Cluster analysis of digital competencies among professors in higher education

Purpose This research focuses on the diagnosis and clustering of professor higher education in relation to digital competencies, based on different levels of digital competency development. Methods The methodology employed in this study involved an Ordinary Least Squares (OLS) regression analysis and cluster analysis using K-means clustering algorithm, considering the Silhouette score, based on the responses obtained through a questionnaire DigComEdu framework, and adjusted according to the experts who conducted a validity analysis. Results The findings indicate that, for the sample professor who answers the questionnaire voluntarily and confidentially, considering margin of error of 5%, a confidence level of 95%, and a response distribution of 50%, corresponding to 314 professors, with a Crombach’s alpha of 0.56, there is no relation between the variables of investigation, age, gender, academic degree, academic hierarchy, and years in academy and the level of digital competencies among professors. Regarding the clustering analysis, specifically using the K-means clustering algorithm, four distinct clusters are identified based on the questionnaire scores, aligning with findings from Silhouette score and Quadratic error by number of clusters. Discussion This research reveals that professors in higher education span all four levels of competency as defined by the DigComEdu model, primarily falling within the intermediate levels of digital competencies. Clustering analysis further provides insights for the implementation of enhancement and development policies, with the aim of guiding professors toward more complex digital activities, ultimately achieving the highest level of digital competencies. This, in turn, fosters improved teaching practices and, consequently, enhances the teaching experience.