🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
illusory_correlation
The perception of a relationship between two variables when no such relationship exists, or the overestimation of the strength of a weak relationship. People are particularly prone to seeing correlations between distinctive events, even when they co-occur no more often than chance would predict. This bias is a key mechanism in stereotype formation.
A person who has had two negative experiences with a specific car brand concludes that the brand is unreliable, while ignoring the many uneventful experiences and not comparing the rate to other brands. The negative events are memorable and create an illusory association.
An office manager notices that the coffee machine breaks down on the same days that a particular remote employee works from the office, and starts privately wondering if that employee somehow causes the machine to malfunction — ignoring dozens of other variables and the random nature of appliance failures.
A parent notices that their child seems to misbehave more on evenings after eating red candy and becomes convinced that artificial dyes cause hyperactivity. They vividly remember the confirming incidents and forget the many times the child ate red candy and behaved perfectly normally.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Is a causal or correlational link claimed without statistical evidence?
Type: binaryAre co-occurrences of two events being noticed while non-co-occurrences are ignored?
Type: binaryWould controlled data support the perceived relationship?
Type: binaryThe perception of a relationship between two variables when no such relationship exists, or the overestimation of the strength of a weak relationship. People are particularly prone to seeing correlations between distinctive events, even when they co-occur no more often than chance would predict. This bias is a key mechanism in stereotype formation.
Distinctive or emotionally salient co-occurrences are more memorable and easier to recall. Confirmation bias then reinforces these associations by selectively attending to confirming instances while ignoring disconfirming ones.
Look at actual data rather than relying on memory. Create a simple 2x2 contingency table to check whether the variables are genuinely associated rather than relying on anecdotal recall.
Illusory correlation sustains stereotypes about minority groups, superstitious beliefs, and alternative medicine claims. It also affects clinical judgment when practitioners associate symptoms with diagnoses based on memorable cases.
The tendency for new information to be pulled toward and assimilated into dominant existing narratives, distorting its interpretation to fit pre-existing stories.
Confusing selection factors with results. The assumption that swimmers have athletic bodies because of swimming, when in reality people with certain body types are selected for (or gravitate toward) competitive swimming. A specific manifestation of the broader confusion between selection effects and causal effects.
Use these tools to detect, analyze, or train this aspect.