Is recognition accuracy really impaired when the target is repeated in the display?

Abstract

Bjork and Murray (1977) have presented a feature-specific interactive channels model that predicts perceptual interference between letters presented simultaneously in visual displays. Maximum interference is predicted when a target letter is presented with an identical letter. In their experiment, Bjork and Murray found support for their model, but their results could have occurred artifactually from response biases of their subjects. Santee and Egeth (1980) used a different paradigm that eliminated the possibility of this particular response bias and reported that their data supported the feature-specific model. However, the present paper shows that analysis of their data by Santee and Egeth was incomplete and the comparisons inappropriate. In two experiments that used the Santee and Egeth paradigm, we failed to find support for the feature-specific model when more detailed analyses of the data were undertaken. In a third experiment, one that used the Bjork and Murray paradigm, but with control of response bias, no significant or suggestive evidence was found that feature similarity between noise and target letters affected recognition of the latter.