The Effects of Targeted Learning Support: Evidence from a Regression Discontinuity Design

Abstract

Low secondary school completion rates are a persistent cause of concern amongst policymakers in high-income countries. In the most recent data, one in four 25-year-olds in OECD countries have yet to pass upper secondary education. This thesis evaluates the short-term effects of a Norwegian policy that aims to increase completion rates. The evaluated program provides learning support to low-performing students at the end of lower secondary school, seeking to improve their basic skills in reading, writing and numeracy. The explicit target group of the program is the bottom ten percent in the average grade distribution. However, the assignment rule has been interpreted differently, creating institution-specific thresholds that determine the participation offers to the students. I develop an approach to identify these thresholds that may also prove useful for other evaluations of targeted policies where lower level administrative units have implemented rules independently. For a relatively small sample the necessary assumptions for a regression discontinuity design are credible. I find no evidence of effects from the program.