Year of Publication
Doctor of Philosophy (PhD)
Arts and Sciences
Dr. C. Melody Carswell
The objective of this study was to examine the use of the cognitive construct of “typicality” to guide design decisions in the development of consumer products. Increasing products that will appeal to consumers, designers strive to balance novelty and familiarity. A potential way to thread this needle is to understand how “typical” a design is of its particular product category. The construct of typicality has been used by psychologists to understand how people create and represent categories. Objects that are more typical of a category are often associated with positive responses from observes (e.g., greater visual appeal, faster recognition). In order to leverage the advantages of typicality in design, however, the construct must be effectively measured. Here, we specifically assess the validity and user acceptance of a single, numeric rating scale (the “legacy” typicality scale) that has been used extensively in cognitive psychology. We also consider whether the construct of typicality should be considered multidimensional rather than unidimensional. Study 1 adopted a criterion-related validity procedure in which typicality ratings from a group of 205 participants were used to predict the choices of products from an independent group of 170 participants who selected the products that they perceived to be the most and least visually appealing and prestigious. Additional unidimensional ratings were collected and used as predictors. These included “visual similarity” (to other members of the product class), “perceived usability,” and “familiarity.” In study 2, 97 subject matter experts (actual product designers and product researchers) responded to scenario-based surveys in which they made design decisions based on the data collected in Study 1. We assessed the quality of the experts’ decisions, as well as their preference for data type (unidimensional or multidimensional). Legacy typicality ratings did not reliably predict which members of a product category were perceived as most appealing or most prestigious. However, typicality ratings did predict which designs would be seen as least appealing and least prestigious. In general, less typical designs were less favorable evaluated. Ratings of additional, theoretically-related constructs did not combine to create a composite measure that was more predictive. A multidimensional scaling analysis of the typicality ratings did, however, reveal that the ratings had a 2D rather than 1D structure. In Study 2, experts strongly preferred using the data from the unidimensional legacy scales compare to the derived multidimensional results (shown as a 2D plot). The presence of typicality data did, in limited cases, improve the quality of experts’ design choices, but these effects depended on the product class being judged and the design objective (maximize visual appeal or maximize prestige). In general, the experts tended to be biased toward highly novel designs over highly typical ones. Low typicality appears to be related to low visual appeal and low prestige; however, high typicality did not ensure favorable reactions from participants. This suggests that typicality scores might be best used to define a minimum criterion which designers should strive to not let their products fall below. We obtained no evidence that a composite scale was a better predictor of design outcomes than the legacy typicality scale, and experts preferred the legacy scale over a multidimensional alternative. Experts’ application of typicality data to their design decisions were highly varied, with novelty seen by them as more important than typicality.
Digital Object Identifier (DOI)
Kent, Travis, "The Measurement of Product Typicality in Design Research. A Basic and Applied Approach" (2022). Theses and Dissertations--Psychology. 211.