Why Can Two Rare Events Seem Related?

After two events coincide a few times, it is easy to feel that they are connected. The impression can be especially strong when both events are uncommon. This is illusory correlation: an observer overestimates an association that is absent or weaker than it appears.

In Hamilton and Gifford's classic 1976 experiments, participants read about two fictional groups. The larger and smaller groups had the same proportions of positive and negative behaviours, while negative behaviours were less frequent overall. When a member of the smaller group performed a negative act, a rare group and a rare behaviour appeared together. That pairing attracted attention and was easier to remember. Participants could then judge the smaller group more negatively even though the presented proportions did not justify the difference.

The effect is related to confirmation bias but is not identical. Confirmation bias emphasises seeking evidence that fits an existing belief; illusory correlation can emerge from the joint distinctiveness of events even without a clear prior expectation. It also comes before the question of whether correlation is causal, because the first mistake may be believing that a correlation exists at all.

The research does not show that everyone makes the same error in every setting. Task design, expectations and memory processes can alter the effect. A useful check is to restore the missing denominator: compare how often common and rare behaviours occur in both groups, rather than counting only the memorable pairings. A striking coincidence is a clue, not proof of a relationship.

https://www.sciencedirect.com/science/article/pii/S0022103176800066
https://pmc.ncbi.nlm.nih.gov/articles/PMC7497493/