PsychologicalDUAL-USE

Illusory Correlation

What it is

Perceiving a relationship between two things that are not actually related, or perceiving a much stronger relationship than exists — typically because both items are distinctive or because the relationship was expected.

How it works

Loren Chapman (1967) found that people shown word pairs later overestimated how often distinctive pairs — long words, or semantically linked ones like bacon and eggs — had appeared together. With Jean Chapman he then showed the effect in professionals: clinicians believed that suspicious patients drew large eyes in a projective drawing test, and naive observers “rediscovered” the same associations from data in which they were absent or even reversed. Hamilton and Gifford (1976) showed that a minority group paired with a rare behavior is judged to display that behavior more often even when the proportions are identical — a mechanism by which stereotypes are formed from nothing. The lever is memory: co-occurrences of two unusual things are encoded vividly, while the far more numerous non-occurrences that would disconfirm the link are not encoded at all. Persuaders exploit this by pairing the thing they want you to fear or buy with a salient outcome and letting your memory do the counting. Crime coverage that names the perpetrator's group, testimonials that pair a product with a recovery, and chart patterns that pair a signal with a rally all work the same way.

Real-world examples

  • •Chapman and Chapman (1967): observers shown Draw-a-Person drawings paired at random with symptom descriptions still reported that suspicious patients drew prominent eyes — the association clinicians believed in but that the data did not contain.
  • •Hamilton and Gifford (1976): a smaller group described with fewer statements was judged to have more undesirable behavior than a larger group with the identical proportion of undesirable acts.
  • •Redelmeier and Tversky (1996) followed arthritis patients for over a year; nearly all believed the weather affected their pain, but the actual correlations were close to zero.
  • •News coverage that foregrounds the ethnicity or immigration status of a perpetrator pairs a distinctive group with a distinctive event; whatever the true rates, the pairing is what memory retains, and the same mechanism works for any minority a political actor wishes to make feared.
  • •Supplement marketing built on testimonials — I took it and the cold went away — presents the product-recovery pairing and never the recoveries that occurred without the product.

Ethical guidelines

Where the line is

Reporting a measured association with its data is honest; the line is crossed when a persuader pairs a group, product, or signal with a vivid outcome so that the audience infers a relationship the persuader has not shown and often knows is absent.

  • ●Before asserting a relationship, count all four cells: with and without the supposed cause, with and without the outcome.
  • ●Do not foreground a group identity in reporting an event unless it is causally relevant; distinctiveness pairing is how stereotypes are manufactured.
  • ●Testimonials that pair a product with an outcome must be accompanied by the rate of that outcome without the product.
  • ●Pointing out a real correlation, with the data, is legitimate; inviting the audience to infer one from a memorable pairing is not.

How to defend against it

  • ►Draw the 2 by 2 table. Ask how often the outcome occurs without the supposed cause; if nobody has that number, the correlation is a feeling.
  • ►Be most suspicious when both things are rare or vivid — an unusual group and an unusual event — because that is exactly when memory over-counts.
  • ►Ask for the base rate of the outcome in the general population before accepting that it is elevated in the highlighted group.
  • ►Keep a written record when you suspect a pattern in your own life (pain and weather, a food and a symptom); recorded data usually disagree with remembered data.
  • ►When a pattern is presented as a chart or list of hits, ask to see the misses.

From the Defense Playbook

Consider the Oppositeminutes

Deliberately generate the reasons your current judgment might be wrong before acting on it, which measurably reduces confirmation-driven error in a way that merely trying to be fair does not.

The Base-Rate Checkminutes

Before accepting what a test result, a profile, or a striking detail seems to prove, ask how common the thing is in the first place, because evidence that sounds accurate can still be wrong most of the time when the underlying condition is rare.

Every playbook entry states how strong its evidence is and when not to use it. Browse the full playbook.

References

  1. Chapman, L. J. (1967). Illusory correlation in observational report. Journal of Verbal Learning and Verbal Behavior, 6(1), 151-155 · link
    The founding word-pair demonstration and the naming of the effect.
  2. Chapman, L. J., & Chapman, J. P. (1967). Genesis of popular but erroneous psychodiagnostic observations. Journal of Abnormal Psychology, 72(3), 193-204 · link
    The Draw-a-Person study showing expectancy-based illusory correlation in clinicians and observers.
  3. Hamilton, D. L., & Gifford, R. K. (1976). Illusory correlation in interpersonal perception: A cognitive basis of stereotypic judgments. Journal of Experimental Social Psychology, 12(4), 392-407 · link
    The minority-group paradigm showing stereotype formation from distinctiveness alone.
  4. Redelmeier, D. A., & Tversky, A. (1996). On the belief that arthritis pain is related to the weather. Proceedings of the National Academy of Sciences, 93(7), 2895-2896
    The arthritis-weather study cited as a real-world example.
Last reviewed
Suggest a correction

Detect Illusory Correlation in any text

Paste any message, email, or article into our free Manipulation Detector to see if Illusory Correlation or other techniques are being used on you.

Try the Detector FreeCreate Free Account