The claim that bizarre mental images facilitate recall more than common images was examined by instructing 32 5s to form a bizarre image for each of 24 noun pairs and a common image for each of 24 other pairs. After a 4-min. filler task, each S was given a cued recall test on the entire list of 48 pairs. Contrary to expectations, bizarre images produced no more correct responses than common images. Since bizarre images took longer to form (5.98 sec. vs. 3.94 sec.), it was concluded that common images are in fact more efficient.

The effects of transfer from a modified free-recall task to verbal-discrimination (VD) learning were investigated. The free-recall procedure was designed to impart increments in experimental frequency to specific words that appeared in VD. Four experimental groups were given prior free-recall exposure to either all right VD items (A-H), all wrong (A-w), both the right and wrong items from same VD pairs (B-S), or both die right and wrong items from different VD pairs (B-D). A fifth group served as a control and learned a free-recall list which did not contain any VD items (c). In terms of number of correct responses during VD learning, the groups were ordered as follows: A-R, c, A-W, B-S, and B-D. The results were interpreted as being consistent with derivations from die frequency theory of VD learning.

A conceptual scheme for reinterpreting the "re-pairing decrement" in verbaldiscrimination (VD) learning was outlined. It was hypothesized that the decrement could be accounted for within a frequency theory by liberalizing the "counting" postulate. The decrement was viewed as a special case of increased VD difficulty resulting from a breakdown in the rule that all correct alternatives are uniformly higher in frequency value than their respective incorrect partners. An experimental manipulation designed to increase the likelihood of a breakdown in this rule was successful in increasing the re-pairing decrement.

No abstract

In a free-recall task there are no restrictions placed on the order in which Ss are allowed to record the items they are able to recall. Yet, even with a list of presumably unrelated words there is a marked consistency in recall orders across trials, and there appears to be some agreement between Ss in specific free-recall orders (Tulving, 1962). When restrictions are placed on order of recall, as occurs with serial-learning tasks, performance is facilitated if the required order parallels the common free-recall orders (Earhard, 1967; Tulving, 1965). The preferred serial order is derived from free-recall protocols by placing in adjacent serial positions items which frequently appeared in adjacent positions during free recall. Thus, the serial order is determined by normative free-recall orders. With lists of associatively related words, similar results have been obtained with variations in serial order designed to maintain the associative structure of the list (Weingartner, 1963).In the present experiment each S learned three serial lists: one with the items arranged randomly, one with the items arranged according to the specific S's preferred serial order, and one with the items arranged according to some other S's preferred serial order. The preferred orders were determined from a prior task in which Ss arranged each list into an "easy-to-Iearn" serial order. The purpose of the present experiment was to determine if this direct method of identifying preferred serial order produced arrangements of words which facilitated serial learning. If Ss utilize idiosyncratic associations in constructing a preferred serial order, then there ought to be some advantage when one learns a serial list that corresponds to his own preferred order, as compared to a list that corresponds to some other S's preferred order. METHOD MaterialsEach S was required to learn three serial lists of meaningful words. There were 15 words in each list. and the lists were selected from Deese (1959). Lists number 2, 8, and II were selected with interitem associative strength values of 4.3, 4.0, and 9.3, respectively. Procedure During the initial phase of the experiment Ss were required to arrange each list into a preferred serial Qrder. The nature of a serial-learning task was briefly explained, and then Ss were asked to arrange the words in a sequence which they believed would be easiest to learn serially. The Ss were not informed that they would be required to learn the lists at a future date.Approximately three weeks intervened between the ordering phase and the learning phase of the experiment. All Ss served in both phases. During the learning phase Ss were Psychon. Sci., 1968. Vol. 13 (2) required to learn all three Ii,!'. in immediate succession. Each S learned one of the three lists in the serial order he had used in arranging the words (Own). one list in the serial order some other S had used in arranging the words (Other), and one list in a random serial order (Random). The lists, conditions. and order of conditions within the learning s...