Person-centered approaches

The person-centered approach, rather than assuming population homogeneity, explicitly models unobserved population heterogeneity. Person-centered methods thus shift the attention from associations between variables to associations – or similarities – between persons. More precisely, the person-centered approach pays explicit attention to the identification of subpopulations displaying distinct configurations on a set of variables (e.g., Latent Profile Analyses) or different associations among variables (e.g., Mixture Regression Analyses). It also seeks to understand how these profiles, or configurations, differ from one another in terms of outcomes, are differentially influenced by predictors, and remain stable or not over time or across different types of individuals. The person-centered approach thus views individuals as complex beings that are best described in a holistic (i.e., describing individuals by looking at configurations across variables) rather than a reductionist (i.e., describing individuals by looking at differences on one or a limited set of variables) manner.