Illusory correlation: why two events happening close together make the brain think one causes the other

What is illusory correlation?

Illusory correlation is a psychological phenomenon in which a person perceives a link between two variables, usually people, events, or behaviors, even when no such link actually exists. This false link can form because rare or novel occurrences stand out more and therefore draw more of the observer's attention than everyday occurrences.

The term was coined by Loren Chapman in 1967, originally to describe people's tendency to overestimate the relationship between two groups when presented with distinctive and unusual information. The concept has also been used to question objectivity in clinical psychological diagnosis, through Chapman's rejection of many Wheeler signs that clinicians of the time widely used to identify homosexuality in the Rorschach test.

Why does the brain create illusory correlations?

Most explanations for illusory correlation involve psychological heuristics, the information-processing shortcuts that underlie many human judgments. One of these is availability: how easily an idea comes to mind. Availability is often used to estimate the likelihood or frequency of an event. This can lead to illusory correlation, because some pairs of events come to mind easily and vividly even though they do not occur often.

Martin Hilbert (2012) proposed an information-processing mechanism, assuming that the transformation from objective observation to subjective judgment always contains noise. He defined noise as the blending of observations during retrieval from memory. Under this model, subjective perceptions or judgments below are identical to noise or objective observations, which can lead to overconfidence or conservative bias: when asked about behavior, participants rate the majority or larger group lower and the minority or smaller group higher. These results are illusory correlation.

The role of working memory

In an experimental study by Eder, Fiedler, and Hamm-Eder (2011), the influence of working memory capacity on illusory correlation was examined. They first looked at individual differences in working memory, then tested whether this affected the formation of illusory correlation. They found that individuals with higher working memory capacity viewed minority group members more positively than those with lower working memory capacity. In the second experiment, the authors studied the effect of memory load in working memory on illusory correlation. They found that increased memory load in working memory led to a higher prevalence of illusory correlation. The experiment was designed specifically to test working memory rather than significant stimulus memory, meaning that the development of illusory correlation was caused by a shortage of central cognitive resources due to working memory load, not by selective retrieval.

The Hamilton and Gifford experiment of 1976

David Hamilton and Robert Gifford (1976) conducted a series of experiments showing how stereotyped beliefs about minority groups can originate from illusory correlation. To test their hypothesis, Hamilton and Gifford had participants read a series of sentences describing desirable or undesirable behaviors, assigned to Group A (the majority) or Group B (the minority). Abstract groups were used so that no pre-existing stereotypes would affect the results. Most sentences were tied to Group A, the rest to Group B.

Each group had the same proportion of positive and negative behaviors, so there was no real link between behavior and group membership. The results showed that positive, desirable behaviors were not seen as distinctive, so participants judged the link accurately. By contrast, when undesirable, distinctive behaviors appeared in the sentences, participants overestimated how much the minority group displayed those behaviors.

A parallel effect occurs when people judge whether two events, such as pain and bad weather, correlate with each other. They rely heavily on the relatively small number of cases in which the two events occurred together. People give relatively little attention to other kinds of observations, such as no pain or good weather.

How is illusory correlation related to stereotypes?

Illusory correlation is one of the ways stereotypes form and are maintained. Hamilton and Rose (1980) found that stereotypes can lead people to expect certain groups and traits to go together, and then to overestimate how often they co-occur. These stereotypes can be learned and passed on without any actual contact between the person holding the stereotype and the group it targets.

The attention theory of learning holds that features of the majority group are learned first, then features of the minority group. This leads to an effort to distinguish the minority from the majority, making these differences learned faster. The attention theory also argues that instead of forming a single stereotype about the minority, two stereotypes are formed: one for the majority and one for the minority.

When does illusory correlation form?

Johnson and Jacobs (2003) ran an experiment to see when in life people begin to form illusory correlation. Children in grades 2 and 5 were exposed to a typical illusory correlation model to see whether negative attributes were attached to the minority group. The authors found that both groups formed illusory correlation.

Another study also found that children produce illusory correlation. In their experiment, children in grades 1, 3, 5, and 7 and adults all viewed the same illusory correlation model. The study found that children did produce significant illusory correlation, but these correlations were weaker than in adults. In a second study, shape groups with different colors were used. Illusory correlation formation still continued, showing that social stimuli are not necessary to produce these correlations.

Explicit and implicit attitudes

Two studies by Ratliff and Nosek examined whether explicit and implicit attitudes affect illusory correlation. In one study, Ratliff and Nosek had two groups: a majority and a minority. They then had three groups of participants, all of whom read about these two groups. One group received readings overwhelmingly favoring the majority, one received readings favoring the minority, and one received neutral readings. The groups that received readings favoring the majority and the minority both favored their respective group both explicitly and implicitly. The group that received neutral readings favored the majority explicitly but not implicitly. The second study was similar, but instead of readings, images of behaviors were presented, and participants wrote a sentence describing the behavior they saw in the images. Findings from both studies supported the authors' argument that the difference between explicit and implicit attitudes results from interpreting covariation and making judgments based on these interpretations (explicit), rather than merely accounting for covariation (implicit).

Can the structure of the experimental model create illusory correlation?

Berndsen et al. (1999) wanted to determine whether the structure of testing for illusory correlation could lead to its formation. The hypothesis was that identifying the test variables as Group A and Group B could lead participants to look for differences between the groups, leading to illusory correlation. An experiment was set up in which one group of participants was told the groups were Group A and Group B, while another group was labeled 1993 or 1994 graduates. This study found that illusory correlation was more likely to be created when the groups were called Group A and B than when they were called the class of 1993 or the class of 1994.

Does learning reduce illusory correlation?

A study was conducted to investigate whether increased learning affects illusory correlation. Results showed that educating people about how illusory correlation occurs led to a lower rate of illusory correlation.

Apophenia and its link to illusory correlation

Apophenia is the tendency to perceive meaningful connections between unrelated things. The term (German: Apophänie, from the Greek verb ἀποφαίνειν) was coined by psychiatrist Klaus Conrad in his 1958 work on the onset stage of schizophrenia. He defined it as "the seeing of connections without motive, accompanied by a special feeling of abnormal meaningfulness." He described the early stages of delusional thinking as excessive self-referential interpretations of real sensory perceptions, as opposed to hallucinations.

Apophenia can be seen as an ordinary effect of brain function. However, when pushed to an extreme, it can be a symptom of mental disorder, for example in paranoid schizophrenia, when patients see hostile patterns (such as a plot to harm them) in ordinary actions. Apophenia is also typical in conspiracy theories, where coincidences can be woven into a clear plot.

Pareidolia

Pareidolia is a type of apophenia involving the perception of images or sounds in random stimuli. A common example is seeing a face in an inanimate object; car headlights and grilles can look like they are "smiling." People around the world see the "Man in the Moon." Sometimes people see the face of a religious figure in a slice of toast or in wood grain. There is strong evidence that hallucinogenic drugs tend to cause or enhance pareidolia. Pareidolia often occurs because the fusiform face area, the part of the brain responsible for seeing faces, misinterprets an object, shape, or configuration with some "face-like" features as a face.

Gambling

Gamblers may imagine they see patterns in numbers appearing in lotteries, card games, or roulette wheels, where no such patterns exist. A common example of this is the gambler's fallacy.

Statistics

In statistics, apophenia is an example of a Type I error, that is, falsely identifying patterns in data. It can be compared to so-called false positives in other testing situations.

Clustering illusion

The clustering illusion is a type of cognitive bias in which a person sees a pattern in a random sequence of numbers or events. Many theories have been rejected because this bias was highlighted. One case in the early 2000s involved the occurrence of breast cancer among employees of ABC Studios in Queensland. A study found that the rate of breast cancer at these studios was six times higher than in the rest of Queensland. An investigation found no correlation between the elevated rate and any location-related factor or any genetic or lifestyle factor of the employees.

Pattern recognition models and causes of apophenia

Although no confirmed reason exists for why apophenia occurs, several respected theories exist. Pattern recognition is a cognitive process involving retrieving information from long-term, short-term, or working memory and matching it with information from stimuli. There are three different ways this can happen and go wrong, leading to apophenia.

Template matching

The stimulus is compared with templates, abstract or partial representations of previously seen stimuli. These templates are stored in long-term memory as a result of past learning or educational experience. For example, D, d, D, d, D, and d are all recognized as the same letter. Template-matching detection processes, when applied to more complex data sets (such as a picture or data clusters), can lead to false template matching. A false-positive finding would lead to apophenia.

Prototype matching

This is similar to template matching, except that prototypes are complete representations of a stimulus. A prototype is not necessarily something seen before; for example, it can be an average or blend of previous stimuli. Importantly, an exact match is not required. An example of prototype matching is that when looking at an animal such as a tiger, instead of recognizing that it has features matching the definition of a tiger (template matching), we recognize that it is similar to a specific mental image of a tiger (prototype matching). This type of pattern recognition can lead to apophenia because the brain is not looking for an exact match; it can grasp some features of a match and assume it fits.

Feature analysis

The stimulus is first broken down into its features and then processed. This pattern recognition model holds that processing goes through four stages: detection, pattern dissection, feature comparison in memory, and recognition.

Evolutionary perspective

One explanation offered by evolutionary psychologists for apophenia is that it is not a defect in human brain cognition but something that emerged over years of evolution. This field of research is called error management theory. One of the most recognized studies in this field is Skinner's box. This experiment involved taking a hungry pigeon, placing it in a box, and releasing food pellets at random times. The pigeon received a pellet while performing some action, and instead of attributing the pellet's arrival to chance, the pigeon repeated that action and kept doing so until another pellet dropped. As the pigeon increased how often it performed the action, it gained the impression that it also increased how often it was "rewarded" with a pellet, even though the food release was in fact still completely random.

Practical applications and ways to reduce it

Illusory correlation and apophenia can affect many areas of life, from medical diagnosis and financial decision-making to judging people at work and in social life. When a person believes two events have a causal link just because they happened close together in time, they can make decisions based on false evidence.

One study showed that educating people about how illusory correlation occurs led to a lower rate of the phenomenon. This suggests that awareness of the bias is the first step to reducing its influence. In addition, collecting complete data and careful statistical testing can help distinguish between a real correlation and one that exists only in perception.

Note on different perspectives

It should be noted that in psychology, there are several different approaches to explaining illusory correlation. Some researchers emphasize the role of availability and information-processing heuristics, while others focus on noise in information processing or working memory capacity. These theories are not mutually exclusive but complementary, each highlighting a different aspect of this complex phenomenon. This article presents the main perspectives based on available references, and readers should consult multiple sources for a comprehensive view.