Kausale Effekte in der Empirischen Bildungsforschung. Ein Vergleich verschiedener Ansätze zur Schätzung des Effekts des Einschulungsalters.

Autor/inn/en	Lüdtke, Oliver; Robitzsch, Alexander; Köller, Olaf; Winkelmann, Henrik
Titel	Kausale Effekte in der Empirischen Bildungsforschung. Ein Vergleich verschiedener Ansätze zur Schätzung des Effekts des Einschulungsalters.
Quelle	Aus: Bos, Wilfried (Hrsg.); Klieme, Eckhard (Hrsg.); Köller, Olaf (Hrsg.): Schulische Lerngelegenheiten und Kompetenzentwicklung. Festschrift für Jürgen Baumert. Münster u.a.: Waxmann (2010) S. 257-284 PDF als Volltext Link als defekt meldenVerfügbarkeit
Beigaben	grafische Darstellungen
Sprache	deutsch
Dokumenttyp	online; gedruckt; Sammelwerksbeitrag
ISBN	3-8309-2358-9; 978-3-8309-2358-9
Schlagwörter	Bildungsforschung; Empirische Forschung; Empirische Untersuchung; Methode; Alter; Deutschland; Deutschunterricht; Kausalität; Mathematikunterricht; Methode; Bildungsforschung; Datenanalyse; Empirische Forschung; Schulanfang; Schuljahr 03; Schuljahr 04; Schülerleistung; Deutschunterricht; Datenanalyse; Mathematikunterricht; Kausalität; Alter; IGLU-E (Internationale Grundschul-Lese-Untersuchung - Ergänzungsstudie); Deutschland
Abstract	This chapter considers the methodological challenges of assessing causal effects in educational research. Based on the question of the age at which children should start school, the chapter applies different strategies (regression analysis, instrumental variable estimation) to estimate the causal effect of age at school entry on later school achievement. Analyses show how the potential outcome framework can be used to define the causal effect of age at school entry and how endogeneous variation in age at school entry (e.g., the time of entry is also influenced by the child's development) complicates the estimation of the causal effect. Regression analysis based on the pilot data used to assess the validity of the German national educational standards in elementary mathematics reveal a negative effect of age at school entry on mathematics achievement in third and fourth grade. However, analyses using the exogeneous variation of month of birth as an instrumental variable for age of school entry show that starting school at a relatively older age had a positive effect on later mathematics achievement. In the next step, a structural model is introduced that explains interindividual differences in school achievement by the decision to start school later, month of birth, and maturity before and after entering school. Sensitivity analyses conducted on the basis of this structural model show that both estimators of the causal effect of school entry are highly sensitive to violations of their assumptions. Implications for using nonexperimental data to assess causal effects in educational research are discussed. (DIPF/Orig.).
Erfasst von	DIPF \| Leibniz-Institut für Bildungsforschung und Bildungsinformation, Frankfurt am Main
Update	2010/4