There is a series of conventions governing how Confirmatory Factor Analysis gets applied, from minimum sample size to the number of items representing each factor, to estimation of factor loadings so they may be interpreted. In their implementation, these rules sometimes lead to unjustified decisions, because they sideline important questions about a model’s practical significance and validity. Conducting a Monte Carlo simulation study, the present research shows the compensatory effects of sample size, number of items, and strength of factor loadings on the stability of parameter estimation when Confirmatory Factor Analysis is conducted. The results point to various scenarios in which bad decisions are easy to make and not detectable through goodness of fit evaluation. In light of the findings, these authors alert researchers to the possible consequences of arbitrary rule following while validating factor models. Before applying the rules, we recommend that the applied researcher conduct their own simulation studies, to determine what conditions would guarantee a stable solution for the particular factor model in question.

En Psicología es frecuente encontrar situaciones en las que se necesita realizar algún tipo de clasificación de personas en subgrupos o clases. Existen técnicas de análisis multivariado como el Análisis Clúster Jerárquico (HCA) que se utilizan habitualmente para este fin. Actualmente, existe un interés creciente por la técnica de Análisis de Clases Latentes (LCA), si bien es una técnica relativamente poco conocida y utilizada. Varios autores han destacado que el LCA presenta importantes ventajas respecto al HCA, en especial que el LCA permite obtener medidas de bondad de ajuste. El objetivo de este trabajo es presentar varias aplicaciones del LCA tanto a partir de un estudio de simulación como a partir de datos reales, y comparar el desempeño de esta técnica frente al HCA. Los resultados a partir de la simulación indican que el LCA tiene una elevada capacidad para detectar estructuras de clase. Los resultados del estudio a partir de datos reales muestran que las distintas clases o mixturas presentes en los datos pueden estar solapadas, lo que dificulta la agrupación de clases al aplicar LCA. El HCA puede ser una buena herramienta de análisis para el investigador aplicado, ya que puede orientar sobre el mejor modelo de LCA que se debería interpretar. En contextos de investigación en los que el modelo teórico no es claro, se recomienda utilizar ambas técnicas con el fin de buscar convergencia de resultados.

Scale Validation Conducting Confirmatory Factor Analysis: A Monte Carlo Simulation Study With LISREL

The Patient Health Questionnaire (PHQ-15) has been widely used to assess somatic symptoms. This study aimed to analyze the psychometric properties of the Spanish version of the PHQ-15, its structure and score distribution across demographic variables in a Spanish sample. In addition, we examined variations in somatic symptoms among different demographic subgroups. 1495 individuals from the Spanish population answered a series of self-reported measures, including PHQ-15. To examine the factorial structure of the PHQ-15, Confirmatory Factor Analysis (CFA) was performed. Additionally, a bifactor CFA model was examined using the Exploratory Structural Equation Modeling (ESEM) framework. Women showed more somatic symptoms than men, and younger individuals showed more somatic symptoms than the older ones. It was also revealed positive associations between somatic symptoms and levels of depression, anxiety, and suspiciousness, while negative associations were found between somatic symptoms and perceived resilience and happiness. Regarding the factorial structure of the PHQ-15, although the one-factor and bifactor models were suitable, the bifactor model underscores the presence of a robust general factor. It is a cross-sectional study, not including non-institutionalized individuals. Somatic symptoms are more frequent in women and younger individuals. Furthermore, the presence of physical symptoms is associated to other psychological aspects, such as depression or anxiety. Finally, bifactor model was the most appropriate to explain the factorial structure of the PHQ-15.

Emotional Intelligence (EI) is considered a fundamental variable for a person’s adequate psychosocial adjustment. In education, its importance transcends the level of interpersonal relationships, and has been proposed as a variable that somehow influences academic performance, although there is controversy about whether its effect is direct or, rather, an intermediate variable. The present research analyses, from a sample of 327 students (52.6% female and mean age = 14.5), the relationship of EI with respect to the knowledge and management of oral communication and reading meta-comprehension strategies, which should directly affect different educational outcomes. In order to assess both the direct and indirect effects of these variables, a Partial Least Squares Structural Equation Modelling (PLS-SEM) approach has been proposed, due to its versatility and the possibility of jointly analysing reflective and formative measures. The results show that EI indirectly affects reading self-concept and reading comprehension, as it is involved in the management and handling of both effective oral communication and reading meta-comprehension strategies.