Assessing science teachers’ digital literacy with survey-based machine learning

Abstract

The COVID-19 pandemic has underscored the urgent need for digital literacy among educators, particularly science teachers who are expected to integrate technology into remote and hybrid learning environments. This study investigates the digital literacy levels of 172 science teachers in the Southeastern Anatolia Region of Turkey using a validated questionnaire, the Digital Literacy Scale. In addition to traditional scoring methods, two machine learning classification algorithms (Support Vector Machine with Sequential Minimal Optimization, SVM-SMO, and Multilayer Perceptron Neural Network, MLPNN) were applied to analyze and classify teachers’ digital literacy levels. The results showed that most participants demonstrated an average level of digital literacy. Among the models tested, the SVM-SMO algorithm achieved the highest classification accuracy of 92%, outperforming MLPNN. This study offers an innovative contribution by demonstrating how machine learning can enhance the analysis of survey-based educational data, providing a more nuanced and automated method for evaluating teacher competencies. The findings highlight both the ongoing need for targeted digital literacy training and the practical utility of AI tools in educational research.