Integrating Deep Learning Models into Civic Education: Effects on Students' Critical Thinking and Democratic Values

Abstract

Civic education is expected to develop critical thinking and democratic values, yet traditional instructional practices often limit these outcomes. This study investigates the effects of deep learning pedagogical models—understood here as instructional approaches emphasizing meaningful engagement, inquiry, and authentic problem-solving, distinct from deep learning in artificial intelligence on students' critical thinking skills and democratic value formation in Indonesian high schools. A mixed-method approach was employed involving 150 students assigned to an experimental group and a control group. The experimental group received instruction through problem-based learning, deliberative discussion, service learning, case analysis, and reflective activities, while the control group followed conventional teaching methods. Quantitative data were collected using the California Critical Thinking Skills Test and the Democratic Values Scale, supported by classroom observations and semi-structured interviews. The results show that students in the experimental group achieved significantly higher critical thinking scores (M = 78.5) than those in the control group (M = 64.8), with a large effect size (d = 1.58). Similar patterns emerged for democratic values, particularly political tolerance, civic engagement, and political efficacy (d = 1.46). Qualitative findings confirm that students became more analytical, open to differing views, and confident in civic participation. These findings contribute to civic education research by providing empirical evidence that integrated deep learning pedagogical models can simultaneously develop both cognitive and democratic competencies, offering a replicable instructional framework for transformative citizenship education in diverse societal contexts.