Signal detection theory
Also known as: Detection theory
Separating true sensitivity from the bias to say 'yes.'
What it means
Signal detection theory models how observers distinguish a meaningful signal from background noise when the two overlap, separating two independent quantities: sensitivity (d-prime), the true ability to tell signal from noise, and the response criterion, the decision threshold for responding 'yes.' By classifying responses as hits, misses, false alarms, and correct rejections, it disentangles perceptual acuity from decision bias, which earlier 'threshold' notions conflated. The criterion shifts with the payoffs and prior probabilities of the situation, explaining why the same observer becomes more or less cautious as stakes change. Originating in radar and engineering, it now pervades perception, memory recognition, medical diagnosis, and lie detection. It reframed detection as a decision under uncertainty rather than a fixed sensory limit.
Examples
A radiologist's miss rate depends both on visual sensitivity and on how readily they call an ambiguous shadow a tumor — two separable factors.
During a heightened security alert, airport screeners pull aside far more harmless bottles. Their eyesight has not changed; the cost of a miss has, so the threshold for 'suspicious' drops.
A witness told the culprit is definitely in the lineup picks someone more readily than one told he may not be there. Same memory, looser criterion, more false identifications.
First described in Green & Swets (1966); roots in Tanner & Swets (1954).