RECOGNITION ASSESSED BY RATING AND RANKING

Recognition tests are ubiquitous in the study of cognition. A simple index of recognition with desirable properties is the proportion R of targets preferred to distractors. The estimation of R by either a rating or a ranking method is straightforward. No distributional assumptions are involved. Three experiments are described in which estimates derived by the two methods were found to be close. The rating method is potentially subject to a bias. A preliminary account is given of calculations which suggest that the bias is small provided at least four rating categories are used.