Students tend to choose a field of study in which they believe they possess the most skills by comparing their performance across different domains, such as math and English. These intraindividual comparisons between domains are known as dimensional comparisons (Möller & Marsh, 2013). There are individual differences in dimensional comparisons, such that some students engage in stronger comparisons than others do, yet few studies have examined the sources of these individual differences. In addition to objective performance, students sometimes also receive subjective feedback (e.g., praise) from parents and teachers. However, it is unknown whether and how this feedback influences dimensional comparisons to shape students’ domain-specific motivational beliefs. Therefore, we first examined whether theories of intelligence (TOIs) moderated dimensional comparison processes in different age groups in two studies. We then investigated whether receiving disproportionate praise in different domains affected students’ domain-specific motivational beliefs. Results showed that incremental TOI moderated math grades’ relation to verbal competence self-concepts in 10th and 11th graders (N = 140). It also moderated verbal grades’ relation to verbal intrinsic values in 1st- to 5th-year college students (N = 136). However, we did not find such moderations in other age groups. Regarding the praise manipulation, 7th to 9th graders (N = 108) showed heightened verbal competence self-concepts after receiving praise on either math or verbal performance. First- to fifth-year college students also showed increased verbal intrinsic values after receiving praise on verbal performance. TOI moderated students’ responses to praise manipulations. These studies shed light on students’ development of domain-specific motivational beliefs and inspire future research. / Psychology

Ziel des Dissertationsprojekts war eine differenzierte Betrachtung des akademischen Selbstkonzepts in den naturwissenschaftlichen Fächern. Unter anderen wurden die Abgrenzung des akademischen Selbstkonzepts von Selbstwirksamkeitserwartungen, seine dimensionale Struktur, Geschlechtsunterschiede sowie seine Beeinflussung durch die Konzeption des Naturwissenschaftsunterrichts und durch dimensionale Vergleiche untersucht. Im Mittelpunkt der ersten Teilstudie stand die Frage nach der empirischen Trennbarkeit von akademischem Selbstkonzept und Selbstwirksamkeitserwartung in den Naturwissenschaften. Es konnte gezeigt werden, dass die beiden Konstrukte messtheoretisch klar abgrenzbar sind und differentielle Zusammenhänge zu anderen Konstrukten aufweisen: Das Selbstkonzept wurde stärker durch soziale Vergleiche, Selbstwirksamkeitserwartungen stärker durch Unterrichtscharakteristika beeinflusst. In der zweiten Teilstudie wurde die fachspezifische Ausdifferenzierung naturwissenschaftlicher Selbstkonzepte untersucht. Es konnte gezeigt werden, dass Schüler zwischen den Selbstkonzeptfacetten in Biologie, Chemie und Physik unterscheiden und diese drei Facetten differentielle Zusammenhänge zu Leistungsmaßen und zum Geschlecht aufweisen. Die dritte Teilstudie beschäftigte sich mit interdisziplinärem Naturwissenschaftsunterricht als Einflussfaktor auf die Selbstkonzeptstruktur. Dabei konnte gezeigt werden, dass Schüler, die interdisziplinär beschult werden, deutlich höhere Zusammenhänge zwischen den Selbstkonzeptfacetten in Biologie, Chemie und Physik aufweisen als Schüler, die getrennt in den drei Fächern unterrichtet werden. In der vierten Teilstudie wurde der Einfluss dimensionaler Leistungsvergleiche auf naturwissenschaftliche Selbstkonzepte sowie Mathematik und Deutsch untersucht. Es zeigten sich Kontrasteffekte zwischen Mathematik, Physik und Chemie auf der einen und Deutsch auf der anderen Seite. Die Befunde haben Implikationen für die Theorie dimensionaler Vergleiche. / In this dissertation project, academic self-concept in the science subjects is scrutinized. Amongst others, its separability from self-efficacy, its dimensionality, and gender differences were examined. The dissertation includes four research articles. All are based on data from large-scale assessment studies. In the first research article, we examined whether academic self-concept and self-efficacy in the science subjects can be distinguished empirically. We found a moderate positive correlation between self-concept and self-efficacy in science, advocating distinguishable constructs. Furthermore, science self-concept was better predicted by the average peer achievement, whereas science self-efficacy was more strongly affected by inquiry-based learning opportunities. The second research article focused on the internal structure of academic self-concept in the sciences. It could be shown that students differentiate between their abilities in biology, chemistry and physics and that the subject-specific self-concept facets are differentially related to achievement and gender. For example, stereotypical gender differences in favor of boys were found for chemistry and physics. The effect of interdisciplinary science teaching on the structure of academic self-concept in the sciences was examined in the third research article. Students who were taught science as an interdisciplinary subject showed substantially higher relations between the self-concept facets in biology, chemistry, and physics than students who were taught these subjects separately. In the fourth research article, the effects of dimensional comparisons on academic self-concepts in the sciences as well as mathematics and German were scrutinized. We found contrast effects between the German and math, physics, and chemistry as well as assimilation effects between math, physics, and chemistry. The results further advance dimensional comparison theory.