Confirmation bias
Confirmation bias refers to tendencies to seek, interpret or remember evidence in ways that favour an existing hypothesis. The term covers several processes, including selectively gathering evidence and discounting contrary findings.[1]
Rule discovery
In Wason’s 1960 experiment, 29 psychology undergraduates saw the numbers 2, 4, 6 and tried to discover the rule by proposing further triples. The experimenter reported whether each triple satisfied the rule: any three numbers in increasing order. Six participants gave the correct rule at their first announcement. Repeatedly testing a narrower rule could produce agreement without revealing what it excluded.[2]
For example, consider the hypothesis consecutive even numbers in increasing order:
| Proposed triple | Experimenter’s answer | What it shows |
|---|---|---|
| 8, 10, 12 | Yes | Fits both the hypothesis and the actual rule. |
| 1, 3, 8 | Yes | Refutes the narrower hypothesis: these are not consecutive even numbers. |
| 6, 4, 2 | No | Both rules reject it, so this result does not distinguish them. |
The informative counterexample here is an accepted triple that the participant’s hypothesis rejects.[2]
Positive tests and confirmation
Klayman and Ha distinguish seeking confirmation from a positive test strategy: examining cases where a hypothesis predicts the property of interest, or where that property is already known to occur. Such a test can expose a mistaken prediction. Its value depends on the task, including how broad the hypothesis is relative to the actual phenomenon. It can be useful in some settings and misleading in others. In probabilistic settings, a single failed prediction need not decisively refute a hypothesis.[1]
See also
References
- ↑ 1.0 1.1 Joshua Klayman and Young-Won Ha, “Confirmation, Disconfirmation, and Information in Hypothesis Testing”, Psychological Review 94(2), 1987, pp. 211–228, especially pp. 211–213, 220 and 225; public text.
- ↑ 2.0 2.1 P. C. Wason, “On the failure to eliminate hypotheses in a conceptual task”, Quarterly Journal of Experimental Psychology 12(3), 1960, pp. 129–140, especially pp. 130–134 and 138–139; published paper.
