Heuristic assumptions

Interpreting any decision requires making auxiliary assumptions regarding how the decision makers viewed their options and related them to their lives. Theories constrain those assumptions. The more general the theory, the fewer constraints it imposes and the more assumptions must be made in any application. Like the rational actor models that they challenged, Daniel Kahneman and Amos Tversky’s heuristics-and-biases and prospect theory research programs were general theories, with broad application. One of the many marvels of their landmark studies is that they rarely test their auxiliary assumptions. Rather, readers were trusted to agree about how people interpret the tasks (e.g., select anchors in studies of that heuristic). Subsequent studies have often accepted those interpretations in order to examine boundary conditions (e.g., extreme anchors). Applying the theories to naturally occurring tasks requires making additional auxiliary assumptions. This article illustrates three ways to evaluate those assumptions: direct assessment, systematic manipulation, and archival analysis. It concludes with proposals for loosely coordinated evaluation of shared and contested assumptions.