Examining big-fish-little-pond-effects across 49 countries: a multilevel latent variable modelling approach

Abstract

Using data from the Trends in International Mathematics and Science Study (TIMSS) 2007, this study examined the big-fish-little-pond-effects (BFLPEs) in 49 countries. In this study, the effect of math ability on math self-concept was decomposed into a within- and a between-level components using implicit mean centring and the complex data structure was considered. Sampling error was thus reduced. In addition, measurement error was minimised by modelling math self-concept as a latent factor at both levels and multilevel measurement invariance was assessed in each country. The BFLPE in each country was a contextual effect modelled as the difference between the between-level and within-level effects. Results suggest that BFLPEs existed in most of those 49 countries but the phenomena were different in those countries. The findings also suggest that when perceived relative standing to classmates was added to the model, BFLPEs decreased. Compared with previous cross-national BFLPE studies, this study demonstrated larger BFLPEs in the academic subject of math in individual countries.

Keywords:

• big-fish-little-pond-effect
• cross-national study
• multilevel latent variable modelling
• TIMSS