The Ecological Rationality of Simple Group Heuristics: Effects of Group Member Strategies on Decision Accuracy

Abstract

The notion of ecological rationality implies that the accuracy of a decision strategy depends on features of the information environment in which it is tested. We demonstrate that the performance of a group may be strongly affected by the decision strategies used by its individual members and specify how this effect is moderated by environmental features. Specifically, in a set of simulation studies, we systematically compared four decision strategies used by the individual group members: two linear, compensatory decision strategies and two simple, noncompensatory heuristics. Individual decisions were aggregated by using a majority rule. To assess the ecological rationality of the strategies, we varied (a) the distribution of cue validities, (b) the quantity, and (c) the quality of shared information. Group performance strongly depended on the distribution of cue validities. When validities were linearly distributed, groups using a compensatory strategy achieved the highest accuracy. Conversely, when cue validities followed a J-shaped distribution, groups using a simple lexicographic heuristic performed best. While these effects were robust across different quantities of shared information, the quality of shared information exerted stronger effects on group performance. Consequences for prescriptive theories on group decision making are discussed Copyright Springer 2006compensatory and noncompensatory decision strategies, group decision making, group performance, simple heuristics,

Fast and Frugal Heuristics: Tools of Social Rationality

“…Analytically, Reimer and Katsikopoulos showed that when the (average) recognition validity α is larger than the (average) knowledge validity β, then a group that recognizes fewer objects (movies) will perform better than a group that recognizes more. Judging from Reimer and Katsikopoulos' results, group members seem to grasp and trust the recognition heuristic and the wisdom encapsulated in partial ignorance (see Reimer & Hoffrage, 2006, for more on simple group heuristics).…”

Naïve Groups Can Solve the Hidden-Profile Problem

“…This characteristic does not hold for all common-cue environments. It is possible to construct common-cue environments in which the best choice alternative can only be found on the basis of more complex decision strategies, such as weighted additive models (see, e.g., Reimer & Hoffrage, 2006). Thus, future studies may address whether the observed differences between unique-and common-cue environments can be generalized across information environments that demand more complex decision strategies.…”

Selecting decision strategies: The differential role of affect

“…The weights used for the implementation of the compensatory WADD model were obtained by rescaling the validities presented on the screen into Goodman-Kruskal validities (i.e., weight = 2 × validity − 1) to control for chance (Reimer & Hoffrage, 2006). The movie pairs were selected such that a simple non-compensatory LEX strategy made opposite predictions to a WADD or EQW strategy (the latter two made similar predictions).…”