Using Multiple (Imperfect) Methods to Test Alternative Views: The Case for Stable Preferences for Not/Compromising

New concepts that go against conventional assumptions are typically hard to “prove” using a single method. In such cases, it is more effective to rely on evidence that employs a number of different methods/data. That is, while changing fundamental assumptions is unlikely to be supported with any single study, a pattern of findings obtained using various approaches and domains can more persuasively advance the alternative view. This approach is illustrated in the context of a basic assumption underlying consumer decision making whereby attribute values are the carriers of utility whereas relative option positions in any given set are irrelevant. Specifically, an assumption that decision researchers have adopted from economic theory is that consumers should have stable utilities for each attribute value. Violations of this principle, such as that consumers prefer Option A over B but prefer B over A when C is added to set, were thus interpreted as evidence that preferences are “constructed” when decisions are made. More generally, decision researchers have assumed that preferences should be based on the options’ absolute values (e.g., 10X magnification, the felt comfort of headphones), and violations of this assumption represent the influence of context, task, and frame on constructed preferences.