What can online traces tell us about students' self-regulated learning? A systematic review of online trace data analysis

"Self-regulated learning" (SRL) is defined as taking responsibility for one's own learning. Selfregulatory skills are crucial to learners' success in the online learning context. Although research on SRL is expanding in recent years, much of the literature has relied on self-reporting tools to measure SRL. Online trace data analysis, an emerging approach, provides the promise of greater authenticity and convenience in measuring SRL as compared to self-reports. We conducted a systematic review of online trace data analysis that measured SRL in various learning platforms to address three research questions: "How did previous studies use online trace data as indicators of SRL?", "What approaches are being used to interpret the online trace data?", and "What are the challenges of using online trace data to measure SRL?". We systematically searched seven bibliographic databases with specific inclusion and exclusion criteria. A total of 38 empirical studies were eventually examined. We leveraged the two most cited SRL models as theoretical basis and mapped the various online trace data into relevant SRL process to answer the first research question. Two commonly adopted approaches to interpret the online trace data were identified. Three key challenges pertaining to the use of trace data to measure SRL were identified: time segmentation, generalization, and validity. We discussed these challenges and the possible means to mitigate them. Finally, we propose a flowchart to guide future studies in conducting online trace data analysis in SRL research.