Towards teaching analytics: a contextual model for analysis of students' evaluation of teaching through text mining...

Area of Science:

• Educational Technology
• Data Mining
• Machine Learning

Background:

• Teaching analytics (TA) is an emerging method in educational technology for improving teaching-learning processes using educational data.
• Student evaluations of teaching (SET) provide valuable data for analyzing teacher performance.
• Understanding the influence of student sentiment and gender on SET is crucial for effective educational feedback.

Purpose of the Study:

• To propose and validate an educational process and data mining plus machine learning (EPDM+ML) model for analyzing teacher performance based on SET data.
• To determine pedagogical factors influencing student recommendations and the role of sentiment and emotions in teacher evaluations, considering teacher gender.
• To predict student recommendations for teachers using student gender, average sentiment, and emotional valence from SET data.

Main Methods:

• Developed an EPDM+ML model integrating Text mining and Machine learning technologies.
• Applied Text mining to extract sentiments and emotions from 85,378 student comments.
• Utilized Analysis of Covariance and Kruskal Wallis Test to analyze quantified sentiment and emotional valence data, considering teacher gender.

Main Results:

• A large majority of comments (76.4% sentiment, 88.2% emotional valence) were positive or neutral.
• Female students' recommendations were significantly influenced by sentiment (p=.000).
• The EPDM+ML model achieved perfect scores (1.00) for precision, recall, specificity, accuracy, and F1-score, validated by k-fold cross-validation.

Conclusions:

• The EPDM+ML model is effective for analyzing SET data and identifying key factors influencing student evaluations.
• The model provides insights into how student sentiment, emotions, and gender impact teacher recommendations.
• This approach supports the advancement of teaching-learning processes and student learning experiences in evolving educational environments.