Subjects were instructed to produce images to either high-imagery or low-imagery word lists. Recall over three trials indicated that hypermnesia occurred in the high-imagery condition, but that a decline in performance occurred in the low-imagery condition. Furthermore, analyses of the recall of individual items indicated that· the difference in recall occurred because there was less forgetting and more reminiscence in the high-imagery condition than in the lowimagery condition. The relationship between imagery and hypermnesia is discussed. Erdelyi and Becker (1974) demonstrated that under certain conditions, memory improves with multitrial recall (hypermnesia). Their experiment compared performance on stimulus lists comprised of either pictures or words. Subjects in the word condition showed a decrease in recall over three recall trials; however, the performance of subjects in the picture condition increased over trials.Erdelyi (Erdelyi, Finkelstein, Herrell, Miller, & Thomas, 1977) reasoned that production of visual images promoted the hypermnesia effect. Such images would readily be produced in the picture condition, but would be less likely to occur in the word condition. This hypothesis was tested by comparing the performance of three groups. One group saw the pictures used in the Erdelyi and Becker (1974) study. The other two groups saw the verbal equivalents of the pictures. In one word condition, subjects were instructed to make images to each word as it was presented (WIM, or "words with imagery instructions"). The other word condition did not receive such instructions (WNO, or "words with no imagery instructions"). As expected, the picture condition and the WIM condition both had increased recall over trials, while the hypermnesia effect did not occur in the WNO condition.The present experiment was designed to test the hypothesis that imagery encoding increases the likelihood of hypermnesia. In this experiment, all subjects were instructed to make images to the words. The imagery ratings of the word lists, however, were varied. Hypermnesia was expected in the high-imagery condition but not in the low-imagery group. This research was submitted as an honors thesis for the BA degree at Bowdoin College by the first author and was supported, in part, by a Langbein Fellowship to the first author. Samuel J. Popkin is now at the Andrus Gerontology Center, University of Southern California, Los Angeles, California 90007.
METHOD MaterialsThe Paivio, Yuille, and Madigan (1968) norms were used to select the two 40-word lists. The words in the high-imagery list (Hi I) had ratings of 5 to 7; those in the low-imagery list (Lo I) were rated 1 to 3. The frequency and meaningfulness of the words in each list were equated by selection of words with Thorndike-Lorge counts between 40 and AA and meaningfulness ratings between 4 and 6. Two random orders of the two lists were constructed.
SubjectsThe subjects were 40 undergraduates who were naive to verbal learning experiments. Twenty subjects were assigned to the Hi I list,...