Effects of age and curing on retrieval from semantic memory

Elderly subjects are known to perform less well than young subjects on laboratory tests of recall from episodic memory. Although the elderly report increased difficulty in recalling information from semantic memory, experimental attempts to demonstrate this deficit are equivocal. It is suggested that studies which use multiple choice tests to measure recall from semantic memory fail to find age-related deficits because the tests provide cues to aid in recall, a procedure known to reduce age-related differences in recall from episodic memory. When time to retrieve a single item of information from semantic memory is measured, some studies show an age-related deficit while others do not. When episodic recall is tested using categorized lists, the elderly show recall deficits largely because they access fewer categories than do young subjects. Semantic cues increase the number of categories recalled by the elderly subjects more than for young subjects in such tasks, Since studies with young subjects show that recall both from categorized lists and from a taxonomic category (a semantic recall task) proceeds via temporal clusters of related items, it was hypothesized that elderly subjects would show increased difficulty in accessing clusters of related items in a semantic recall task, just as they do in recall of categorized lists. Further, it was hypothesized that semantic cues would reduce the time taken by the elderly to access sequential clusters of information from semantic memory. In one experiment, healthy, well-educated young (ages 19-21) and old (ages 67-72) subjects were required to perform a Bousfield task: to generate examples from two taxonomic categories, foods and animals, for 15 minutes. The slope-difference algorithm, a procedure developed by Gruenewald and Lockhead, was used to categorize each subject's inter- item times (IIT's) into times between temporal clusters (BIIT's) and times between items within temporal clusters (WIIT). In a second experiment, a group of old subjects were given semantic differential labels as cues for recall on one of their two experimental trials. Results for the first experiment showed no age effect on mean BUT, number of clusters, or average cluster size for recall of food items. There were also no age effects during the first 5 minutes of recall of animals. Later in the task old subjects had longer mean BIIT's for animals than did young subjects. The differences appeared to result because old subjects tended to report primarily mammals, while young subjects reported birds, fish, reptiles/amphibians, and insects as well, A trend toward slower mean WIIT's for old subjects was attributed to slower vocalization rates. Thus, Experiment 1 failed to demonstrate age- related differences in time to access successive clusters of related items in semantic memory or in the rate at which items in a cluster are emitted. Higher repetition rates observed for the old subjects do support an age-related deficit in recognition. In the second experiment, only half the subjects reported that the semantic -differential cues were helpful in finding new items. No effect of cuing was observed for the food category. Cuing did significantly reduce mean BIIT for animals during the last 5 minutes of recall. However, the actual effect of cuing on number of clusters produced was minimal. It was suggested that more practice with the cues might have led to higher cue usage and a greater impact on BIIT.