A quantitative investigation into student progression in STEM subjects at a post-92 UK university

Previous research (e.g. ECU (2008)) has indicated an “attainment gap” between certain groups of students in Higher Education. Even when factors such as socio-economic background and pre-University attainment levels have been controlled, Black, Asian and other Minority Ethnic (BAME) students tend to achieve lower grades than their White peers. In some disciplines (notably some sciences), there is also a marked gap between genders. A previous project at Kingston (2015) indicated that STEM students’ performance at one level of their HE studies correlated significantly and positively with their performance at later levels. However, the regression coefficients (notably the slope) relating their grades at different levels varied considerably between subject disciplines, and subject-specific simple regression only explained a modest proportion of inter-person variation. However, that study did not take any account of students’ individual attributes, e.g. age, gender, ethnicity, socio-economic background. In this talk, we describe our recent work extending the previous project by taking such additional factors into account, to investigate how students from different backgrounds and of different genders and ethnicities perform at different levels of their degree studies. Use of multivariate (multiple linear regression and logistic regression) and multi-level regression models have yielded “all else being equal” results, indicating that BAME students typically achieve marks around 3% lower than those of their white peers, and older students tend to perform less well than younger ones. We are also investigating the influence of students’ socio-economic backgrounds and entry qualifications on their performance at University, and the influence of all these factors on student progression and attainment of “good” (First or Upper Second Class Honours) final degrees. We also had available more “fine grained” information on the profiles of first year mathematics degree students, including their attendance rates at classes, genders, more detail ethnicity information and socio-economic backgrounds, so we were able to perform a more detailed analysis of these students in relation to their attainment. In the first year of their higher education studies, female students of mathematics actually performed slightly better on average than their male peers, and BAME students slightly better than while students, although neither of those differences was statistically significant. Noting that BAME students are more likely to continue living at home during their studies, and thus possibly have long commutes to University, we hypothesised that this could negatively affect their attendance at classes, and hence possibly have a detrimental effect on their academic performance. A student’s average first year module mark proved to be highly significantly correlated with their attendance rate (R = 0.486, df = 84, p < 0.001), although there was substantial variation between students with similar attendance rates. These results could be of considerable value for monitoring and addressing the BAME and gender attainment gaps at Kingston University and elsewhere, and to identifying and rectifying the origins of the problem.