Análisis de Clases Latentes como Enfoque de Modelamiento no Observado: fundamentos teóricos, supuestos y relaciones con los modelos de medición de Rasch

Latent Class Analysis (LCA) is a statistical approach aimed at identifying unobserved subpopulations based on patterns of responses to categorical variables. Unlike continuous latent trait models, LCA assumes that population heterogeneity can be represented through a finite number of qualitatively distinct latent classes. This article presents a theoretical review of latent class analysis, addressing its conceptual foundations, statistical formalization, model assumptions, and the criteria commonly used to determine the optimal number of classes. The conceptual and methodological relationship between LCA and Rasch measurement models is also examined, highlighting their similarities, differences, and potential complementary uses in educational research. The paper concludes with a critical reflection on the advantages, limitations, and current challenges of latent class approaches in the study of complex educational phenomena.