4. Quality education
High-stake exams change teacher grading standards: evidence from a policy reform
Abstract
The introduction of high-stake standardized exams has been shown to be strongly associated with improvements in an average performance in international standardized assessments [Bergbauer et al., 2018] as well as with proficiency in mathematics and reading [Jacob and Lefgren, 2004]. Despite legitimate concerns with the incentives to teach to the exam and the limited set of skills under evaluation, the fact that idiosyncratic grading biases tend to vanish through the application of blind, anonymous evaluation mechanisms have ensured their validity and continuity in multiple education systems (OECD, 2013). By means of a difference-in-difference approach, we exploit a policy reform that introduced high-stakes exams at the end of the 6th grade in Portuguese and Mathematics, in Portugal, for a short period of 4 years. We study how the introduction of these high-stake exams changed teacher grading standards and teacher scores distribution. We find that teachers in courses covered by external high-stakes exams tend to give a significantly higher percentage of failing scores and a significantly lower percentage of higher scores, an effect which shows to be stronger for Mathematics and not homogeneous for male and female students.
Access this paper here