Importance models of the physical self: improved methodology supports a normative-cultural importance model but not the individual importance model

We examine theoretical and methodological issues associated with the roles of individual and group-normative importance in self-esteem determination. Critical issues include multicollinearity among the physical self-subdomains, which may have affected previous results, and the need for a multidimensional perspective on importance models. We apply state-of-the-art methodologies, including exploratory structural equation modelling and the product-of-indicators approach to latent interactions. Positive interactions would be required to support the individually importance-weighted average model, but none were observed in the multidimensional model estimated on the full sample. Nonetheless, some interaction effects were found in the country-specific version of the model. Rather, we found support for the alternative group importance-weighted average model. We conclude that domain-specific self-concepts are weighted differently and thus differentially affect self-esteem, but these weights do not seem to depend on individual differences in importance. Although awaiting confirmation from further studies, our results suggest the idea that individuals use mainly normative importance processes based on cultural factors in weighting each domain-specific component of self-concept.