Validez de constructo y consistencia interna de tres estructuras factoriales y dos sistemas de puntuación del Cuestionario General de Salud-12

Abstract

Introduction. Scale psychometric properties analysis is important due to it allows decreasing the measurement bias in the research context. Objective. To determine the General Health Questionnaire's - 12-construct validity and internal consistency of three factor structures and two scoring methods. Materials and methods. Validation study in a nested sample of 483 dental students in Cartagena, Colombia. The scale was applied with a set of personal and academic questions as well. Validity was assessed for the following structures: one-dimensional, one-dimensional with the response-correction bias, two and three factors. Assessed scoring methods included Likert (0-1-2-3) and dichotomous (0-0-1-1). Validity was assessed through confirmatory factor analysis and internal consistency with Cronbach's Alpha coefficient for the Likert scoring method and Kuder-Richardson's formulae 20 for the dichotomous system. Analysis was performed using Stata v. 13.2 (StataCorp. TX, USA) and Mplus v. 7.11 (Muthén & Muthén, Los Angeles, USA). Results. Fit indexes showed as the best model the one-dimensional with response-correction bias with the following statistics degrees of freedom: 36; ?2: 52.432; root mean square error of approximation: 0.03; 90[%] CI: 0.008-0.048; comparative fit index: 0.982; Tucker-Lewis index: 0.966. Internal consistency was 0.70. Conclusions. Obtained model fit allows controlling measurement bias and then guarantees results external validity when GHQ-12 is employed for mental screening in this population. Scales psychometric properties should be critically assessed before any result analysis.