Search

Accountability, autonomy, and choice play a leading role in recent school reforms in many countries. This report provides new evidence on whether students perform better in school systems that have such institutional measures in place. We implement an internationally comparative approach within a rigorous micro-econometric framework that accounts for the influences of a large set of student, family, school, and country characteristics. The student-level data used in the analysis comes from the PISA 2003 international student achievement test that encompasses up to 265,000 students from 37 countries. Our results reveal that different facets of accountability, autonomy, and choice are strongly associated with the level of student achievement across countries. With respect to accountability, students perform better where policies are in place that aim at students (external exit exams), teachers (monitoring of lessons), and schools (assessment-based comparisons). The combined achievement differences amount to more than one and a half PISA grade-level equivalents. Students in schools with hiring autonomy perform better on average, while they perform worse in schools with autonomy in formulating their budget. School autonomy over the budget, salaries, and course contents appears to be more beneficial when external exit exams hold schools accountable for their decisions. Students perform better in countries with more choice and competition as measured by the share of privately managed schools, the share of total school funding from government sources, and the equality of government funding between public and private schools. Cross-country differences in private school operation account for up to two PISA grade-level equivalents. The performance advantage of privately operated schools within countries is stronger where schools face external accountability measures and are autonomous. In urban areas, indicators of choice among public schools are also associated with superior outcomes. Several aspects of accountability, autonomy, and choice are also associated with superior noncognitive outcomes such as student morale and commitment, non-disruptive behaviour, disciplinary climate, and tardiness. We find no evidence that these policies have led schools to focus on raising student achievement at the expense of non-cognitive skills. (Three appendices include: (1) Database and Descriptive Statistics; (2) Econometric Modeling; and (3) Additional Tables. Contains 32 footnotes, 11 figures and 16 tables.)