Learning Analytics Dataset on Student Attendance, Correctness, and Assessment Performance in a Paid Non-Formal Online Learning Platform

This dataset contains anonymized learning analytics logs collected from 7,415 students enrolled in CoLearn, a large-scale, subscription-based nonformal online learning platform operating throughout Indonesia, from July through December 2025. The platform provides online tutoring sessions in mathematics and natural sciences for students in grades 4-11, covering elementary, middle, and high school levels. This dataset was compiled from the platform’s learning management system and attendance records to support research on student retention and learning performance in non-formal online education. This dataset consists of student-level learning logs summarizing their participation over the course of one academic semester. Included variables are user identifiers (which have been anonymized), educational level, grade level, number of classes attended, attendance rate, attendance consistency (including attendance lasting more than 30 minutes), average participation duration, participation in assessments, number of assessment attempts, number of correct answers, accuracy rate, and other derived learning analytics indicators used to evaluate student engagement and academic achievement. These variables are generated directly from students’ interactions with the learning platform and assessment activities. These tutoring sessions were conducted synchronously via Zoom using an interactive teaching approach that emphasized active participation and formative assessment. As a subscription-based nonformal education platform, CoLearn allows students to voluntarily participate in online learning outside of regular school hours. Therefore, the attendance records in this dataset reflect ongoing participation, not mandatory school attendance. All personally identifiable information was removed prior to data preparation. Student identities were replaced with anonymized user identifiers, ensuring that no names, contact information, school names, or other sensitive data were included, while maintaining the integrity of the learning analysis variables. This dataset supports the study "Does Correctness Matter? Investigating Its Influence on Student Retention in Paid Non-Formal Online Learning Platforms." It provides a valuable resource for examining the relationships between attendance, assessment performance, and student retention in paid non-formal online learning environments. The dataset may also be used to explore learning behaviors across educational levels and develop data-driven strategies to improve student engagement and retention.