Incorporating Desirable Difficulties into the design of digital learning: A think-aloud study

Background The use of digital learning platforms is becoming increasingly common in health professions education, but there are few studies evaluating their effectiveness based on cognitive science principles. Here, we draw upon well-established learning concepts grounded in cognitive science – desirable difficulties, productive struggle and cognitive load theory – to explore whether and how digital learning platforms can enhance learning. Our study aims were to i) identify points of struggle when participants processed learning material on a digital platform compared to traditional PDF learning materials and ii) explore how design factors influenced their subsequent responses. Methods Using in-depth think-aloud interviews, we compared how medical student participants engaged with learning material in traditional PDF format versus similar content on an online learning platform (five participants, two interviews each for a total of ten interviews). Participants were instructed to navigate the learning material as they would in their usual practice while verbalizing their thoughts during the process. Interviews were conducted on Zoom and video-recorded. Transcripts and videos were analysed using thematic analysis. Results We identified three themes centred around non-struggle, generative struggle and non-generative struggle. In the absence of struggle, learners tended to learn superficially and remained disengaged from the material. Desirable difficulties, such as in the form of online quizzes, enhanced learning through introducing points of struggle that led to deeper processing of information, which we term generative struggle. However, struggle that resulted from increased extraneous load due to flawed design was counter-productive and led to disengagement, which we term non-generative struggle. Conclusions Based on our work, we propose a model for how increased cognitive load due to desirable difficulties promotes generative processing and greater engagement with the learning material. This work can guide the future design and evaluation of digital learning platforms for more effective learning based on cognitive science principles.