Abstract
Over the past 2 decades, a large number of countries have been engaged in the decentralization of decision-making to schools. Although the motives and incentives for school autonomy are often diverse, it is commonly believed that decentralization will enhance the quality of schooling. Based on a secondary analysis of data from OECD's Programme for International Student Assessment (PISA 2000), this study aims at testing to what degree this assumption is supported by empirical data. For that purpose, 4 domains of autonomy and internal decentralization are taken into account: personnel management, financial resources, student policies, and curriculum. The results indicate that more autonomy in personnel management is related to higher levels of reading literacy. This relationship, however, disappears when differences between schools' student composition are taken into account. No significant associations were found for the other 3 domains.
Notes
1. The analyses are based on five plausible values for each student's reading literacy. These plausible values are random numbers that are drawn from the distribution of test scores that could reasonably be assigned to each individual student, based on their Item Response Theory (IRT) proficiency score. The variance between test scores refers to the variation in plausible values for each of the students, and can be regarded as variance in students' test scores due to measurement error. For more information on the construction of these plausible values, the reader is referred to the PISA 2000 Technical Report (OECD, Citation2002c).