Evaluation Method and Empirical Analysis of the Achievement of Ideological and Political Education Goals in Digital Teaching Scenarios

Abstract

In digital and intelligent teaching scenarios, the goals of ideological and political education in courses face the dilemma of being “implicit and difficult to reveal, difficult to trace the process, and difficult to evaluate the effectiveness.” Traditional assessments struggle to dynamically capture the evolution of students’ ideological and political literacy. To address this, this paper proposes a method for assessing achievement by integrating multimodal process data. First, the three-dimensional goals of “cognitive identification—emotional internalization—behavioral practice” are deconstructed, and multi-source data mapping rules are established. Next, improved fuzzy hierarchical analysis is introduced for weight self-calibration, and an emotional temporal attention network is designed to extract deep emotional evolution features. Finally, a dynamic quantitative model is constructed based on evidence-based reasoning to achieve interpretable attribution. Experiments show that this method achieves a Kappa score of 0.92 with expert evaluation, and the three-dimensional internal reliability Cronbach’s α is higher than 0.80. The correlation coefficient between process assessment and final exam scores (0.78) is superior to that of traditional questionnaires (0.44), and the attribution accuracy is 85%. Ablation experiments confirm the key role of multimodal integration and emotional temporal mechanisms. This method provides an effective technical path for the process evaluation of ideological and political education in courses.