bias

By /

In decision science, bias refers to systematic and predictable deviations from accepted models of rationality.

In scientific language, “systematic” and “predictable” are not meant to imply that everyone exhibits the bias (or even that most people do, though that tends to be the conclusion), but rather just that they occur more than probability theory would predict if there were random errors around the (rational) mean.

Biases can be inbuilt or learned and they can be driven by cognitive, emotional, or social factors. In decision science, however, the focus tends to be on more universal, inbuilt, cognitive biases. Examples include loss aversion (where losses are felt more intensely than equivalent gains), anchoring (giving disproportionate weight to initial information), and confirmation bias (favoring information that confirms pre-existing beliefs).

The term is not inherently pejorative, since one could recommend the use of a heuristic, for example, while recognizing it leads to systematic deviations from accepted models of rational choice. See glossary entries for bounded rationality or fast and frugal heuristics for more on this idea.

That said, the term bias tends to be pejorative even among decision scientists, for a few distinct reasons.

First, just as rational and normative are already widely used in the English language outside of decision science and they imply good thinking, terms like irrational, non-normative, and biased have inherently pejorative connotations in the English language, regardless of whether decision scientists mean to imply anything pejorative with the terms. See the wikipedia entry on the term bias, for example, already biased (pun intended) to emphasize more scientific uses of the term. It nonetheless recognizes the pejorative meaning, which is common in social scientific uses of the word. This is a good example of where scientific definitions can become disconnected from the common-sense meaning of words).

Second, decision scientists often use the term—along with rational and irrational—interchangeably with the more widely used, pejorative sense of the term, implying the decision makers had sufficient information available such that “they should have known better.”

Third, since the normative models of rational choice are often taken by decision scientists to correspond to actual rational choice (rather than just being simplifying models that may be based on unrealistic or false assumptions), the pejorative sense of the term is often seen as factually interchangeable with the more scientific sense of the term, without much careful deliberation as to whether or when that might be true or false.

Two commonly studied biases in the casino gambling domain are the gambler’s fallacy and the hot hand cognitive illusion, although there are several other putative biases that have commonly been observed in the casino gambling context.

The terms cognitive illusion and fallacy are often used synonymously with bias, and they are sometimes used to get around the more pejorative meaning inherent to the term bias. Cognitive illusion is used to imply the parallel with optical illusions, since we generally accept that optical illusions result from how the visual system has been adaptively designed. Optical illusions are recognized to be a result of the unusual (biased?) environment and to rely on processes that usually work well except in those usually-artificial circumstances. There is not a tendency to frame them as biased or irrational. The term fallacy is another synonym that retains much of the pejorative but that suggests the bias results from the incorrect application of logic or probability. Thus, “belief in the hot hand” (when it is false or exaggerated) is often called “the hot hand cognitive illusion,” “the hot hand fallacy,” or “the hot hand bias.”

See the wikipedia entry on cognitive bias for more about the concept. See the glossary entry on heuristics and biases for more about the inherent relationship between heuristics and biases conveyed by many decision scientists.

Scroll to Top