Tipping the scale: The role of discriminability in conjoint analysis

Conjoint analysis is a widely used method for determining how much certain attributes matter to consumers by observing a series of their choices. However, how those attributes are expressed has important consequences for their choices and thus for conclusions drawn by market researchers about attribute importance. Expanded attribute scales (e.g., expressing exercise time in minutes) leads consumers to perceive greater differences between scale levels than contracted scales (e.g., expressing exercise time in hours). The authors show in two domains that simply expanding an attribute's scale can shift choice toward alternatives that perform well on a scale that is expanded and thus can impact conjoint results such as attribute importance and screening. Thus, practitioners should take care when they choose precisely how to elicit preferences or how to describe their products: the extent of the scale's expansion will determine researchers' inferences about the importance of the attribute it describes. By illustrating the curvilinear relationship between scale expansion and multiple measures, the authors also offer practitioners some insight into the limits of scale expansion.

Duke Scholars

Author Richard P. Larrick Fuqua School of Business