La teoría de respuesta a los itemes: una moderna alternativa para el análisis psicométrico de instrumentos de medición

Abstract

Compared to Classical Test Theory, Item Response Theory (IRT) represents a more powerful approach for the technical quality analysis of psychometric measurement instruments. Item Response Theory allows us to obtain more information about the test items and the construct or latent variable that is intended to be measured in the examinees. Likewise, with IRT, it is possible to generate tests adapted to the examinees levels. Furthermore, indicators of the instrument and the individual items psychometric properties are calculated according to the different values of the construct. These characteristics give IRT essential advantages over Classical Test Theory. However, there are still unsolved estimation problems, which sometimes make it impossible to adjust certain models on specific sets of data. To correctly understand the nature, potentials and limitations of IRT, knowledge in mathematics and statistics is required. Perhaps due to this fact, IRT approaches are not well known among social scientists in the Latin American region. In general, the two disciplines that can benefit the most from the application of these models are Psychology and Education.