Data with character: towards student-centered learning in tertiary education
Options
BORIS DOI
Abstract
Student numbers in tertiary education continue to grow, alongside increasing societal and individual expectations for educational quality. Within this context, blended learning frameworks, the combination of online and in-person instruction, present substantial potential to identify predictors of academic performance. Combined with learning analytics, they enable personalized, data-driven student support. Over several years, an Integrated Blended Learning Analytics Framework (IBLAF) in a real-world university setting was applied to examine student-related variables, including prior knowledge, level of understanding, self-efficacy, learning behavior, and personality traits, as predictors of academic performance. By advancing understanding of the variables shaping academic performance, this research provides a foundation for targeted interventions to support individual and flexible learning pathways.
Three empirical studies were conducted. Study 1 investigated the stability of predictive variables across four university courses from three different disciplines (N = 1,186) and over two academic years. Prior knowledge and performance in online formative assessments (OFAs) emerged as stable and course-independent predictors of academic performance. Study 2 examined the relationship between students’ self-assessed and objectively measured understanding in OFAs during the term and explored self-efficacy at varying levels of specificity and its joint impact on academic performance (N = 271). Objectively measured understanding was the strongest predictor, with academic self-efficacy providing an additional direct effect in predicting academic performance. Study 3 focused on conscientiousness, examining whether it can be validly assessed through behavioral indicators derived from learning activity data within the IBLAF (N = 221). This behavioral measure showed stronger predictive power for academic performance than traditional self-report questionnaires, suggesting a more context-sensitive and adaptable assessment approach. In addition, an exploratory analysis of students' evaluations of the IBLAF experience revealed a generally high level of satisfaction, despite considerable heterogeneity in students' needs and study preferences.
Overall, the findings demonstrate the potential of blended learning frameworks enriched with learning analytics to generate empirically grounded, psychologically informed insights into key variables influencing student learning and performance, thereby providing a foundation for targeted, flexible, and personalized educational interventions.
Three empirical studies were conducted. Study 1 investigated the stability of predictive variables across four university courses from three different disciplines (N = 1,186) and over two academic years. Prior knowledge and performance in online formative assessments (OFAs) emerged as stable and course-independent predictors of academic performance. Study 2 examined the relationship between students’ self-assessed and objectively measured understanding in OFAs during the term and explored self-efficacy at varying levels of specificity and its joint impact on academic performance (N = 271). Objectively measured understanding was the strongest predictor, with academic self-efficacy providing an additional direct effect in predicting academic performance. Study 3 focused on conscientiousness, examining whether it can be validly assessed through behavioral indicators derived from learning activity data within the IBLAF (N = 221). This behavioral measure showed stronger predictive power for academic performance than traditional self-report questionnaires, suggesting a more context-sensitive and adaptable assessment approach. In addition, an exploratory analysis of students' evaluations of the IBLAF experience revealed a generally high level of satisfaction, despite considerable heterogeneity in students' needs and study preferences.
Overall, the findings demonstrate the potential of blended learning frameworks enriched with learning analytics to generate empirically grounded, psychologically informed insights into key variables influencing student learning and performance, thereby providing a foundation for targeted, flexible, and personalized educational interventions.
Date of Publication
2025
Year of graduation
2025
Theses Type
dissertation
Subject(s)
Keyword(s)
Learning Analytics
•
Higher Education
•
Blended Learning
•
Personality
Language(s)
en
Author(s)
Faculty/Graduate School
Access(Rights)
open.access
Primary OA Publication
true