Learner choice in a smart learning environment: insights into rationales and alignment with adaptive assignments

Abstract

Smart learning environments (SLEs) increasingly integrate adaptive technologies to personalize instruction. A central design question concerns the balance between adaptivity (system-controlled adaptation) and adaptability (learner-controlled modification). Hybrid approaches, such as adaptive recommendations that learners may override, aim to combine instructional optimization with learner autonomy. Information about the choices from students in these hybrid approaches can further inform future design of SLEs. This study investigates learners' task difficulty selections in a web-based SLE using worked examples for quantitative physics problem solving. A total of 530 secondary students were randomly assigned to either a fully adaptive condition, or an adaptable condition with adaptive recommendations. Log data and survey responses were analyzed to examine alignment with recommendations and underlying rationales. Learning gains were analyzed using an atomized problem-solving test. The adaptability condition shows adherence rates of approximately 84% resulting in similar learning paths for both conditions. Learning gains show no significant differences between both conditions. Students' rationales revealed reasons like trust in the system, effort avoidance, and self-challenging. Several appreciated the possibility to make choices. These findings suggest that adaptive recommendations can effectively nudge learner decisions in an adaptable SLE while the possibility for choice is appreciated, yielding learning outcomes comparable to a fully adaptive system. The study contributes empirical evidence on learner-system interaction in adaptable SLEs with adaptive recommendations and thus informs future designs of personalized learning environments. Limitations arise from many students that did not provide rationales and future research could focus more precisely on concepts like perceived autonomy.