Comparison (Two-Distribution) Designs for Discrimination

In comparison designs for discrimination, each trial contains two stimulus sequences. In two-alternative forced-choice (2AFC), the two stimulus classes are presented in either of two orders. Performance is expected to be better than in the yes-no design. In the yes-no reminder design, a yes-no interval containing one of the two stimulus classes is preceded by a constant “reminder,” and performance is expected to be worse than yes-no if the observer compares the two intervals. In 2AFC with reminder (2AFCR), the last two stimuli are always different, while the first stimulus, the reminder, is always the same. The likelihood-ratio and difference decision strategies result in the same performance as the 2AFC task. However, application of the absolute-difference strategy leads to reduced performance. Accuracy in 2AFC depends on the interstimulus interval and the range of stimuli. Long intervals and wide ranges reduce performance. Yes-no accuracy tends to be lower, relative to 2AFC, than detection theory predicts.

Distribution (mathematics) · Mathematical analysis · Computer Science · Mathematics · Optimal Experimental Design Methods