Acceso usuarios
Alejandro Sanín Posada
Se presenta en este artículo una discusión acerca de la calidad psicométrica de una batería de instrumentos que miden autoestima, autoeficacia, motivación, estilos de afrontamiento, crecimiento psicológico, locus de control y estabilidad emocional. Estas son características de la persona que, en interacción con el contexto de trabajo, pueden constituirse en riesgos psicosociales o factores protectores dependiendo del nivel en que se encuentren. Este proceso de verificación se realizó con base en la teoría psicométrica clásica. Se procedió a verificar la calidad de los reactivos diseñados en términos de su capacidad de discriminación y asociación con la variable que mide. Se calculó el nivel de confiabilidad general de las pruebas y se procedió a explorar su validez de construcción a partir de análisis factorial. Los datos obtenidos a través de estos métodos de verificación mostraron resultados favorables en todas las escalas, con excepción de la de Locus de Control y dos subpruebas de Motivación. En general se apreció una capacidad de discriminación de los ítems aceptable, índices de correlación Ítem-Test altos y significativos, niveles de confiabilidad muy favorables y una estructura factorial clara, indicadora de validez de construcción de los instrumentos. Se encontró además que algunas variables estaban conformadas por varios factores no conocidos previamente ni documentados en la literatura científica. También que algunos de los esperados y referenciados en la literatura no estaban presentes.AbstractIs presented in this paper a discussion about the psychometric quality of a battery of instruments that measure self-esteem, self-efficacy, motivation, coping styles, psychological growth, locus of control and emotional stability. These are characteristics of the person, interacting with the work context can become a psychosocial risk or protective factors depending on the level they are. This process of testing was carried out based on classical psychometric theory. We proceeded to check the quality of the items in terms of their agreement and association with the variable being measured. We calculated the level of overall reliability of the evidence and proceeded to explore its construct validity from factor analysis. The data obtained through these verification methods showed favorable results in all scales, with the exception of locus of control and two subscales of Motivation. Results show an acceptable item discriminate index, as well as favorable reliability levels and a clear factor structures suggesting appropriated construct validity. It was found that some variables were shaped by several factors not previously known or documented in the scientific literature. Also some of the expected and referenced in the literature were not present.
Is presented in this paper a discussion about the psychometric quality of a battery of instruments that measure self-esteem, self-efficacy, motivation, coping styles, psychological growth, locus of control and emotional stability. These are characteristics of the person, interacting with the work context can become a psychosocial risk or protective factors depending on the level they are. This process of testing was carried out based on classical psychometric theory. We proceeded to check the quality of the items in terms of their agreement and association with the variable being measured. We calculated the level of overall reliability of the evidence and proceeded to explore its construct validity from factor analysis. The data obtained through these verification methods showed favorable results in all scales, with the exception of locus of control and two subscales of Motivation. Results show an acceptable item discriminate index, as well as favorable reliability levels and a clear factor structures suggesting appropriated construct validity. It was found that some variables were shaped by several factors not previously known or documented in the scientific literature. Also some of the expected and referenced in the literature were not present.
