Exploring Indonesian Teachers’ Understanding and Attitudes Toward Deep Learning

Abstract The study investigates Indonesian teachers’ knowledge and attitude towards deep learning principles through a quantitative descriptive approach. Selected through accidental sampling techniques, a total of 33,394 teachers in Indonesia became the respondents. The data were collected using questionnaires; fifteen questions with a 7-point semantic differentiated scale contained knowledge (K) aspects, and twenty questions with a 5-point Likert scale contained attitude (A). The data were analyzed using k-means clustering, correlation analysis, and heatmap visualization techniques. The results showed that teachers generally understand the deep learning principles, especially reflective, collaborative, and contextual learning. However, their attitude towards the implementation indicates dualism – some are positive, while others are resistant to innovations, such as technology use and interdisciplinary approaches. Cluster analysis reveals a significant attitude-behavior gap with two distinct profiles: (1) teachers with a high understanding but low attitude, and (2) teachers with a positive attitude but a moderate level of understanding. The study highlights the importance of differentiated and cluster-based teacher professional development programs, contextual and practice-based training, as well as the affirmation of teachers in the frontier, the outermost, and the remote areas (3T). These allow the success of the Merdeka Curriculum and the educational transformation based on deep learning in Indonesia.