[pt] Com base nos estudos já realizados, onde se comprova a influência de fatores escolares no desempenho dos alunos, o estudo aborda a política de composição de turmas, estabelecida pelas escolas para os alunos que apresentam defasagem na idade adequada à série cursada. A pesquisa explora os resultados dos testes e instrumentos contextuais do Programa de Avaliação da Educação Básica (PROEB)do Sistema de Avaliação da Educação Pública de Minas Gerais (SIMAVE), realizados em 2003 e mostra, por meio de análises de modelos estatísticos multiníveis, a associação entre homogeneidade/heterogeneidade das salas de aula, em relação à defasagem idade-série, e o desempenho dos alunos. / [en] Based on previous studies which have proved the influence of school characteristics on pupils´ performance, this study examines the current policy determining classroom composition established by schools for those pupils who are older than their peer group. The study explores test results and contextualized questionnaires used by the Basic Education Evaluation Program an organ of the State Public Education Evaluation System of Minas Gerais, carried out in 2003, and which demonstrates, by means of analysis of statistical hierarchical linear models, the relationship between classroom homogeneity/heterogeneity in relation to advanced age for a specific school year and pupils´ performance.

Students’ perceptions of normative support for positive intergroup relations from teachers and school staff have been linked to a number of positive intergroup outcomes (Green, Adams, & Turner, 1988; Jugert, Noack, & Rutland, 2011). Additional studies testing the effects of ethnic proportions in classrooms show evidence for positive and negative intergroup outcomes between ethnic majority and ethnic minority students (e.g., Durkin et al., 2011; Vervoort, Scholte, & Scheepers, 2011). Still, research has yet to test simultaneously the effects of ethnic proportions in a classroom in conjunction with students’ subjective perceptions of normative support for positive intergroup relations. With a sample of Latino and White students from 44 classrooms in two public middle schools the current research tests (1) how two sets of perceived school norms (promoting positive intergroup relations and promoting fairness) predict levels of comfort with outgroup members and greater interest in cross-ethnic friendship, and (2) whether proportions of outgroup members in the classroom moderates the relationships between each set of perceived norms and the outcomes.

The present study uses secondary data from the Positive, Engaged, Achieving Kids (PEAK) Project which was a large-scale effectiveness trial of the Social Skills Improvement System, SEL Edition Classwide Intervention Program (SSIS SEL CIP; Elliot & Gresham, 2017 [PI: DiPerna]). The purpose of the present study was to examine the effectiveness of this intervention at further developing social-emotional competencies for a sample of linguistically-diverse students. Furthermore, this study aimed to understand the impact that various classroom contextual factors, such as teacher-student interactions and classroom composition, had on social-emotional outcomes, as well as the potential moderating role they served between the intervention condition and social-emotional outcomes. This research was guided by the following questions: (1) To what extent does a classwide SEL program, the Social Skills Improvement System, SEL Edition, Classwide Intervention Program (SSIS SEL CIP), improve social-emotional outcomes for a sample of linguistically-diverse students? (2) To what extent do teacher and classroom contextual factors (i.e., teacher-student interactions, teacher characteristics, and classroom characteristics) predict social-emotional outcomes for linguistically-diverse students? (3) Do contextual characteristics moderate the relation between SEL programming and student outcomes, such that the effects of the intervention are magnified for linguistically-diverse students when other contextual factors are also present in the classroom environment (e.g., quality teacher-student interactions)? Hierarchical linear modeling (HLM) revealed that receipt of the SSIS SEL CIP did not serve as a statistically significant predictor of social-emotional outcomes. Teachers’ emotional support positively predicted social-emotional outcomes, while classroom organization served as a negative predictor. Teacher-student interactions (i.e., emotional support, classroom organization, and instructional support) did not have a moderating effect on outcomes, but classroom composition variables did have statistically significant moderating effects such that students who received the intervention and were in classrooms with a higher percentage of racial minority or EL students had higher social-emotional outcomes. Limitations of this study which offer avenues for future research are discussed, along with implications of this study’s findings. / Psychological Studies in Education

In der Dissertation wurden Effekte der zuwanderungsbezogenen Klassenzusammensetzung auf verschiedene Leistungsmaße und psychosoziale Schülermerkmale im Rahmen von drei Zeitschriftenbeiträgen untersucht. In Teilstudie eins wurden Effekte der sozialen und zuwanderungsbezogenen Zusammensetzung auf das Leseverständnis im Verlauf der neunten Klasse und deren Vermittlung durch Merkmale der Unterrichtsqualität analysiert. Der Klassenanteil von Heranwachsenden mit Zuwanderungshintergrund war unter Kontrolle des mittleren sozioökonomischen Status nicht mit der Leseleistung assoziiert. Es zeigten sich Hinweise einer partiellen Vermittlung des Effekts sozialer Zusammensetzung durch das Ausmaß herausfordernden Sprachunterrichts. Teilstudie zwei beschäftigte sich mit Effekten der Zusammensetzung auf die Lernmotivation im Verlauf der neunten Klasse und deren Vermittlung durch das Ausmaß schülerorientierten Unterrichtsklimas. Heranwachsende mit Zuwanderungshintergrund berichteten höhere Motivation, wenn sie Klassen mit einem höheren Anteil von Heranwachsenden mit Zuwanderungshintergrund besuchten. Es zeigten sich keine vermittelten Effekte durch Merkmale des Unterrichtsklimas. In Teilstudie drei wurden der Anteil von Kindern mit Zuwanderungshintergrund und Maße zuwanderungsbezogener Heterogenität in ihrer Assoziation mit Mathematik- und Leseleistung sowie dem Verbundenheitsgefühl mit Peers von Kindern der vierten Klasse verglichen. Es bestanden stärkere negative Effekte des Klassenanteils im Vergleich zu Heterogenitätsmaßen. Bei simultaner Berücksichtigung von Klassenanteil und Heterogenitätsmaßen waren Heterogenitätseffekte größtenteils nicht mehr signifikant. Es zeigten sich jedoch leicht positive Heterogenitätseffekte auf die Mathematikleistung. Es bestanden keine Effektunterschiede zwischen Heterogenitätsmaßen. Die Befunde verweisen auf die Bedeutung der Schulklasse für die akademische Entwicklung sowie für die Erklärung zuwanderungsbezogener Disparitäten im Bildungssystem. / The thesis at hand explored effects of the language-related and ethnic makeup of classrooms on several achievement outcomes as well as psychosocial outcomes in three research articles. The first article investigated effects of socioeconomic and language minority classroom composition on ninth graders’ reading achievement across two measurement points and explored the mediating role of instructional quality. The proportion of minority students had no effect on achievement after controlling for the socioeconomic composition. The results suggest that the effect of the socioeconomic composition on achievement may be mediated partially by the teacher’s focus on language during instruction. The second article focused on effects of the proportion of language minority students on the development of ninth graders’ motivation and the mediating role of a student-oriented climate. The analyses revealed that minority students were increasingly motivated in classrooms with a higher proportion of minority students. The compositional effect was not mediated by the instructional climate. The third article explored various measures of the ethnic makeup of classrooms and their relationship with student outcomes. The proportion of minority students and measures of diversity were compared in their prediction of fourth graders’ achievement in math and reading as well as the feeling of belonging with one’s peers. The proportion of minority students showed stronger negative effects on student outcomes than the diversity measures. Including diversity measures and the proportion of minority students, the diversity effects mostly lost their significance. However, students showed slightly higher levels of math achievement in more diverse classrooms. The various measures of diversity led to similar results. The findings add to the evidence indicating that the classroom learning environment affects students’ academic development and contributes to ethnic disparities in the education system.

Daten der amtlichen Statistik deuten auf eine Überrepräsentation von Kindern mit ausländischer Staatsbürgerschaft an Förderschulen hin. Dies wird in der Literatur häufig als Hinweis auf eine Benachteiligung von Kindern mit Zuwanderungshintergrund (ZHG) gedeutet. Hier anknüpfend geht die vorliegende Arbeit in drei Teilstudien der Frage nach, inwiefern die Diagnose eines sonderpädagogischen Förderbedarfs im Schwerpunkt Lernen (SPF-L) und die besuchte Schulart mit dem ZHG eines Kindes zusammenhängen. Teilstudie I zeigt anhand logistischer Regressionsanalysen, dass Kinder aus zugewanderten Familien unter Berücksichtigung der sozialen Herkunft, der kognitiven Grundfähigkeiten und der schulischen Kompetenzen seltener einen diagnostizierten SPF-L aufweisen und seltener eine Förderschule besuchen als Kinder ohne ZHG. Vertiefend werden zwei Erklärungen für diese Befunde näher betrachtet: Eine Erklärung für die geringere Diagnosehäufigkeit eines SPF-L liegt in der Klassenzusammensetzung. Diese bildet den Referenzrahmen für die Leistungsbeurteilung und könnte somit auch bei der Einleitung von Diagnoseverfahren bedeutsam sein. Teilstudie II untersucht daher, inwiefern die Diagnose eines SPF-L mit der zuwanderungsbezogenen, der sozialen und der leistungsbezogenen Klassenzusammensetzung assoziiert ist. Logistische Mehrebenen-Regressionsanalysen weisen auf entsprechende Zusammenhänge hin. Zur Erklärung des selteneren Förderschulbesuchs von Kindern mit ZHG könnten Schulwahlmotive von Eltern beitragen. Teilstudie III zeigt daher anhand von Strukturgleichungsanalysen, dass sich Schulwahlmotive nach Schulart und ZHG unterscheiden und dass diese Motive den Zusammenhang zwischen einem ZHG und der besuchten Schulart vollständig mediieren. Die Dissertation leistet einen Beitrag zur empirischen Fundierung der Diskussion über Objektivität bei der Einleitung von Diagnoseverfahren und die Relevanz von Hintergrundmerkmalen für die Diagnose eines SPF-L und Schulartentscheidungen. / Data from official statistics indicate that children with migrant background are overrepresented in special schools. In previous studies this finding has often been interpreted as an indication of ethnic discrimination. The present thesis investigates in three studies whether the migrant background of a child is associated with special educational needs in the area of learning (SEN-L) and with the type of school attended. The results of logistic regression analyses in Study I indicate that controlling for differences in social background, cognitive abilities and academic achievement, children with migrant background are less likely to have diagnosed SEN-L and to attend a special school than non-migrant children. Two explanations for these findings are examined in more detail: One possible explanation for the lower frequency of SEN-L diagnosis could be found in the effects of classroom composition which forms a frame of reference for achievement assessment. This may also be relevant for the initiation of diagnostic procedures. Study II investigates whether a SEN-L is associated with the migrant, socioeconomic, and academic classroom composition. Multilevel logistic regression models confirmed the relationship between classroom composition and SEN-L. One possible explanation for the finding that children with migrant background are less likely to attend a special school in the area of learning than comparable children without migrant background could lie in parents' motives for school choice. Study III therefore uses structural equation analysis to show that the parent’s motives for school choice differ according to school type and migrant background and also mediate the relationship between migrant background and school type. The thesis contributes to the empirical foundation of the discussion on objectivity in the initiation of diagnostic procedures by examining the relevance of background characteristics for the diagnosis of SEN-L and decisions on the school type.