Predictions of self-regulated cognitive and metacognitve learning strategy use and access to online learning activities to university students’ academic success in a blended context

Research has consistently shown that students' self-reported self-regulated learning behaviors affect their level of academic success. However, self-reported data alone may not always represent what students actually do in learning. In modern tertiary education where learning management systems are widely adopted, the advancement of learning analytic data extracted by the complex algorithms from the learning management system make it possible to observe what students actually do in the online learning. Hence, this study used a combination of students' self-report data on the self-regulated learning strategy use, including both positive and negative cognitive and metacognitive strategy use; and the learning analytic data on the frequency of access to the online learning activities, to predict students' academic success among a cohort of 145 first year engineering students. Hierarchical regression analyses showed that while self-reported self-regulated learning strategy use alone could explain around 13% of variance in the students' final exam scores, adding learning analytic data could explained an extra 10% of variance in the students' final exam scores.