Comparison of Construct Mean Scores Across Populations: A Conceptual Framework
Applied researchers who wish to compare groups of people (e.g. different cultures) on the basis of a theoretically meaningful concept (e.g. individualism / collectivism) are faced with a number of methodological problems (e.g. measurement issues/cross-group comparability issues, ... etc.). Unfortunately, the availability of ‘coping strategies’ is largely dependent on the nature of the underlying concept (either represented by an ‘emergent’ or a ‘latent’ construct). If the concept represents an emergent construct (i.e. as represented by formative indicators), both measurement and cross-group comparability of construct quantifications are problematical. With latent constructs (as represented by reflective indicators), however, these methodological problems are much easier to overcome. By using multigroup Confirmatory Factor Analysis (CFA) models invariance of measurement across groups can be explicitly tested, and cross-group comparisons can be made (provided sufficient evidence exists for claiming cross-group comparability of construct scores).
KeywordsMeasurement Invariance Measurement Model Latent Construct Formative Indicator Meaningful Concept
Unable to display preview. Download preview PDF.
- Akaike, H. (1973), Information theory and an extension of the maximum likelihood principle. In Second international symposium in information theory, Petrov, N & F. Csaki (Eds.), 267–281, Akademiai Kiado, Budapest.Google Scholar
- Bollen, K.A. and Long, J.S. (Eds.)(1993), Testing structural equation models, Sage, Newbury Park, California.Google Scholar
- Hox, J.J. (1997). From theoretical concept to survey question. In Survey measurement and process quality, Lyberg, L. et al. (Eds.), 47–69, Wiley, New York.Google Scholar
- Sörbom, D. (1981). Structural equation models with structured means. In Systems under indirect observation: causality structure, and prediction, Jöreskog, K.G. and Wold, D. (Eds.), 183–195, North Holland, Amsterdam.Google Scholar
- Stelzl, I and Schnabel, K. (1992). The two-group MANOVA problem with unequal covariance matrices: a simulation study comparing Hotelling’s T2 to the LISREL approach. Methodika, 6, 54–75.Google Scholar