El uso de la invarianza de medición con variables dicotómicas como evidencia de validez

This paper explains and exemplifies how an analysis of the measurement invariance with dichotomous items may be implemented in order to obtain evidence of validity. The example was developed with data from the Academic Aptitude Test (PAA, for its initials in Spanish; PPPAA, 2014) at the University of Costa Rica (UCR). Measurement invariance was analyzed according to two classifications, by high school kind (public or private) and by sex. The method used was a confirmatory factor analysis with two correlated factors defined from the two components of the PAA (N = 11304). Each group analysis presented strict invariance. It was concluded that the test scores have the same unit of measurement for both men and women, and among students from public and private high schools. Finally, these results constitute evidence of validity of the PAA scores in terms of their internal structure from the perspective of fairness.