Blended course designs have been increasingly adopted in medical education. However, research on the relations between the key aspects of students' learning experience and their learning outcomes often only measure students' self-reported experience, neglecting what they actually do in learning. This study combined both self-reported and observational measures of students' learning experiences and examined the relations between the two sets of measures and their contributions to learning outcomes. Australian medical students were asked to report their approaches to, and perceptions of, learning. Their frequency and duration of their interactions with both online formative and summative tasks were observed and recorded. Correlation analyses showed that the learning outcomes was positively related to deep approaches to using online technologies and duration of interactions with online summative tasks. The hierarchical multiple regression analysis found that the self-reported approaches and duration of interactions jointed explained the learning outcomes, accounting for 6% of the variance. The study demonstrated the complementary nature of using both self-reported and observational measures of students' learning experiences to explain the learning outcomes in blended course designs.