The role of academic ability indicators in big-fish-little-pond effect research: A comparison study

ABSTRACT

The authors examined the big-fish-little-pond effect (BFLPE) on academic self-concept (ASC) using different indicators of academic ability (i.e., achievement test, cognitive ability test, grades corrected for grading-on-a-curve effects, uncorrected grades). They investigated under what circumstances grades are suitable indicators of academic ability in BFLPE research. The sample comprised 730 sixth-grade students from 30 classes belonging to the top track of the German secondary high school system. Using multilevel models, all indicators of academic ability exhibited negative contrast effects on ASC at class level (i.e., BFLPE). The authors found the strongest effects for corrected grades, followed by achievement tests, cognitive ability, and, finally, uncorrected grades. Thus, the study provides evidence for the usage of grades within BFLPE research for investigating the BFLPE.

KEYWORDS:

• Academic achievement
• academic self-concept
• big-fish-little-pond effect
• contrast effect
• multilevel modeling

Note

Acknowledgments

The data on which this study was based stem from the PULSS Project (team in alphabetical order: Bettina Harder, Monika Motschenbacher, Franzis Preckel, Wolfgang Schneider, Eva Stumpf, Susanne Trottler, Katharina Vogl, Christina Weiß, and Albert Ziegler), a two-cohort longitudinal study on school achievement and academic development at the beginning of secondary education. The project was funded by the Bavarian Ministry for Education and Culture; the Ministry for Culture, Youth, and Sports of the Federal State of Baden-Württemberg; and the Karg-Foundation (Frankfurt).

Notes

1. One prerequisite for the comparison of results from different studies is the computation of effect sizes, because the beta weights might be misleading. Unfortunately, the report of effect sizes is not yet a standard practice regarding reference group effects of class-average academic ability Marsh et al. (2009). We met those requirements in our study by presenting two effect sizes.