Satisficing

Satisficing in Decision Making and Inductive Reasoning

Herbert Simon introduced the concept of satisficing—a melding of “satisfy” and “suffice”—to describe how people make decisions under real-world constraints. Rather than optimizing every choice, which would be impossibly time- and energy-consuming, they stop when they reach a solution that is good enough. Simon’s broader account of “bounded rationality” concerns the limits on information, time, and computational capacity. Satisficing is one way to make a choice within those limits. This corner-cutting practice has become a standard defense for departures from formal decision models, and it is often extended to explain why people fail to reason statistically—the assumption being that intuitive shortcuts are less laborious than normative statistical reasoning.

However, psychological research reveals two important complications. First, the satisficing defense may not transfer automatically to inductive reasoning. Statistical thinking, once learned, can be fast and even automatic—often less laborious than constructing elaborate causal accounts. For example, in a study by Krantz et al. (1983) on sports statistics, some participants persisted in generating strained causal explanations for phenomena like a “rookie of the year” playing worse the following season, while others quickly produced crisp statistical explanations (e.g., “the first year was just above their lifetime average”) that still acknowledged room for causal factors. This suggests that the effortfulness of a reasoning strategy depends on the thinker’s training and the immediate context, not on a simple satisficing vs. optimizing trade-off.

Second, subsequent work by Daniel Kahneman and Amos Tversky showed that intuitive judgment often deviates from rational models in qualitative ways, not merely as a rough approximation that stops early. Their introspection and experiments uncovered systematic biases and heuristics that cannot be explained solely by a bounded optimizing machine. Thus, while Simon’s satisficing captures one aspect of real-world rationality, it may understate the deeper, predictable departures from normative thinking that shape both decisions and inductive inferences.

Ecological Representativeness of Problems Showing Errors

A related criticism of the literature on errors in inductive reasoning is that these errors result from examining people’s judgment in particular kinds of problems and contexts where judgments are particularly fallible (e.g., Cohen, 1981; Dennett, 1981; Lopes, 1982). The criticism suggests that studies reveal more about experimenters' cleverness than about real-world failures. The authors offer several responses to this criticism:

• Some subjects showed an appreciation of statistical principles, suggesting differences in skill rather than deception.
• Factors that make statistical reasoning more or less likely (e.g., recognition of heterogeneity and the role of chance) are not factors that make people more dupable but rather factors that influence the obviousness of statistical reasoning.
• Statistical training markedly influences answers, suggesting that knowledge matters; this does not rule out an additional role for context.