Discovering Insights in Learning Analytics Through a Mixed-Methods Framework: Application to Computer Programming Education

Aim/Purpose: This article proposes a framework based on sequential explanatory mixed-methods design in the learning analytics domain to enhance models used support success of process and learner. The consists three main phases: (1) quantitative data analysis; (2) qualitative (3) integration discussion results. Furthermore, we illustrated application this by examining relationships between metrics academic performance subject Computer Programming coupled with content analysis responses students’ perception questionnaire their experiences subject. Background: There is prevalence research designs analytics, which limits understanding processes. due abundance ease collection virtual environments management systems compared data. Methodology: study uses mixed-methods, non-experimental, design. phase aims analyze identify behaviors, trends, measures using correlation or regression analysis. On other hand, focuses conducting was applied historical from use an automated feedback evaluation platform for programming exercises course at National University Colombia during 2019 2020. question is: How can generate better variables generated throughout students Programming? Contribution: contribution work proposal applicable computer courses, allows complementing, corroborating, refuting quantitatively evidenced results generating hypotheses about possible causes explanations student behavior. In addition, provide processes Colombia. Findings: A field has been proposed improve answer posed corresponds that mixed methods effectively complement From framework, it appears data, representing perceptions students, generally supported extended consistency two phases allowed us behavior course. Recommendations Practitioners: We suggest implementing various educational contexts populations. By doing so, practitioners gather more diverse insights, lead different settings groups learners. Recommendation Researchers: Researchers approach projects, usually exclusively analysis, results, find understand depth thanks information provided complementary Impact Society: limit our Instead, propose suggests comprehensive includes deeper insight into Ultimately, effective interventions improvements teaching practices. Future Research: Potential lines continue mixed-method methodology include following: first, population sample, such as universities knowledge areas; second, techniques correct unbalanced sets studies; third, analyzing interactions grading activities relation activity grades; last, findings positively impact evaluating statistically through experimental designs. context introductory education, AI/large language have potential revolutionize enhancing experience, providing personalized support, enabling efficient assessment mechanisms. area implement these models.