The purpose of this study is to continue to explore which function identification methods work best for specific design tasks. Prior literature describes the top-down and bottom-up approaches as equivalent methods for functional decomposition. Building on our prior work, this study tests the bottom-up method against the top-down and enumeration methods. We used a 3-factor within-subject study (n=136). While most of our diagramoriented metrics were not statistically different, we found statistical support that: 1.) students reported that the dissection activity was more useful when using bottom-up, and 2.) that student engineers committed many more syntax errors when using the bottom-up method (by listing parts instead of functions). We believe that both these results are due to the increased focus on individual parts. We do not know if an increased attention to the parts would increase novelty or fixation, and recommend future studies to find out.
