confusability

Confusability is the likelihood that one stimulus, symbol, word, or memory item is mistaken for another because of similarity in appearance, sound, or meaning. It is a common source of errors in perception, language processing, memory tasks, and everyday communication.

In perception, confusability arises when visually similar glyphs (such as the letters I, l, and 1; O and 0) or closely related phonemes (for example p and b; t and d) are misidentified. In memory and language processing, items with overlapping features can be confused with one another, contributing to intrusion errors or misattribution.

Confusability is measured with methods such as confusion matrices, which record correct identifications and misclassifications, and from these derive error rates. In psychology, signal detection theory provides discriminability measures (d') and decision criteria to quantify how easily stimuli are distinguished under uncertainty.

Applications include minimizing confusability in typography, icons, and user interfaces by using distinct shapes, colors, and spacing. In branding, reducing brand confusability helps prevent trademark confusion. In data processing and machine learning, reducing confusability through feature design and preprocessing improves classification and recognition accuracy.

See also: confusion, perception, signal detection theory.