Individual differences in the representation of sentences in memory.

A cued recall procedure was used to assess the nature of the memory representation that underlies the ability of mentally retarded and nonretarded individuals to remember single sentences. Mentally retarded, equal-CA, and equal-MA subjects listened to a list of sentences after which their ability to recall the object noun of the sentence was assessed when they were provided recall cues that contained (a) only the subject noun of the original sentence, (b) only the verb of the sentence, or (c) both the subject and verb. As expected, performance for all groups was best when they were provided the subject plus verb cue relative to the single word cues. In addition, the groups differed in the magnitude of this two-word cue advantage, with the retarded subjects exhibiting the smallest and the equal-CA subjects exhibiting the largest advantage. This finding reflects a difference in the degree to which mentally retarded and nonretarded individuals construct sentence representations that more precisely specify the meaning of the sentence through the integration of its constituents.

The instantiation of general terms by deaf adolescents/adults.

Two studies were conducted using severely and profoundly deaf high school students to determine their ability to instantiate particular exemplars of general nouns and to use those instantiations as retrieval cues. The results indicated that the deaf adolescents/adults could instantiate when asked to do so but did not do so spontaneously; sentence recall was best when the retrieval cue matched the word used in the original sentence; and recall of sentences in which all information was explicit was better than of sentences in which some information had to be inferred. Impoverished semantic representations, difficulty in integrating semantic representations, and insufficient strategy use were suggested as possible alternative and competing explanations for the obtained results.

Sources of difficulty in the solution of verbal arithmetic problems by mentally retarded and nonretarded individuals.

Effects of several logical (i.e., operation type and amount of extraneous information), memory (i.e., availability of memory aids and number of problem presentations), and semantic variables (i.e., problem text type) on verbal math problem-solving performance were assessed. Results revealed that the overall problem-solving performance of mildly mentally retarded adolescents was inferior to that of nonretarded fourth graders in spite of comparable performance on a computational screening test. Although the retarded individuals experienced particular difficulty with subtraction and static problem texts, the two groups responded similarly to the other experimental variables. The possibly important role of comprehension in problem-solving was discussed.

Semantic processing can facilitate free recall in mildly retarded adolescents.

The efficacy of semantic processing in free recall was investigated in two experiments with EMR adolescents. In Experiment 1, they were taught to use one of two semantic strategies for memorizing a 15-word list. Compared with controls, neither strategy helped recall either in original learning or transfer. In Experiment 2, one of the semantic strategies, a story mnemonic, was investigated further. Rather than being taught to construct their own stories as in Experiment 1, subjects in Experiment 2 were provided with experimenter-composed stories. They showed better immediate recall and retention after 2 months than did no-strategy controls. However, about 1 year after original learning, the retention of experimental and control subjects no longer differed. Discussion focused on the story mnemonic's potential utility and the criteria for judging such potential, e.g., amount of facilitation, ease of training and performance of the strategy, and the degree of its generalizability.

Comprehension and recall of sentences by mentally retarded and nonretarded individuals.

A cued sentence-recall task was used to determine the extent to which 24 mildly mentally retarded adolescents and 24 equal-MA nonretarded children differed in their ability to recall sentences and to infer and utilize particular exemplars of general nouns as retrieval cues. We found that the sentence recall performance of the retarded adolescents was poor relative to that of the nonretarded children; however, both groups found general and particular cues to be equally effective retrieval aids for target sentences. Differential sentence reconstruction and editing strategies were suggested as possible sources of the obtained recall differences.

Use of categorical knowledge to aid recall.

Recall performance of 84 mildly retarded adolescents and 84 non-retarded fourth graders was compared under various encoding conditions. The seven encoding conditions consisted of the following format-cueing combinations: simultaneous-subject cues, simultaneous-experimenter cues, sort-subjects cues, sort-experimenter cues, sequential-subject cues, sequential-experimenter cues, and sequential-no cues. Stimuli were 24 colour pictures from six categories. Results indicated generally comparable levels of performance for retarded and non-retarded subjects. However, retarded subjects were less responsive to sort and experimenter cue conditions, which led to increased recall in the non-retarded subjects. It was suggested that retrieval difficulties may be interfering with the ability of retarded individuals to utilize categorical knowledge to facilitate recall.

Reasoning by trainable mentally retarded and young non-retarded individuals.

Three groups of trainable retarded children, and three groups of non-retarded children matched for mental age, were presented with two problem-solving tasks. The tasks required the problem-solver to combine relevant parts of the problem and independently arrive at a solution from their combination. The data was analysed to determine: (1) whether the ability to reason in trainable retarded children showed developmental changes, and (2) to compare the development of reasoning behaviour in trainable retarded children with that of comparable MA non-retarded children. Implications were made from the findings for classroom teaching.

Facilitation of class-inclusion performance in mildly retarded adolescents: feedback and strategy training.

Two experiments were conducted to explore the feasibility of training mildly retarded adolescents to solve class-inclusion problems. In Experiment 1, the effects of training consisting primarily of feedback were investigated. In Experiment 2, training with feedback alone vs. training with feedback plus a counting strategy was compared. Both experiments assessed the generality of training effects to specific training and/or posttest formats. Results suggested that training was generally effective in improving performance with untrained class-inclusion problems on immediate and delayed posttests. Feedback training with pictorial stimuli appeared to be effective regardless of posttest format. The effects of the other training method/format combinations varied according to posttest format.

Sentence mediation and stimulus blocking in free recall.

In a two-phasing training study, EMR adolescents were initially provided with sentences linking each pair of 2 words in a 20-word list to be free recalled. In a second phase, subjects were required to create and verbalize their own sentences. This sentence-mediation condition was compared with (a) one in which items were presented in the same fixed pairs from trial-to-trial, but without sentences and (b) a group that received items in changing pairs. Results indicated superiority of both sentence-mediation and fixed-pairs treatments in comparison to the changing-pairs condition on both correct recall and subjective organization. Sentences did not facilitate recall more than did consistent ordering. The results were discussed in the context of the growing evidence that providing input structure can remediate organization deficits in retarded persons.

Cognitive processes and academic achievement in EMR adolescents.

Twelve cognitive process variables were investigated as predictors of reading and arithmetic achievement and studied for internal structure, with 115 EMR adolescent subjects. Stepwise regression and factor analyses were employed to study prediction and structure, respectively. Memory variables were the most important predictors of reading. They were also involved in arithmetic but to a lesser extent. The data suggested that the ability to generate and utilize strategies facilitating the recall of unstructured material as well as the capacity to be sensitive to strategy-relevant structure embedded in stimuli were important prerequisites for the acquisition of reading skills. An oddity task, measuring the ability to maintain the same relational focus for successive applications to new stimulus material, was found to be most pertinent for predicting arithmetic computational skills.

Transfer of categorial clustering set in mildly retarded adolescents.

The relative effectiveness of instructions to adopt an organizational strategy vs. exposure to category labels was investigated. Twenty-four retarded adolescents were assigned to each of three conditions (blocking plus instructions, blocking plus labels, and blocking) for pretraining and training. Immediately after training, all subjects received a transfer task consisting of a new randomly organized list containing words from the same (training) categories for one-half of the subjects and from different categories for the other half. When exposure to category labels accompanied blocking, results indicated increased clustering and recall during the training phase and a trend toward increased clustering during the transfer phase.

Effects of letter-reversals training on the discrimination performance of EMR children.

Ninety mentally retarded elementary-school children who committed letter-reversal errors on a pretest were randomly assigned to one of three (one experimental, two control) training conditions consisting of seven sessions given over a 5-week period. Subsequent statistical analyses revealed that for one of four letter-discrimination posttests, the experimental subjects performed significantly better than the placebo control subjects. Differences between the experimental and the method-specific control subjects were in the expected direction but not statistically significant. Aptitude times Treatment interactions for three of the four posttests revealed the training effect to be most pronounced with younger and more retarded children.

Relative effectiveness of various letter discrimination procedures in directionality pretraining.

Two investigations were conducted to provide information relevant to programming basic reading skills for retarded individuals. Both experiments utilized various procedures to teach subjects to discriminate between the letters b and d. The role of feedback and task format conditions was investigated in Experiment 1. Results indicated that two-choice discrimination learning was significantly more difficult than either oddity or matching-to-sample conditions. In Experiment 2 the effects of problem type and sample reduncancy were examined. Analyses indicated that simultaneous matching-to-sample procedures were significantly easier than successive matching-to-sample procedures. Performance with a delayed matching-to-sample transfer task did not reflect differential effects of training in either study.