Group paired-associate learning: applicability to the study of mediated generalization.

Abstract

Es interested in verbal learning have been somewhat handicapped by the inefficiency of their procedures. In the traditional paired-associate anticipation learning procedure only one S may be tested at a time, requiring that a large amount of E's time must be devoted to the collection of data. Where several experimental treaunencs are involved, several hundred hours of E's time may be necessary before any results are available. The result is that an experimental program is frequently limited by the amount of time available as well as by the quantity and quality of E's hypotheses. The inefficiency of the traditional paired-associate procedures has led a number of Es to seek ways in which more than one S can be tested at a time while maintaining comparability with the findings to be expected from an individual testing procedure. Such a procedure has recently been reported by Cieutat (1960, 1961, 1962). Each S in the group was given a booklet containing the material to be learned. The correct pairings of items were llsted at the top of each page, and below them were two columns of items. In each column the stimulus items were presented in random order at the left, and with each stimulus were presented four of the response alternatives. S's cask was to circle the correct response for each stimulus, and the measure of learning was the total number of items circled in successive 1-min. intervals. Cieutat (1960) has shown that when lists of pairs differing in meaningfulness are used, the difference between groups in number of items circled in - successive minutes increases, indicating that high meaningfulness pairs are learned more rapidly than low meaningfulness pairs. Since this result was consistent with previous findings on the effects of meaningfulness obtained with individual paired-associate learning procedures (see Underwood & Schulz, 1960), Cieutat inferred that che group procedure is comparable to the individual procedure. A further experiment by Cieutat ( 1961 ) supported this conclusion, again showing that when stimulus and response meaningfulness are varied simultaneously learning is faster for high meaningfulness pairs than for low meaningfulness pairs. In this experiment, however, Cieutat also varied stimulus and response meaningfulness independently, and the results showed that stimulus meaningfulness consistently yielded a greater facilitative effect than response