Abstract
•Academic self-concepts are formed by dimensional, social and temporal comparisons.•In contrasting school subjects, better achievement can lead to lower self-concepts.•School grades drive academic self-concepts more than test scores.•Dimensional comparison theory helps teacher provide more effective feedback.
In a comprehensive study (15,356 Dutch 9th grade students from 651 classes in 95 schools) we empirically tested the dimensional comparison theory (DCT) propositions formulated by Möller & Marsh (2013) as an extension of I/E theory, exploring methodological, theoretical, and substantive insights. According to DCT, academic self-concepts (ASC) are formed in relation to dimensional comparisons in different school subjects, as well as to social and temporal comparisons. In support of DCT predictions, paths from achievement to ASC in matching domains were substantially positive, but paths to non-matching domains (e.g., math achievement to verbal self-concept) were significantly negative. Extending DCT, we show that the more dissimilar the subjects, the more negative the cross paths (far comparisons), whereas cross paths relating more similar subjects (near comparisons) are much less negative and sometimes positive. Extending previous self-concept research and its integration with DCT, we found that positive paths for matching domains and negative paths for non-matching domains were larger for class marks based on classroom performance than for standardized test scores. Controlling for direct measures of social comparison (meVclass ratings of how each student compares to classmates) substantially reduced positive paths from achievement to ASC in matching domains, but also reduced the size of the negative paths from non-matching domains. Supplemental analyses suggest that dimensional comparison processes in both subjective rankings and actual class marks are consistent with those found in ASCs.