Purpose: The purpose of this study is to examine effects of classroom-level and student-level factors on student engagement in the context of a higher education system vertically differentiated into research, applied and comprehensive university types. Design/methodology/approach: The study used a cross-sectional multilevel design to explain student engagement based on class and student variables. Specifically, the study collects data from 656 students and 61 randomly selected teachers at both levels and uses multilevel modeling to explain relationship patterns. Findings: The results show that institutions vary significantly in student engagement scores. In addition, while a significant variation is found at the student and classroom level, the effects of academic achievement, instructional quality, teaching experience and teacher qualifications on student engagement vary across classrooms in institutions. However, the interaction effect of classroom and student-level variables on student engagement remains non-significant. Originality/value: The main contribution of this lies in the explanation of student engagement using classroom and student level factors in a vertically differentiated higher education system using multilevel modeling. Student engagement varied in classrooms research universities applied and comprehensive universities.