Several researchers posit the tendency of adults and children to behave positively towards individuals similar to themselves (in-group bias) and to behave hostilely towards those dissimilar to themselves (out-group bias) is not learned, but is instead innate.! Using infant-parent dyads, Mahajan and Wynn (2012) examined this question by asking infants (n = 32), seated in their parents' lap, to choose between two foods, watch a puppet show during which two puppets verbally stated a liking or disliking of these foods, and then choose one of the puppets. They found more infants chose the puppet that "liked" the same food the infants chose. Based on these results, Mahajan and Wynn suggest this tendency is innate. However, methodolgical limitations cling the unintentional effects of parent bias and use of a single choice-trial make this conclusion premature. The purpose of the current study was to replicate and extend the methodology of Mahajan and Wynn by addressing these limitations. Twenty-four. infant-parent dyads were separated into two groups. Affer choosing a food and watching the puppet show, infants in Group 1 (mutiple-baseline across participants design) chose a puppet 3-5 times before their parents were exposed to the parent bias measure and then chose an additional five times; parents in Group 2 (between-subjects comparison group) were exposed to the parent bias measure first, then infants chose a puppet five times. In Group 1, 7 of 12 infants (58%) selected the similar puppet on the first choice trial In Group 2, 2 of 12 infants (17%) selected the similar puppet on the first choice trial. Repeated choice triat showed no influence of parent bias but did show patterns of side stability with 18 of 24 infants making a majority of their puppet selections on the same side.

Several researchers posit the tendency of adults and children to behave positively towards individuals similar to themselves (in-group bias) and to behave hostilely towards those dissimilar to themselves (out-group bias) is not learned, but is instead innate. Using infant-parent dyads, Mahajan and Wynn (2012) examined this question by asking infants ( n = 32), seated in their parents' lap, to choose between two foods, watch a puppet show during which two puppets verbally stated a liking or disliking of these foods, and then choose one of the puppets. They found more infants chose the puppet that “liked” the same food the infants chose. Based on these results, Mahajan and Wynn suggest this tendency is innate. However, methodological limitations including the unintentional effects of parent bias and use of a single choice-trial make this conclusion premature. The purpose of the current study was to replicate and extend the methodology of Mahajan and Wynn by addressing these limitations. Twenty-four infant-parent dyads were separated into two groups. After choosing a food and watching the puppet show, infants in Group 1 (multiple-baseline across participants design) chose a puppet 3-5 times before their parents were exposed to the parent bias measure and then chose an additional five times; parents in Group 2 (between-subjects comparison group) were exposed to the parent bias measure first, then infants chose a puppet five times. In Group 1, 7 of 12 infants (58%) selected the similar puppet on the first choice trial. In Group 2, 2 of 12 infants (17%) selected the similar puppet on the first choice trial. Repeated choice trials showed no influence of parent bias but did show patterns of side stability with 18 of 24 infants making a majority of their puppet selections on the same side.

Die Dissertation untersucht Leistungserwartungen von Lehrkräften. Teilstudie 1 und 2 gehen der Frage nach, welche Merkmale von Schüler*innen Lehrkräfte in ihre Erwartungen an die Leistungen von Lernenden einbeziehen. Teilstudie 3 untersucht, welche Bedeutung inakkurat hohe oder niedrige Leistungserwartungen von Lehrkräften für ihr unterrichtliches Handeln und für die Leistungsentwicklungen der Lernenden haben. Die Analysen basieren auf Daten der Längsschnittstudie „Kompetenzerwerb und Lernvoraussetzungen“, an der im Schuljahr 2013/2014 insgesamt 1065 Schüler*innen aus 64 ersten Grundschulklassen teilgenommen haben. Teilstudie 1 ergab, dass Erwartungen von Lehrkräften an die sprachlichen und mathematischen Leistungen von Erstklässler*innen in Abhängigkeit vom ethnischen Hintergrund, sozialen Hintergrund und Geschlecht der Lernenden variieren. Diese Unterschiede blieben zum Teil auch nach Kontrolle der Ausgangsfähigkeiten und selbsteingeschätzten Motivation der Schüler*innen bestehen und kennzeichnen daher Verzerrungen. Den Ergebnissen von Teilstudie 2 zufolge gehen solche sozialen und geschlechtsbezogenen Verzerrungen zum Teil auf Lehrkrafteinschätzungen der Motivation und des Arbeitsverhaltens zurück. Ethnische Verzerrungen bestanden hingegen unabhängig von diesen Lehrkrafteinschätzungen. Dass inakkurat hohe oder niedrige Leistungserwartungen die Leistungsentwicklung von Lernenden bereits im ersten Grundschuljahr beeinflussen können, zeigten die Ergebnisse der Teilstudie 3. Die zugehörige Videostudie ergab, dass sich inakkurate Leistungserwartungen von Lehrkräften in ihrem Feedbackverhalten niederschlagen. Eine bedeutsame Mediation der Erwartungseffekte durch das Feedback ließ sich aber nicht nachweisen. Die Befunde der Dissertation sind für die pädagogische Praxis sehr relevant. Sie sprechen einerseits für den Vorteil hoher Erwartungen und sensibilisieren andererseits für das Risiko verzerrender Einflüsse von Hintergrundmerkmalen der Schüler*innen. / The dissertation deals with teacher expectations for student achievement. The first two studies investigate student characteristics teachers rely on when forming their achievement expectations. Study 3 investigates the associations of inaccurately high or low teacher expectations with their teaching behavior and students’ subsequent achievement development. The analyses are based on data from the longitudinal research project “competence acquisition and learning preconditions” in which 1065 students from 64 first grade classes participated in the school year 2013/2014. Study 1 revealed that teachers’ expectations for students’ language and mathematics achievement differed depending on students’ ethnic background, social background and gender. These differences partly persisted even after controlling for students’ actual achievement, general cognitive abilities and motivation, and therefore indicate biased expectations. The results of study 2 showed that social bias and gender bias in teacher expectations were partly due to differences in teachers’ perceptions of students’ motivation and learning behavior. Ethnic bias, however, appeared independent of these teacher perceptions. The results of study 3 support the assumption that inaccurately high or low teacher expectations may result in self-fulfilling prophecies and influence students’ achievement development as early as in the first school grade. The video study of study 3 further revealed that inaccurately high or low teacher expectations may result in different teacher feedback for the students. Teacher feedback, however, did not substantially mediate teacher expectancy effects on student achievement. The results of the present dissertation are of great relevance for educational practice. First, they highlight the advantages of high expectations and, second, they sensitize for biasing influences of student background characteristics.