MEASURING STUDENTS’ FRAMING AND REFRAMING IN DESIGN: A SEMANTIC DISTANCE APPROACH

Understanding the cognitive processes that support designing is essential for fostering effective decision-making and creativity in design education. This study presents a novel approach to quantitatively characterize and measure framing and reframing (F-RF) behaviors during design problem-solving. F-RF is an essential cognitive process in problem-solving, influencing how problems and solutions are viewed. A frame, shaped by the problem-solver's knowledge, experience, and values, acts as a conceptual lens, affecting how information is interpreted and decisions are made. In design, framing plays a critical role by identifying and defining constraints, goals, and potential solutions, guiding designers in how they perceive and approach their tasks. Effective framing allows for the strategic redefinition of problems, fostering the exploration of ideas. Quantitative measurement methods for F-RF are underexplored. This approach presented here uses Natural Language Processing (NLP) techniques to analyze F-RF behaviors during design tasks. By employing verbal protocols, this methodology provides an objective means to track and assess designers’ F-RF behavior through their relative semantic distances. A case study involved a controlled experiment with five independent undergraduate students, tasked with designing a small museum. During a 55-minute design session, they verbalized their thoughts, which were recorded for analysis. The results demonstrated that the data-driven NLP approach effectively characterized and tracked F-RF behaviors throughout the design process. It facilitated identifying and measuring F-RF occurrences in students' verbal protocols. This approach also enabled the quantification of semantic distance values associated with design F-RF, revealing insights into the students' designing behavior. The design sessions had Designing and Finalizing phases. No significant differences were found in the number of frames generated during these phases. However, the semantic distance value of F-RF was higher in the Finalizing phase. The analysis revealed different patterns, with cyclical phases of high-semantic F-RF value – associated with an increase in divergent thinking– alternating with low-semantic F-RF– associated with convergent thinking. Unlike traditional models, which suggest initial divergence at the early stages of the process, followed by a final stage of convergence, this study revealed a more dynamic F-RF process involving a cyclical interplay of expansion of the design space. The study supports innovative assessment methods in design education, emphasizing the importance of understanding framing-related cognitive processes to foster effective design practices. The proposed NLP approach advances design research and enables new assessment strategies for design education, which may be used to enhance student performance. By tracking F-RF activity using NLP, educators gain deeper insights into students' design behaviors. Reflecting on these cognitive processes promotes the development of problem-framing skills, which are associated with high-quality creative design outcomes. Integrating this methodology in design studios can strengthen students' framing skills, improving their outcomes and preparing them for complex real-world design challenges. Additionally, it can address the need for assessment methods in multidisciplinary team-based design projects that emphasize collaborative design and framing. Additionally, developing automated feedback systems based on this methodology could provide timely relevant insights for educators and students. Such systems would increase students' awareness of framing behaviors, supporting their learning experience.