Multidimensional models involving mastery, self-esteem, sensation seeking and perceived stress, its relations with behavioral and emotional problems and invariances across different pubertal stages

Abstract

It's notorious the lack of instruments with construct validity evidences that we have in the country to evaluate sensible and crucial stages of mental and intellectual development of young Brazilians. Robust analyses via structural equation modeling (confirmatory factor analysis, invariance analysis, latent class analysis) are still narrowly spread, which impair the knowledge of psychological processes, its stabilities and comparability between groups (model invariance tests). This subproject aims to evaluate different measurement and invariance models of the questionnaires proposed in the main project (mastery, perceived stress, self-esteem, sensation seeking, risk behavior and Pubertal Development Scale) across different pubertal stages, ages and sexes. With construct validity evidences of these measurement models, they will be compared with the measurement model of behavioral and emotional problems proposed by the Child Behavior Checklist (CBCL). Confirmatory factor analyses will be used for each one of the measurement models individually. Adequation indexes will evaluate the models' adjustment. Invariance analyses will be used to evaluate the measurement models' stability through two different techniques: multigroup confirmatory factor analysis (for dichotomous variables such as sex) and multiple causes, multiple indicators (MIMIC) for continuous and ordinal measures. Mplus software will be used for modeling. (AU)