Comparing teacher and external assessments: Are boys, immigrants, and poorer students undergraded?

Abstract

In this paper, we study the presence of systematic differences between teacher non-blind assessments and external blindly graded standardized tests as a measure of grading misalignment. Using a large administrative database covering two student cohorts (N 1⁄4 31, 183 pupils) from publicly-funded schools in the Basque Country (Spain), we explore the grading gaps found between these two type of assess- ments for several student characteristics using fixed effects modeling. We find that, after controlling for standardized achievement, systematic teachers’ under-assessment exists for student groups that, on average, lag behind in school: boys, children from an immigrant background, and low SES students. The observed data patterns withstand several robustness checks, including the use of instrumental variables approach (IV) and other alternative regression specifications.