Predicting students' academic performance based on school and socio-demographic characteristics

Abstract

Students' trajectories into university are often uniquely dependent on school qualifications though these alone are limited as predictors of academic potential. This study endorses this, examining associations between school grades, school type, school performance, socio-economic deprivation, neighbourhood participation, sex and academic achievement at a British university. Consistent with past research, large entry-level differences between students are generally narrowed by final year at university. Students from the most deprived areas performed less well than more affluent students. Asian and black students performed less well than white students. Female students performed better than their male counterparts. Contrasting with past research, though school performance was positively associated with entry grades, students from lowperforming schools were more likely to achieve the highest degree classifications. Additionally, independent school students performed less well than comprehensive school students at final year despite entering with higher grades. These variations exemplify how patterns observed nationally may differ between universities.Keywords: education; attainment; contextual background; inequality Despite a dramatic increase in higher education (HE) participation in England over the last half century, the under-representation of students from socio-economically disadvantaged backgrounds remains a glaring reality (Blanden and Machin 2004;Breen and Jonsson 2005;Croxford and Raffe 2013;Haveman and Smeeding 2006;Singleton 2010a). These students are known as Widening Participation (WP) students, who along with students with disabilities and some ethnic minority groups are currently under-represented in HE (Gorard 2008;Mason and Sparkes 2002). Differences in HE participation are largely attributed to the poorer school-level academic qualifications obtained by a large proportion of students within low socio-economic status (SES) classifications and are associated with educational disadvantage (Chowdry et al. 2013;Steele, Vignoles, and Jenkins 2007; Sutton Trust 2005). Further, research comparing the academic performance of students from different school types and

Using data mining techniques to predict students at risk of poor performance

“…In the later subjects, Home students have higher levels of attainment than OS students. Additionally, our findings were consistent with other studies [23], [25] in terms of gender: female students are more likely to graduate with GH degrees than male students, although they are minorities in some disciplines such as science subjects compared to art subjects. For instance, our first dataset which relates to a science subject has 17% female students, compared to 83% male students; our validation dataset has 39% females and 61% males.…”

“…For instance, our first dataset which relates to a science subject has 17% female students, compared to 83% male students; our validation dataset has 39% females and 61% males. Some studies [23] have found that students who come from areas with the lowest levels of participation in HE, and those who come from less affluent areas, are more likely to have lower attainment. In contrast, some other studies in [23] have found no significant difference.…”

“…[23], [24], [25], and [26]). Previous studies have found that Home students are associated with higher attainment than OS students ( [25], [23] and [24]).…”

“…[23], [24], [25], and [26]). Previous studies have found that Home students are associated with higher attainment than OS students ( [25], [23] and [24]). In contrast, some studies [25], [27] found that there were no significant differences in the class of degree obtained by OS students compared to Home students.…”

Systematic mapping review on student’s performance analysis using big data predictive model

“…Limited papers fall under extended [19-20, 27, 41, 43, 55, 58-60, 63-65] and holistic [14,[22][23][24]39] categories. In [25] included distance of schools from a district office to predict school's performance and accreditation within the vicinity of the district office.…”