Why is support for Jamesian actual–ideal discrepancy model so elusive? A latent-variable approach

•We investigate the Actual–Ideal Discrepancy (AID) model of self-esteem determination.•We apply a latent difference approach to self-concept domains and sub-domains.•Young adolescents from two different countries (UK: 402; Italy: 250) are involved.•The effects of the AID remained elusive for most domains and subdomains.•AIDs did not explain the more general self-concepts beyond actual domains.

We investigate the Actual–Ideal Discrepancy (AID) model of self-esteem determination dating back to James (1890/1963). Although intuitively appealing, this model received weak support from rigorous empirical research. We propose a multiple-item latent difference approach to AID as applied to a range of self-concept domains and sub-domains in young adolescents from two different countries (UK: N = 402; Italy: N = 250). The effects of the AID remained elusive for most domains and sub-domains; indeed, the effects of the specific AIDs were generally trivial and only appearance-AID became significant in the combined sample. Also AIDs did not substantially explain the more general self-concepts beyond what was explained by the actual domains. Even if they had been significant, AID effects would have been, at best, trivially small and detectable only using appropriate latent-variable methodologies coupled with large sample sizes; thus undermining the psychological meaning of the AID model.