Engagement and learning benefits of goal setting with rewards in human-ai tutoring

Active learning promises improved educational outcomes yet depends on students’ sustained motivation to engage in practice. Goal setting can enhance learner engagement. However, past evidence of the effectiveness of setting goals tends to be limited to non-digital learning settings and does not scale well as it requires active teacher or parent involvement. We study goal setting in a hybrid human-AI tutoring context, where students engage with personalized learning software with feedback and hints while being supported by a human tutor. Each student set a weekly goal (e.g., how much time or how many skills to master) and was rewarded for goal achievement, while human tutors regularly checked in about goal progress. We investigate whether this intervention improves student engagement and skill mastery. We observed 110 middle school students in a hybrid tutoring program over 12 weeks, with goal-setting support integrated into the program after six weeks. Using a quasi-experimental interrupted time series model, we estimated the intervention’s impact on engagement and learning. Weekly practice time increased, on average, by about 25% after introducing goal setting, and skills mastered per week increased by about 40%. This effect remained stable over time. These findings suggest that goal-setting support can improve the quantity and quality of practice in hybrid tutoring contexts with minimal added teacher workload. The present study advances the understanding of strategies that can increase student engagement with personalized learning systems in a low-cost and scalable manner, improving the learning benefits of such systems.