Measuring educational inequality of opportunity: pupil’s effort matters

•Effort, broadly defined, contributes between 31% and 40% of predicted variation in student performance in rural Bangladesh.•Effort indicators are the main source of within-school variation in mathematics and English performance.•Social background and parental effort are a key source of between-schools variation in performance.•Accounting for the correlation between effort and circumstances beyond students’ control does not matter much for the analysis of inequality of opportunity in the context of rural Bangladeshi schools.•In the context of rural Bangladesh, it is not sufficient to erase the influence of family and social background to level the educational playing field. The distinction between effort and other factors, such as family background, matters for correcting policies and normative reasons when we appeal to inequality of opportunity. We take advantage of a purposefully designed survey on secondary schools in rural Bangladesh to offer a comprehensive view of the importance of overall effort when measuring inequalities of opportunity in education. The analysis comprises decomposition exercises of the predicted variance of student performance in mathematics and English by source (effort, circumstances, etc.) and subgroup (within- and between-schools) based on parametric estimates of educational production functions. Pupils’ effort, preferences, and talents contribute between 31% and 40% of the total predicted variances in performance scores. The contribution of overall effort falls by 10% when the correlation between effort and circumstances is taken into account. These findings are robust to the choice of estimation strategy (i.e. combined within- and between-schools variation models versus multilevel random-effect models). All in all, these results advocate that social determinism in education can be mitigated by individual effort at school.