The preliminary studies to be described had the objective of discovering whether it would be feasible to employ the techniques of free recall with the subjects with whom we were to deal. Specifically, it was hoped to show that simultaneous visual presentation of items allows subjects ample opportunity to organise their recall; for it provides them with the chance to perceive relations between items while those items are present. Further, it was intended to demonstrate that various experimenter manipulations could have significant effects on degree of organisation. Since these studies were conceived as experiments in their own right, the results obtained may be treated with some confidence. The major aim of Experiment 1(1) was to discover whether significant

clustering (O-E) and ITR (O-E) did occur with subjects of low VA given random order presentation. Secondary aims were to investigate relations between these two measures; to investigate subjects’ spontaneous naming at presentation; and to assess any differences between mongoloid and other subjects. Specific details of Experiment 1 are presented in Table 1. Its

essential features are that it presented visually and simultaneously a very few items in a multi-trial procedure to fairly young subjects (pupils at Special Schools). The material consisted of three sets, two of apparently unrelated and one of categorically related items. The latter comprised three categories with two items in each category. All subjects received the three sets of material on successive days, with the order of sets randomised. The only other feature of this experiment in which it differed from most of the other experiments was that subjects were not forced to name items in the presentation phase of the task. They were merely told to look at the pictures carefully; however, many of them spontaneously named, perhaps as a result of the preliminary naming trial. Thus in general the situation was set up in such a way as to provide

every opportunity for subjects to organise the randomly ordered items. In particular, the use of the visual modality and of simultaneous rather

than successive presentation of items had this objective (2, 3). The results indicated that the objective was attained. The scores expected by chance and those actually obtained were compared for the clustering and ITR measures both for the mongoloid and for the non-mongoloid groups. For the mongoloid group, there was a highly significant difference between obtained and expected by chance clustering and ITR; for the non-mongoloid group, the difference was highly significant in the case of clustering, and just significant in the case of ITR (all t tests with correlated means). These results have been replicated in essence (4). However, it must be stressed that the ITR scores over all subjects for

the two sets of unrelated material were only just significantly correlated (r = 0.25). The vast variability of ITR scores over trials for each subject also suggested that it was a fragile dependent variable. Moreover, it cannot be said that the significant clustering scores

necessarily indicate that hierarchical coding processes were operating. Indeed, there were significant sequential constancies in the recall of the categorisable set of material, since obtained and expected ITR scores calculated for that set differed highly significantly. When an analysis was carried out on all the pairs of items which occurred in the same order on successive recall trials, it was found that 53% of these pairs consisted of items from the same category; only 20% would be predicted on the basis of chance. Next, analysis of all those pairs of items from the same category which occurred on successive trials were undertaken. It was found that in 63% of these pairs, the two items were in the same order on the two trials (e.g. car, train). Clearly, certain strong sequential constraints were operating. Further, it must be stressed that although significant differences

between obtained and expected measures were found, the percentage of total possible clustering was still only 37.5% and of ITR, 15%. Nevertheless, there are indications that hierarchical processes may also have been operating in the case of the categorisable material; for clustering (O-E) was highly significantly correlated with amount recalled for both mongoloid and non-mongoloid subjects; this correlation was significant in the case of non-mongoloid subjects and just significant in the case of mongoloid subjects even .when the recall-independent percentage measure was employed. In terms of the secondary aims of the experiment, no correlation

between clustering scores for categorisable material and ITR scores for unrelated material was obtained. Further, ITR, unlike clustering, did not correlate with amount recalled. When subjects spontaneously

named, the order of their naming bore no relation to the order of their recall; rather, it was in some directional sequence (e.g. left-to-right, top-to-bottom) corresponding to the arrangement of items on the slide. Differences between mongoloid and non-mongoloid subjects will be

discussed in Chapter 10.