Why Do We Seek Supporting Evidence Once We Have a Hypothesis?

Once we form a hypothesis, we readily ask which facts fit it rather than what result would show it to be wrong. In Wason’s classic rule-discovery experiment, participants were given the sequence “2, 4, 6”, which followed an unknown rule, and proposed new sequences while trying to identify that rule. Many first guessed “add two” and tested “8, 10, 12”. A positive answer increased their confidence, although the actual rule could simply be “any three increasing numbers”.

The problem is not that supporting evidence has no value. It is that the test lacks discrimination. If both “add two” and “any increasing sequence” predict the same example, a positive answer cannot choose between them. A test such as “6, 5, 4”, on which the two hypotheses predict different answers, carries more information.

This tendency is often called confirmation bias, but its limits matter. People are not always consciously defending a preferred belief. Often they are using a positive test strategy: searching first for cases that match the category they have in mind. In environments where the target event is rare or search is costly, this can even be efficient. It becomes misleading when easy support is treated as a severe test, or when we never design a condition under which our idea could fail.

Better judgement therefore requires more than forcing ourselves to read one opposing opinion. A sharper question is: if this hypothesis were false, what should I observe? The most useful evidence often makes competing explanations predict different outcomes.

https://doi.org/10.1080/17470216008416717
https://doi.org/10.1037/0033-295X.94.2.211