the upper or lower tails of the distribution of grades, something that biases a correlation downwards.
The presence of SES students in Australian universities has increased as a result of the government’s declared target. This increase has unlikely been uniform since those universities specially concern with academic excellence may have (possibly unfoundedly) resisted the recommendations of the government. Future research could exploit these differences in order to provide more solid evidence that could answer the main question postulated by the author.
One may also question the value of the question itself. What if further research shows that SES students seem to perform worse? This could encourage practices of statistical discrimination within certain universities. If the goal is to guarantee equality of opportunities, research informing efforts to give special support to SES university students may be a more desirable focus of attention.
Text: Iñaki Rodríguez Longarela
