Re-evaluation of the D21-digital-index assessment instrument for measuring higher-level digital competences

Abstract

The development of reliable tools for assessing digital competences is of great importance. This is why we set out a re-evaluation of the measurement quality of the D21-Digital-Index assessment instrument. The D21-Digital-Index is a biannual and influential study held in Germany. The instrument used in the D21-surveys is based on the theoretical framework DigComp. In our analyses we used data of 1142 participants from vocational training and higher education institutions to estimate item parameters and the quality of the instrument using Item Response Theory. Because choosing an appropriate IRT-model is crucial for instrument evaluation, we calculated and compared two types of models, the Rasch and the Birnbaum model of which the latter turned out to achieve the better fit. In a unidimensional analysis the five scales of the instrument with 24 items in total yield acceptable measures. Multidimensional analysis shows a dimensional separation and hence confirms the construct validity of the instrument.