Análisis de datos en gestión deportivala conmensurabilidad entre tamaños de efecto proveniente de escalas de medida con diferente rango

Abstract

This study analyses the problem of commensurability between effect sizes calculated from Likert scales with different rank. Standardized mean difference has been the index considered for that purpose. We have defined the scale invariance term and we have deduced the expressions that relate the possible distortion provoked by changing the scale rank on the mean and variance statistics. The analysis of these equations drives to the following conclusions: (1) Different rating scales distort measurement; (2) Effect sizes are commensurable to the extent that scale invariance holds. This is achieved when distortion is negligible; (3) When scale invariance is not guaranteed, the comparison between different studies can yield misleading results.