Hybrid Regression Analysis of Predictors of Teachers' ICT Use in Teaching Practices: Evidence from ICILS

The effective integration of Information and Communication Technology (ICT) in education is crucial for student success in the digital age, and large-scale international assessments provide valuable data for understanding and improving these practices. This study investigates the factors influencing teachers' ICT integration in teaching using the ICILS 2018 dataset. The study combines traditional methods, such as multiple linear regression, with machine learning techniques like the random forest algorithm, to comprehensively understand the factors driving teachers' ICT use. The findings reveal that the models successfully explained a significant portion of the variance in teachers' ICT practices. The random forest analysis identified country context as the most influential predictor, followed by teachers' positive views on the benefits of ICT, teacher collaboration, ICT self-efficacy, and experience with online lessons. Factors such as experience in preparing online materials and teacher age had a moderate influence, while access to computer resources and gender showed minimal impact. The results encourage educators and policymakers to invest in modifiable factors, such as teacher collaboration, self-efficacy, and positive perceptions of ICT, to enhance teachers' ICT integration.