A Researcher's Guide to the Swedish Compulsory School Reform. CEE DP 87

Centre for the Economics of Education. London School of Economics and Political Science, Houghton Street, London, WC2A 2AE, UK. Tel: +44-20-7955-7673; Fax: +44-20-7955-7595; e-mail: [email protected]; Web site: http://cee.lse.ac.uk

N

57

Researchers

School Restructuring; Compulsory Education; Outcomes of Education; Educational Attainment; Educational Change; Gender Differences; Achievement Gains; Program Effectiveness; Program Evaluation; Educational Trends; Municipalities; Program Implementation; School Statistics; Check Lists; Predictor Variables; Outcome Measures; Statistical Analysis; Statistical Data; Cohort Analysis; Educational Assessment; Educational Indicators; Foreign Countries; Evaluation Methods

2045-6557

When studying different types of returns to education, educational reforms are commonly used in the economics literature as a source of exogenous variation in education. The Swedish compulsory school reform is one example; the reform extended compulsory education throughout the country, in different municipalities at different points in time. Such variation across cohorts and regions can be used in a differences-in-differences framework, in order to estimate causal effects of education. This paper provides a guide to researchers who consider using the Swedish reform in an empirical analysis: I present a description and background of the reform, provide some baseline results, a reliability analysis of the reform coding, a discussion of whether the reform is a valid instrument, and comment on the interpretation of IV estimates of returns to schooling. The main conclusions are the following: i) a reliability analysis of the reform coding finds a lower bound reliability estimate of 0.66-0.91; ii) the reform indeed raised educational attainment, more so for boys than for girls, and iii) with careful consideration of region-specific trends, the reform can be considered a valid instrument for education. Appended are: (1) Reform coding for register data; and (2) Reliability analysis. (Contains 4 figures, 9 tables and 16 footnotes.) [Part of this research has been financed by the Swedish Council for Working Life and Social Research.]

As Provided

34

2012

ED530136