Background

The prevention of early leaving from education is a central objective of education policy within the European Union. Student dropout can be conceptualized as the endpoint of a gradual process of disengagement from school, arising from cognitive, affective, and behavioral antecedents. Prior research situates disengagement within processes of declining motivation, negative affect, and truancy, which collectively undermine students’ academic trajectories. As risk factors such as gender or migration history are largely immutable, researchers should aim to identify early indicators of dropout that are more amenable to intervention. In this study, we examined the social context of the classroom and its potentially protective role against disengagement. Drawing on social identity theory, we hypothesized that a sense of belonging—operationalized as social identification with one’s class—and a positive classroom climate would be associated with lower levels of disengagement at both individual and classroom levels. To account for classroom dynamics and student composition, we employed a multilevel analytical approach.

Method

We drew on two cross-sectional samples from 16 German middle schools with high annual dropout rates (>10%). Sample 1 comprised 255 students in 24 classrooms, and Sample 2 included 287 students across 21 classrooms. Students self-reported affective (daily mood at school), cognitive (amotivation), and behavioural (truancy) disengagement, as well as social identification with the class and classroom climate. Gender, grade level, and migration background were included as individual-level covariates in both samples; economic learning resources (ELR) were additionally controlled for in Sample 2. Multilevel mixed-effects models with random intercepts were estimated to examine associations at both the individual and classroom level.

Results

Across both samples, stronger individual social identification with the class was consistently associated with a more positive daily mood at school, indicating lower affective disengagement. In Sample 2, social identification was significantly negatively associated with amotivation. These findings suggest that a stronger identification with the classroom corresponds with reduced affective and cognitive disengagement among students at risk of dropout. At the classroom level, however, neither aggregated social identification nor classroom climate showed protective associations with disengagement outcomes. Notably, higher grade level and a greater proportion of students with a migration background were associated with increased truancy at the classroom level in Sample 2, indicating that classroom composition may contribute to behavioural disengagement. Classroom climate was not significantly associated with any disengagement dimension in either sample.

Conclusion

Our study highlights the importance of individual students’ social identification with their classroom as a correlate of lower disengagement—particularly affective and cognitive disengagement—in educational contexts characterized by elevated dropout risk. Classroom-level aggregates of identification and perceived classroom climate did not independently buffer against disengagement, suggesting that protective social processes in heterogeneous, high-risk classrooms may operate primarily at the individual level. However, grade level and migration background as classroom composition factors were associated with behavioral disengagement. Interventions targeting the prevention of dropout may therefore benefit from strengthening students’ sense of belonging and individual social integration within classrooms, while also considering demographic and compositional influences at both the individual and classroom level.


Back to Spotlights

Eara