<![CDATA[This study aims to evaluate the effectiveness of hybrid teaching in enhancing learning outcomes, learning engagement, and student satisfaction in university-level data science education. A convergent mixed-methods design was employed to integrate quantitative and qualitative evidence from a Data Preprocessing course at Y University, China. The study involved 242 undergraduate students, comprising 122 students in the hybrid teaching group and 120 students in the conventional lecture-based group. Data were collected using validated five-point Likert-scale questionnaires, academic performance records, and semi-structured interviews. Quantitative data were analyzed using descriptive statistics, reliability and validity analyses, independent-samples t-tests, Pearson correlation analysis, and multiple linear regression, while qualitative data were analyzed using thematic analysis. The results showed that students in the hybrid teaching group achieved significantly higher theoretical examination scores (M = 76.45 vs. 74.04, p = .023), practical performance scores (M = 84.91 vs. 75.97, p < .001), and overall academic achievement scores (M = 81.53 vs. 75.20, p < .001) than those in the conventional instruction group. Learning engagement emerged as the strongest predictor of overall satisfaction (β = .455, p < .001), followed by knowledge mastery and skill enhancement (β = .228, p < .05). Qualitative findings further indicated that students valued the flexibility of self-paced learning, collaborative classroom interactions, and authentic programming experiences. Overall, the findings suggest that hybrid teaching is a promising approach for supporting academic achievement, practical competencies, learning engagement, and student satisfaction. The findings support the integration of hybrid learning into data science curricula by combining flexible digital learning resources with face-to-face programming practice, collaborative problem-solving, and authentic data-processing activities.]]>
