Can 'at risk students' be identified based on demographics?

The success of students in their first year of university can be quite variable, and it is often difficult to anticipate which students require additional support. These students are classified as ‘at risk’ students. The ‘at risk’ students can be identified by monitoring the students' academic performance within early semester assessment items. Students with poor performance in these assessment items can be identified, and faculty and student advisors can better provide these students with support mechanisms in order to minimise the likelihood of student performance issues. PURPOSE The purpose for this study is to determine if other indicators are available to further identify these ‘at risk’ students. For early identification, the indicators present before students start their university degree will be investigated. The beginning step in this study is to determine if ‘at risk’ students can be identified based on the student demographic information, that the student provides when enrolling at the university. The chosen demographic subgroups of this study were: age, gender, residency, entry ranking, and chosen engineering discipline. DESIGN/METHOD The investigation determined if student performance issues could be identified from these available demographics. The method adopted was to perform grade statistical analysis on demographic subgroups, from a single population of all first year engineering students. That is, the study consists of tracking the grades within the eight common first year subjects, that all first year engineering students undertake, within all engineering disciplines. Statistical significance tests were carried out for each tested demographic, to determine if poor grades are attributable to certain demographic groups. RESULTS The analysis revealed that neither gender nor residence status of the students were significant in detecting ‘at risk’ students. The analysis did conclude that both student age and student entry ranking was significant in detecting ‘at risk’ status. In this case, students who were younger and students who had poorer entry ranking, were more likely to be ‘at risk’. The analysis also revealed that students who had not pre-selected an engineering discipline, were more likely to be ‘at risk’. Finally, there was no appreciable performance issues between students of different disciplines. CONCLUSIONS It was found that for some of the particular demographics, significant indicators were observed, whereas no significant indicators were observed for others. Based on the study’s outcome, it is possible to identify some groups of students, before the start of their studies, who will be most likely be in need of additional support in order to succeed in their first year of university.