Effects of contiguity and similarity on the learning of concepts.

Publisher Website
Google Scholar
Abstract
In this study the validity of the following hypotheses was tested: (1) the greater the temporal contiguity among representations of the same concept, the faster the learning of that concept, and (2) the greater the similarity among the relevant concept cues of different concepts, the slower the learning of these concepts. A total of 120 Ss was randomly assigned to 6 groups of 20 Ss each of which learned a practice and experimental list. The experimental list consisted of line and dot combinations. The azimuth of the line (3 azimuths) and the position of the dot (s) (high, middle, and low) were the concept determinants. The number of dots (1, 2, or 3) and the color of the line (red, green, or blue) were the distractors. Learning was by a modified paired-associate method. Three variations in similarity and 2 in contiguity were employed. Similarity was varied by changing the azimuth differences between the lines, contiguity by changing the cluster of the 3 concept representations in a group of 27 stimuli. Ss were given 45 trials on the experimental list. The following results were obtained: (1) statistically significant effects of contiguity in terms both of number of concepts learned and number of intrusions during learning were demonstrated. (2) No statistically significant effects of similarity were demonstrated either in terms of number of concepts learned or of number of within-list intrusions during learning. (3) The most within-list intrusions were made to stimuli of the slight slant concepts, then to those of the extreme slant, and vertical concepts, respectively. It was suggested that closer grouping of representations of the same concept led to a greater likelihood that responses made by S to the concept-revelant features of this concept would be temporally contiguous. Grouping would also lead to a decreased probability of interference from responses to other concept stimuli. The concept of stimulus generalization was employed in explaining the frequencies of errors involving the 3 different classes of concepts. Finally, it was suggested that the points on the similarity dimension may have been too close together to lead to differences in the rate of concept learning.
Keywords

This publication has 0 references indexed in Scilit: