HIGH-STAKES EXAMS CHANGE TEACHER GRADING STANDARDS: EVIDENCE FROM A POLICY REFORM
Autor: Pedro Freitas
Resumo: The existence of external assessment mechanisms remains a controversial issue in education policy. In this work we study how student assessment performed by class teachers is affected by the strengthening of external accountability mechanisms, through the introduction of high-stakes exams. By means of a difference-indifferences
approach, we exploit a policy reform to measure how the introduction of high-stakes exams in Portuguese and Mathematics at the end of the 6th grade
changed teacher grading standards. It is shown that the introduction of high-stakes exams leads to a change in teacher scores distribution, namely a significantly higher share of failing scores and a lower percentage of higher scores. Consequently, in the
short-run, we estimate that the introduction of high-stakes exams increased grade repetition by 3 pp. This effect is stronger for Mathematics and for schools with a lower concentration of college graduated mothers. We additionally conclude on potential positive effects on long-run academic results for students who took the high-stakes exams.
ALL MEASURES LEAD TO ROME: ABILITY PEER EFFECTS USING TEACHER VS EXAM SCORES
Autores: David Marques da Costa, Pedro Freitas, Luís Catela Nunes, Ana Balcão Reis
Resumo: We study whether different measures of students’ achievement produce different estimates of ability peer effects. We compare peer effects of classmates when ability is measured through standardized exam scores in Mathematics and Portuguese and teacher scores for a broader set of subjects and grades. Our dataset refers to the population of 9th grade students in public schools in Portugal. To address identification threats stemming from unobserved sorting of students to peer groups, we resort to school-by-year and teacher-by-school fixed effects. The results point to positive and very similar ability peer effects regardless of the subject and assessment measure.
NOT HAPPY WITH MY SCORE! DETERMINANTS OF EXAM APPEALS IN HIGH SCHOOL
Autores: Pedro Freitas, Luís Catela Nunes, Ana Balcão Reis
Resumo: Using a large administrative database of upper secondary students, we examine how exam appeals and their outcomes are related to students’ characteristics. We conclude that the higher the family’s socioeconomic status, the greater the likelihood of the student filing an exam appeal. However, the likelihood that the revision will
result in an improvement in the exam score is not determined by the students’ original exam scores. Our findings show that, while exam grading is blinded, the presence of barriers to exam appeals may create biases favouring students from higher socioeconomic backgrounds.
WHAT DRIVES THE EVOLUTION IN PISA SCORES? A QUANTILE DECOMPOSITION APPROACH
Autores: Pedro Freitas, Luís Catela Nunes, Ana Balcão Reis
Resumo: In recent years PISA has become one of the most inuential instruments to compare dierent educational systems across time. The structure of the PISA data allows the identication of the evolution of 15-year-old students' academic performance in the long-run. This evolution can be driven by changes in the country's socio-demographic structure or due to the higher ability of educational systems to reach higher results given the students' characteristics. By using data from Portugal, a country with one the sharpest positive evolutions in PISA scores between 2003 and 2018, we test these hypotheses by means of a quantile decomposition approach. We estimate that, despite the signicant changes in the socio-economic background of Portuguese students in recent years, this was not the main driver in the evolution of Portuguese results. Additionally, we conclude that the change in results is higher for high-achieving students and that the evolution is stronger for students in public schools and in non-vocational tracks.
UNDERSTANDING THE PRIVATE-PUBLIC SCHOOL PERFORMANCE GAP IN PISA: EVIDENCE FROM PORTUGAL
Autores: Ricardo Colaço, Pedro Freitas, Luís Catela Nunes, Ana Balcão Reis
Resumo: Education policies often become the subject of scrutiny when the results of international large-scale assessments are published. However, the quality of thoseassessments depends critically on the representativeness of the relevant samples. Namely, for sound comparisons between public and private schools, it is required that the respective samples are representative. We analyse the recently reported
convergence in performance of private and public schools in Portugal. We find that part of the diminishing gap in scores between the two types of schools is actually explained by sampling issues. In addition, when changes in the population structure
are taken into account and score eects are isolated, we find that the convergence in scores is much less impressive than reported. For instance, in Science, after correcting for sampling issues and removing population composition eects, the reported convergence of 45.6 points between private and public schools from 2015 to 2018, amounts to only 8.7 points. The decomposition and sampling correction methods used in this paper can be easily adapted to other contexts.
CAN VOCATIONAL EDUCATION IMPROVE SCHOOLING AND LABOUR OUTCOMES? EVIDENCE FROM A LARGE EXPANSION
Autores: João R. Ferreira, Pedro S. Martins
Resumo: We evaluate the education and labour impact of vocational education and training (VET). Identification draws on a reform to reduce early school leaving, which involved a large-scale, staggered introduction of VET courses. Drawing on comprehensive student-school matched panel data, we find that VET increased upper secondary graduation rates considerably: our LATE estimates are as large as 50 percentage points. These effects are even stronger for low-achieving students and welfare recipients; and also hold when exploiting the large gender differences of VET, with many courses selected almost only by either boys or girls. Moreover, we find evidence of regional youth employment growth and VET wage premiums following VET expansion.
Leia aqui: Nova SBE Working Paper Series wp658