Cognitive Effects of Physical Models in Engineering Idea Generation

Abstract

Designers use various representations to externalize their ideas, physical models being an important one. Physical models are widely used by designers and their use is promoted as an effective design tool by industry and government agencies. However, very little is known about the cognitive effects of physical models in the design process; the available guidelines are conflicting. Some researchers argue for the frequent implementation of physical models, while others observe that the use of physical models fixates designers. In light of these conflicts, the research discussed in this dissertation focuses on understanding the cognitive effects of physical models and developing guidelines for aiding designers in their implementation. A combination of controlled lab studies and qualitative studies is adopted to achieve said goal. The results from the controlled studies show that physical models supplement designers? erroneous mental models and help them to come up with more ideas satisfying the problem requirements. These studies also demonstrate that design fixation is not inherent in physical modeling, but it is caused by the Sunk Cost Effect. According to Sunk Cost Effect, as designers spend more time building physical models of their initial ideas, they tend to fixate more to the variations of those ideas. A qualitative study on industry-sponsored projects and development cases of award-winning products further supports these results in more realistic situations. Further, the studies reported in this dissertation show that physical models can be effective tools for the mitigation of fixation to undesirable design features in a flawed example; however, these results can also depend upon the experience level of a designer in solving open-ended design problems. With these insights from the series of studies, a set of guidelines and a Model Error Reeducation Method (MERM) are formulated and tested with novice designers. MERM helps designers in identifying critical loads and interface designs they miss in their original designs, before prototyping. The results from the testing of this method show that this method is very useful in avoiding said errors in physical modeling.