Learning Analytics

Description

Learning analytics uses the data generated in educational settings to improve teaching and learning. In programming courses these applications are often effective at guiding interventions; however, many rely only on quantitative data and leave aside qualitative approaches and mixed methods.

Goal

To determine the relationships between learning-process metrics and academic performance in the Computer Programming course, through an analytics strategy grounded in mixed methods.

Methodology

We used an explanatory sequential mixed-methods design:

1. Data preparation and metric design
2. Quantitative and qualitative analysis
3. Identification of relationships with academic performance

We obtained 55 variables describing the learning process (submissions, verdict rates, software metrics on the code, and use of platform tools).

Results

• The number of submissions, the success rate and the number of correct and incorrect answers correlate positively with performance.
• Students who failed the course used UNCode more as an automatic grading tool; those who passed used it more to get feedback.

Keywords

Learning analytics · Computer programming · Mixed methods · Correlation analysis · Academic performance

Authors

Edna Johanna Chaparro Amaya
Master's student in Systems and Computing Engineering
Universidad Nacional de Colombia
Email: edchaparroa@unal.edu.co