The use of latent construct measurement scales is widespread in the medical literature. An essential condition is that the instrument be validated, in addition to being highly reliable.
Objective: To evaluate the level of compliance with various methodological and statistical premises of validation studies of psychometric instruments in pediatrics, which are necessary to ensure adequate interpretation of the results, in order to offer recommendations for future research.
Methods: Articles published in two of the most prestigious journals in Spanish (Andes Pediatrica and Anales de Pediatría) were identified between December 2000 and December 2023, where factor analysis was performed on the items of measurement scales, and in which the reliability and validity of these scales were studied. Twenty-five articles that met the selection criteria were selected and compared with the analysis procedure recommendations.
Results: Five key points were identified to improve: 1) Reflective and formative indicators; 2) Determination of the best reflective indicator; 3) Validity of a scale linked to the fit of the theoretical model; 4) Reliability using Cronbach's alpha; 5) Factor analysis versus principal components. Factor rotation.
Conclusions: The analysis processes of measurement instruments in pediatrics must be improved, especially regarding the selection and specification of indicators, the use of factor analysis, and the fit of the proposed models.