Systematic differences across evaluation schemes and educational choice #

• Disadvantaged groups get systematically lower teacher scores for equal exam scores.
• This may affect education choice if external scores are unknown when choices are made.
• Eliminating disparities closes 13% of low-SES pupils’ high-school enrolment gap.

Using large-scale register data from Denmark in a difference-in-differences framework, I analyse whether systematic disparities between internal teacher scores and external exam scores in the school-leaving certificates are linked to pupil characteristics. Such differences may matter for post-compulsory education choices in school systems where external assessments are not available when these choices are made. I use sibling fixed effects methods to simulate changes in educational choices for disadvantaged groups if they were graded by their teachers as their advantaged peers. The results suggest that probabilities of enrolling in more ambitious tracks would increase across the whole ability distribution and for all disadvantaged groups. In particular, in the middle of the ability distribution the results predict an increase of 5% points in migrant pupils’ probability of enrolling in high-school. The corresponding increase for pupils with low educated parents is 4% points, closing 13% of the high-school enrolment gap to pupils with high educated parents.