対連合学習における意味的媒介要因の分析

抄録

Various factors have been found to determine the difficulty of paired-associate learning. The present study was designed to examine the relationships among these factors and the difficulty of learning by a factor-analytical method.
Forty common nouns were used and four experiments were conducted, as follows: (1) Paired-associate learning and retention: Ss learned two lists of 10 pairs of nouns for five trials and were tested by anticipation method after one week. (2) Free association test: Ss were given a written association task to all stimulus nouns. (3) Restricted association test: Ss were asked to give sensory impressions to the same nouns. (4) Semantic differential: Ss rated the connotative meaning of each noun on 20 semantic scales. Ss were 280 undergraduate students in all.
The centroid method of factor analysis and orthogonal rotation of axis were applied to the correlation matrix based on 15 measures for 20 paired-associates which were obtained from the four tests; four factors were extracted.
The variables positively loaded with Factor I were difficulty of learning (during acquisition), amount of retention, differences of rating scores between stimulus items (S) and response items (R), and associative overlap between S and R. Thus Factor I was interpreted as “similarity between S and R”. Factor II had high positive loadings on rating deviation and associative score of R, arid Factor III on rating deviation and associative score of S; hence these two factors were considered as “meaningfulness of either S or R”. Because of its high positive loadings on difficulty of learning, amount of retention, differences of rating scores between S and R, associative overlap between S and R, and associative overlap of sensory impressions between S and R, Factor IV was interpreted as “associative similarity between S and R”. Results showed that both associative and connotative similarities are major factors which mediate between S and R in paired-associate learning. Meaningfulness of either S or R had no significant relationship to difficulty of learning as well as to amount of retention in this study, but these findings are of course restricted within the present experimental setup and materials.