Are assessments in higher education fair for all? Investigating the association between assessment load and type and students’ academic performance: a longitudinal large-scale data set analysis

The persistent awarding gap between advantaged groups and marginalised groups remains a critical concern for higher education. Moving beyond individual-deficit explanations, this study investigates how assessment design, including types and loads, functions as structural impediments for success. Drawing on a large-scale longitudinal data set (2017/18–2024/25) involving 52,420 students represented across 3,606 modules (over 700,000 student-module observations) at a UK university, we employ linear mixed-effects modelling to examine the relationship between assessment design and student performance. Our results reveal that traditional, exam-based assessments are positively associated with academic performance among White students, but negatively associated with the performance of Black students and those from socio-economically disadvantaged backgrounds. Conversely, practice-based and relational assessments appear to function as a pedagogical leveller, particularly for Black students. Furthermore, this study finds a positive association between assessment load and module marks, a result which is significantly more pronounced for socio-economically disadvantaged students. These findings have crucial implications for assessment policies and educational practices in higher education.