Research projects

Following recent literature, we consider testing and survey behaviours, in particular, decline in performance during the test, as our main measures of non-cognitive skills. Information on these abilities is obtained from external evaluations of the schools, the PISA 2012 tests. We compare non-cognitive skills of students who attended schools that participated in the PAE with the hypothetical outcomes that these same students would have obtained had they not attended PAE schools. The counterfactual outcomes are inferred using schools that did not participate in the PAE but took the PISA exams. We employ two complementary estimation strategies: the matching procedure weighting method and a difference-indifferences approach.

We find that the PAE has a substantial positive effect on non-cognitive skills: the estimated increase on the ability to sustain test performance is between 0.042 and 0.047 of one standard deviation. In addition, it reduces the probability of falling behind into the bottom part of the ability to sustain test performance distribution by about 2 percentage points. The results are driven by girls and are explained by the positive gender gap in sustained test performance prior to the remedial intervention combined with a more effective participation in it among girls.