Übereinstimmung, Variabilität und Reliabilität von Schülerurteilen zur Unterrichtsqualität auf Schulebene: Ergebnisse aus 81 Ländern

Using student ratings to assess instructional quality of schools should fulfill three requirements: (1) an appropriate level of inter-rater agreement within schools, (2) systematic variance of student ratings between schools, (3) an adequate reliability level of aggregated student ratings. Using international PISA-data (2000–2012; 81 countries, over 55,300 schools, over 1.3 million 15-year olds) this study investigated how these requirements were met regarding indicators of instructional quality (classroom management, cognitive activation, individual learning support). We computed the interrater agreement index r WG(J) , as well as the intraclass correlations ICC(1) and ICC(2). Our results showed that (1) student ratings demonstrated a moderate or strong level of agreement for most indicators of instructional quality and (2) instructional quality assessed by students varied systematically between schools. Yet, (3) reliability of aggregated student ratings was not sufficient in many countries. We discuss these results regarding conventions to evaluate agreement, variability, and reliability of student ratings at the school level.