On the correspondence between inter- and intra-individual measurement models: An application

The implicit assumption in most psychological studies focusing on group-level interpretations is that the interindividual (i.e., between-person) variability can mirror the intraindividual (i.e., within-person) variability over time. In other words, measuring a hundred individuals at one time point is considered equivalent to measuring one individual a hundred times. Such an assumption needs to be tested in the context of psychological constructs of interest. In this thesis, we aimed to test this assumption in the context of self-esteem, to illustrate the heterogeneity among individuals, and to highlight the contribution of the person-oriented research paradigm for psychological science. For these aims, we collected intensive longitudinal data, specifically from a small sample of individuals (N = 32), asking about their current experience of their self-esteem at multiple measurement occasions (T = 112). In the first part of the thesis, as a group-level examination, we employed the multilevel random coefficient modeling to explore the decomposition of trait and state variance in self-esteem. In the second part, as an individual-level examination, we used latent variable group iterative multiple model estimation (LV-GIMME) to investigate the extent of the correspondence between interindividual and intraindividual measurement models of self-esteem. Specifically, we focused on the uniqueness of each individual, explored the heterogeneity in their measurement models of self-esteem, and discussed the generalizability of the group-level findings on the decomposition of variance in self-esteem. The findings indicated that a group-level analysis for self-esteem could not be generalized to everyone in the sample due to the non-correspondence between intraindividual and interindividual measurement structures, as well as the presence of heterogeneity among individuals at different levels. All in all, we recommended the meticulous examination of individuals in a sample before engaging in a group-level generalization of findings, highlighting the importance of the person-oriented paradigm.