A measure of stimulus similarity and errors in some paired-associate learning tasks.

Publisher Website
Google Scholar
Abstract
Similarity among aural signals of the International Morse was determined by use of the method of constant stimuli with 2 response categories (same: different). 598 human Ss, without prior code training, judged the 1296 two-signal permutations which may be formed from the 36 common Morse signals. All signals were presented at high tone speed. The measure of stimulus similarity provided by this method was evaluated in terms of substitution error data. It was found that the number of substitution errors made to a stimulus is an increasing function of the average similarity of that stimulus to the other stimuli in the paired-associate list. This function for the greater part of the observed similarity range is linear. The number of substitution errors made to a stimulus is also a decreasing function of the relative number of times that stimulus is judged "same" when presented twice in rapid succession.
Keywords

This publication has 1 reference indexed in Scilit: