Judgment under uncertainty: heuristics and biases

Judgment Under Uncertainty: Heuristics and Biases

Tversky and Kahneman's article Judgment under uncertainty: Heuristics and Biases, was published in 1974 and it occupies a pivotal place in the literature of decision making, judgment process. The core premise of this article revolves around the argument that people are not influenced as much by factual probability data as they are by cognitive heuristics when making a decision or judgment. "In general, these heuristics are quite useful, but sometimes they lead to severe and systematic errors"

The authors maintain that when a person is making a decision about something, he is likely to assign them probabilities. In the process, he will eliminate those possibilities, which have a zero probability. Once they are removed, the rest of the possibilities will be arranged according to a sub-process, which is based on three important heuristics. These heuristics are defined as availability, representative-ness and anchoring. Availability is the heuristic device that is used to assess the frequency and possibility of some event. In this case frequency is connected with the number of times this event occurred in the past and possibility is connected with likelihood of it happening in the future. A person living in Australia doesn't need to check any data or statistics to know that snow is more likely in June than in December. Similarly a person living in America would for example think just the opposite because of the number of times it has happened in the past.

Availability is a useful device and the authors are absolutely correct in including this in list of useful heuristics that people commonly rely on. But authors feel that heuristics can result in systemic errors because of biases that influence them. In the case of availability for example, Tversky and Kahneman 1974 argue that there are six biases that would affect recollection of events. We understand that availability is heavily depended on how accurately a person is capable of recalling an event therefore it is not entirely wrong to assume that recollection may not always be precise and accurate since some biases might come into play. The six biases identified by the authors are:

1.How familiar a person is with an event raises the probability of it being recollected with greater ease. The more familiar an event, the easier it is to recall it. A person may have seen snow in his city all his life but he may have heard of snow in New York and thus while he can judgment more accurately in his own case, he may not be able to predict snow with same accuracy in NY.

2. The magnitude of an event also bears on a person's ability to recall. The greater or more dramatic an event was, the sharper its image in the memory is.

3. The duration between now and the event. If the event were fairly recent, it would be recalled with more ease than an event that happened years ago.

4. The search sets created for this purpose may not always be effective. Recollection is easier when search sets are effective. A person can easily recall how often snowfall was witnessed if it happens each year in December instead of if it happens sporadically any time throughout the year.

5. Scenarios in which the events took place.

6. Illusory connection in agent's mind.

Authors argue that these biases are common and can affect an agent's judgment. The second important heuristic is representative-ness. This is a process where two objects or events are found to be interrelated or dependent on each other. If a person feels that a is connected with B. Or B. had given rise to a, they will make judgment in that manner. The simple belief that operates here is that as long as two things are closely connected or resemble each other a lot, there is a likelihood of them taking place together. For example a black person may be associated with a number of events. Or if a robbery occurred and among the suspects was a black person, what is the likelihood of people accusing him of the offence. Very high, we would say. This is because of availability and stereotypical connection between people of minorities and crime. In this heuristic again, five kinds of errors or biases can emerge:

1. Some errors are purely theoretical in nature. The two events may not be as related to each other as were initially assumed. For example all Muslims with beard are not fanatic or extremists but there is a likelihood that these two would go together in a person's mind.

2. People with show insensitivity to previous results. They may fail to take into account prior probability outcomes and instead rely on their own judgment based on representative-ness.

3. No attention to the size of the group examined. In other words, people tend to foget about sample sizes. They feel that a result that was true for a group of 50 would also be true for a group of 500 and this can lead to errors in judgment.

4. Inability to understand the role of chance. Tversky and Kahneman maintain that, "people expect that a sequence of events generated by a random process will represent the essential characteristics of that process even when the sequence is short"

5. Misconceptions about regression and their role in the judgment process. People may often be so heavily influenced by representative-ness that they may forget law of regression completely. Hence people "do not expect regression in many contexts where it is bound to occur... (and) when they recognize the occurrence of regression, they often invent spurious causal explanations for it"

The third heuristic of anchoring comes into play when adjustment or estimates are being calculated. When a person is to depend on calculation of probabilities or assigning of the same, they may not move very far from the starting point of calculation. This is called anchoring where people anchor their estimates to the starting point, which is always bigger to them than the latter. Interestingly the authors with the help of two sequences have illustrated this:

8 x 7 x 6 x 5 x 4 x 3 x 2 x 1 or

1 x 2 x 3 x 4 x 5 x 6 x 7 x 8