Virtual learning environments give students more autonomy over their learning than traditional face-to-face classes and require that students adapt the ways they consume and assimilate new information. One theory of this process is self-regulated learning, which is illustrated in Efklides' Metacognitive and Affective model of Self-Regulated Learning (MASRL). MASRL represents the interplay between cognition, metacognition, and affect, both within a learner and between a learner and a task. This study uses learning analytics to operationalize Efklides' MASRL model in order to investigate the extent to which a combination of cognitive, metacognitive, and affective variables explains students' learning outcomes. This research was conducted at a private American university with 119 undergraduate students enrolled in four sections of an online or hybrid Calculus I course in fall 2020 and spring 2021. Five cognitive variables were defined and measured according to the Cognitive Operational framework for Analytics (COPA). Three metacognitive variables measured students' engagement with the course, and three affective variables measured students' affective states, as evidenced by digital traces in the LMS. Learning outcomes in this study were measured by students' final course grades. Binary logistic regression revealed that two cognitive, one metacognitive, and two affective variables were significant in explaining whether students' learning outcomes would be above or below the median. The confusion matrix and the area under the Receiver Operating Characteristics (ROC) curve showed high accuracy and usefulness for this regression model. The implications of these findings for online/hybrid learners and Efklides' MASRL model are subsequently explored.