Supporting Deep Learning in Complex Problem-Solving Contexts through Process Visualization and Adaptive Feedback

Abstract

Objectives: Learning through problem solving has received wide attentions especially in complex domains such as medical education, where problem-solving experience is regarded as crucial to learning and expertise development (Schmidt et al., 1990). Meanwhile, research on expertise development has shown that desired learning outcomes in problem-solving contexts cannot be achieved by a mere accumulation of experience, but require systematic and deliberate effort with expert help (Ericsson, 2008; Jarodzka et al., 2010). It is important to engage learners in reflective thinking and practice with expert help to enable continuous learning and improvement. This study examined the effects of a computer-supported reflective learning approach that allowed students to review their own problem-solving process in a visual format and receive the feedback about the gap between their performance and that of the expert when working with a number of simulated clinical cases. Theoretical framework: Problem solving provides opportunities for deep learning through practical experience with real-world problems and authentic tasks. According to experiential learning theory, the outcomes of experiential learning depend not only on participation, but also on meaningful reflection on the experience (Dewey, 1938; Moon, 1999). Problem solving, especially in ill-structured domains, often involves sophisticated processes that are difficult to capture and master (Jonassen, 1997; Kirschner et al., 2006). Reflective learning requires necessary support that helps students to self-reflect on their problem-solving and thinking process, observe experts’ processes for comparison and contrast, and receive timely prompts (Lin et al., 1999). Cognitive apprenticeship theory highlights the importance of externalizing the process of complex tasks for novices to observe and practice with expert help (Collins et al., 1991). Methods: Twenty-five third year students from a medical school participated in the study. They used a computer-based reflective learning approach implemented in an online system for their independent study with five simulated glaucoma cases within five weeks. For each case, students could access initial information, select clinical examinations, and make intermediate judgments based on examination results; based on available information, students determined to perform further examinations or make a diagnostic conclusion. They could diagnose a case more than once. After completing a diagnosis, they could view their own diagnostic process in a visual format, and then determine the degree of similarity between their performance and that of the expert on a set of indicators, all generated by the computer. Once the similarity reached 60%, the expert’s problem-solving process could be viewed by students. Students’ learning outcomes were assessed via knowledge tests, diagnostic tasks, and professional tests before and after the learning program. Results: A significant pre-post improvement was found in students’ performance in knowledge tests and diagnostic tasks. Students also made pre-post improvement in two professional tests for diagnostic pattern recognition and clinical data interpretation. Significance: It is critical to externalize complex cognitive processes in problem-solving contexts and provide learners with timely adaptive feedback to support reflective thinking and practice. The results of the study have shown the potential of computer-supported process visualization and adaptive feedback to support individual performance based on expert knowledge.