Modelling nonresponse in educational multi-informant studies. A multilevel approach using bivariate probit models.

Autor/inn/en	Steinhauer, Hans Walter; Aßmann, Christian
Institution	Leibniz-Institut für Bildungsverläufe
Titel	Modelling nonresponse in educational multi-informant studies. A multilevel approach using bivariate probit models.
Quelle	Bamberg: Leibniz-Institut für Bildungsverläufe (2018), 31 S. PDF als Volltext kostenfreie Datei Link als defekt meldenVerfügbarkeit
Reihe	LIfBi Working paper. 74
Beigaben	Literaturangaben
Sprache	englisch
Dokumenttyp	online; Monographie
Schlagwörter	Längsschnittuntersuchung; Telefoninterview; Test; Eltern; Schule; Schüler; Bivarianz; Studie; NEPS (National Educational Panel Study)
Abstract	Large-scale educational surveys open use a multi-informant survey approach. Adapting such an approach, the National Educational Panel Study enriches students' tests and survey data with information obtained within a parent telephone interview. Both, students and parents may refuse to participate. We adapt a bivariate probit model with multilevel structure allowing for clustering at the school level to model the participation process. Using simulated maximum likelihood estimation, the empirical results point at significance of explaining factors like school type and family background. Specification tests highlight the importance of considering correlation as well as clustering structures when modelling participation processes. (Orig.).
Erfasst von	DIPF \| Leibniz-Institut für Bildungsforschung und Bildungsinformation, Frankfurt am Main
Update	2021/2