Heuristic (psychology)

Heuristics is the process by which humans use mental shortcuts to arrive at decisions. Heuristics are simple strategies that humans, animals, organizations, and even machines use to quickly form judgments, make decisions, and find solutions to complex problems. Often this involves focusing on the most relevant aspects of a problem or situation to formulate a solution. While heuristic processes are used to find the answers and solutions that are most likely to work or be correct, they are not always right or the most accurate. Judgments and decisions based on heuristics are simply good enough to satisfy a pressing need in situations of uncertainty, where information is incomplete. In that sense they can differ from answers given by logic and probability.
The economist and cognitive psychologist Herbert A. Simon introduced the concept of heuristics in the 1950s, suggesting there were limitations to rational decision making. In the 1970s, psychologists Amos Tversky and Daniel Kahneman added to the field with their research on cognitive bias. It was their work that introduced specific heuristic models, a field which has only expanded since. While some argue that pure laziness is behind the heuristics process, this could just be a simplified explanation for why people don't act the way we expected them to. Other theories argue that it can be more accurate than decisions based on every known factor and consequence, such as the less-is-more effect.

History

formulated one of the first models of heuristics, known as satisficing. His more general research program posed the question of how humans make decisions when the conditions for rational choice theory are not met, that is how people decide under uncertainty. Simon is also known as the father of bounded rationality, which he understood as the study of the match between heuristics and decision environments. This program was later extended into the study of ecological rationality.
In the early 1970s, psychologists Amos Tversky and Daniel Kahneman took a different approach, linking heuristics to cognitive biases. Their typical experimental setup consisted of a rule of logic or probability, embedded in a verbal description of a judgement problem, and demonstrated that people's intuitive judgement deviated from the rule. The "Linda problem" [|below] gives an [|example]. The deviation is then explained by a heuristic. This research, called the heuristics-and-biases program, challenged the idea that human beings are rational actors and first gained worldwide attention in 1974 with the Science paper "Judgment Under Uncertainty: Heuristics and Biases" and although the originally proposed heuristics have been refined over time, this research program has changed the field by permanently setting the research questions.
The original ideas by Herbert Simon were taken up in the 1990s by Gerd Gigerenzer and others. According to their perspective, the study of heuristics requires formal models that allow predictions of behavior to be made ex ante. Their program has three aspects:

What are the heuristics humans use?
Under what conditions should humans rely on a given heuristic?
How to design heuristic decision aids that are easy to understand and execute?

Among others, this program has shown that heuristics can lead to fast, frugal, and accurate decisions in many real-world situations that are characterized by uncertainty.
These two different research programs have led to two kinds of models of heuristics, formal models and informal ones. Formal models describe the decision process in terms of an algorithm, which allows for mathematical proofs and computer simulations. In contrast, informal models are verbal descriptions.

Formal models of heuristics

List of formal models of heuristics:

Herbert Simon's satisficing heuristic can be used to choose one alternative from a set of alternatives in situations of uncertainty. Here, uncertainty means that the total set of alternatives and their consequences is not known or knowable. For instance, professional real-estate entrepreneurs rely on satisficing to decide in which location to invest to develop new commercial areas: "If I believe I can get at least x return within y years, then I take the option." In general, satisficing is defined as:

Step 1: Set an aspiration level α
Step 2: Choose the first alternative that satisfies α

If no alternative is found, then the aspiration level can be adapted.

Step 3: If after time β no alternative has satisfied α, then decrease α by some amount δ and return to step 1.

Satisficing has been reported across many domains, for instance as a heuristic car dealers use to price used BMWs.

Elimination by aspects

Unlike satisficing, Amos Tversky's elimination-by-aspect heuristic can be used when all alternatives are simultaneously available. The decision-maker gradually reduces the number of alternatives by eliminating alternatives that do not meet the aspiration level of a specific attribute. During a series of selections, people tend to experience uncertainty and exhibit inconsistency. Elimination by aspects could be used when facing selections. In general, the process of elimination by aspects is as follows:

Step 1: Select one attribute related to decision making
Step 2: Eliminate all alternatives that exclude this specific attribute
Step 3: Use another attribute in order to further eliminate alternatives
Step 4: Repeat step 3 until only one option is left, a decision has then been made

Elimination by aspects does not speculate that choosing alternatives could help consumers to maximize utility, on the contrary, it holds that selection is the result of a probabilistic process that gradually eliminates alternatives. A simple example is given by Amos Tversky: when someone wants to purchase a new car, the first aspect they will take into account might be the automatic transmission, this will eliminate all alternatives that do not contain such an aspect. Then, when all the alternatives that do not have this feature are eliminated, another aspect will be given such as a $3000 price limit. The process of elimination continues to occur until all alternatives are eliminated.
Elimination by aspects is well used in the early stage of business angels' decision-making process since it facilitates a fast-decision-making tool - alternatives will be eliminated when investors find a critical defect of the potential opportunities. Another research also demonstrated that elimination by aspects has widely been used in electricity contract choice. The logic behind these two examples is that elimination by aspects helps to make decisions when facing a series of complicated choices. One may need to make a decision among all alternatives while he or she only has limited intuitive computational facilities and time. However, elimination by aspects as a compensatory model could help to make such complex decisions since it is easier to apply and involves nonnumerical computations.

Recognition heuristic

The recognition heuristic exploits the basic psychological capacity for recognition in order to make inferences about unknown quantities in the world. For two alternatives, the heuristic is:

If one of two alternatives is recognized and the other not, then infer that the recognized alternative has the higher value with respect to the criterion.

For example, in the 2003 Wimbledon tennis tournament, Andy Roddick played Tommy Robredo. If one has heard of Roddick but not of Robredo, the recognition heuristic leads to the prediction that Roddick will win. The recognition heuristic exploits partial ignorance, if one has heard of both or no player, a different strategy is needed. Studies of Wimbledon 2003 and 2005 have shown that the recognition heuristic applied by semi-ignorant amateur players predicted the outcomes of all gentlemen single games as well and better than the seedings of the Wimbledon experts, as well as the ATP rankings. The recognition heuristic is ecologically rational when the recognition validity is substantially above chance. In the present case, recognition of players' names is highly correlated with their chances of winning.

Take-the-best

The take-the-best heuristic exploits the basic psychological capacity for retrieving cues from memory in the order of their validity. Based on the cue values, it infers which of two alternatives has a higher value on a criterion. Unlike the recognition heuristic, it requires that all alternatives are recognized, and it thus can be applied when the recognition heuristic cannot. For binary cues, the heuristic is defined as:
The validity v_i of a cue i is defined as the proportion of correct decisions c_i:
v_i = c_i / t_i
where ti is the number of cases the values of the two alternatives differ on cue i. The validity of each cue can be estimated from samples of observation.
Take-the-best has remarkable properties. In comparison with complex machine learning models, it has been shown that it can often predict better than regression models, classification-and-regression trees, neural networks, and support vector machines.
Similarly, psychological studies have shown that in situations where take-the-best is ecologically rational, a large proportion of people tend to rely on it. This includes decision making by airport custom officers, professional burglars and police officers and student populations. The conditions under which take-the-best is ecologically rational are mostly known. Take-the-best shows that the previous view that ignoring part of the information would be generally irrational is incorrect. Less can be more.