[ CogSci Summaries home | UP | email ]

Morrow, D. G., Leirer, V. O., & Altieri, P. A. (1992). Aging, expertise, and narrative processing. Psychology and Aging, 7, 376-388.

Author of the summary: David Zach Hambrick, 1998, gt8781a@prism.gatech.edu

Morrow, Leirer, Altieri, & Fitzsimmons (1994)


The goal of this investigation was to examine "whether aviation expertise reduces age differences in a task that is similar to air traffic control (ATC) communication" (p. 134). The task was to read ATC commands, and then to repeat them back. The assumption in the administration of this task was that "expertise should improve performance only to the degree that the task is compatible with, or organized to take advantage of, expert knowledge" (p. 134). Age-related declines in complex task performance are well-documented. However, older experts may show preserved levels of functioning on specific tasks for one of at least three reasons. First, experts may be more select in terms of level of cognitive functioning than non-selected adults (e.g., age x intelligence interaction). Second, experience may preserve high levels of functioning (maintenance). Finally, older experts may compensate for decline in certain aspect of functioning through gains in other aspects of functioning.


Morrow et al. argue that "The possibility of compensation should depend on properties of the task" (p. 135). They argue that compensation should only occur for highly relevant tasks, and define relevant tasks as those "that are compatible with expert knowledge and procedures that are familiar to experts . . . " (p. 135). Morrow et al. use this definition to explain why, on the one hand, Salthouse (1984) found evidence for compensation in typing but, on the other, Salthouse et al. did not find evidence for maintenance in architects and why Lindenberger et al. found that graphic design experience did not reduce age differences in a Method of Loci task. In short, according to Morrow et al., the degree of domain relevance will determine whether age differences are reduced for experts relative to non-experts. The problem is that degree of domain-relevance is intuitive: How is one task more domain-relevant that another and to what degree?


The subjects were 48 pilots (24 young, 24 old) and 48 non-pilots (24 young, 24 old). Interpretation of an age x expertise interaction on an aviation task would be complicated by possible selection biases. For example, pilots may be a higher functioning group than non-pilots. However, all age x pilot status interactions on WAIS variables were non-significant. The most interesting finding was an age x expertise (pilot status) interaction predicting readback accuracy for heading (e.g., 320, 270) in the expected direction: pilots were more accurate than non-pilots, and older pilots were as accurate as younger pilots, whereas younger non-pilots were more accurate than older non-pilots. The age x expertise interaction was predicted for heading readback accuracy, but not for altitude and speed, for the following reason: "nonpilots may find heading commands particularly difficult, whereas pilots can rely on their domain knowledge" (p. 136). This amounts to an age x expertise x task difficulty interaction (which was, in fact significant in the form of age x expertise x command type.) In other words, expertise reduces age differences only when the task is difficult enough to really depend on domain-relevant knowledge and skill. Analysis of recall order (e.g., last-first strategy) and error type (e.g., close to correct) suggested that expertise benefits were attributable to use of domain-specific knowledge.


Morrow et al. state, "We found evidence that expertise reduced age differences in heading readback performance, suggesting that aviation knowledge compensated hypothesized age-related processing declines for older adults" (p. 142). This, however, is only supported if there were positive relations between processing resources and task performance and negative relations between age and task performance. In fact, in terms of score on the Block design portion of the WAIS, the former correlation was .23 and the latter correlation was -.45 (both significant). The interpretation of why the age x expertise interaction was found for heading readback, but not for altitude or speed readback, is interesting:


They were less predictable . . . they required mental rotation to be interpreted with respect to the mental map . . . and perhaps because they were first mentioned in the message. This difficulty is more likely to influence older nonpilots than older pilots, who can rely on domain knowledge to reduce processing demands on working memory.


Study 2 was identical, except that commands were presented via tape recorder. Also, message length was varied (short vs. long), and a fourth command was added (radio frequency). Analyses revealed some interesting results. First, age x experience interactions were significant for long messages (all comands). (However, it should be noted that older pilots were no more accurate than younger non-pilots.) Age x expertise interaction was also significant for heading at both message length, although the difference between young and old pilots was this time significant. For short commands only, the age x expertise interaction was marginally significant for heading, but not for the other commands. (This comparison is important because of the earlier suggestion that the age x expertise interaction emerges only at higher workload levels.)


Comments and Questions


The authors offer the following conclusion: "The . . . results suggest that expertise (domain knowledge) can compensate hypothesized age declines in the cognitive resources necessary to perform pilot tasks when both materials and procedures are highly relevant to piloting" (p. 145). The evidence that knowledge was the compensatory factor was indirect. That is, there was no measure of knowledge. It would be interesting to see whether the variance attributable to age x expertise interactions could be accounted for by a knowledge measure. What pattern of correlations would support the compensation for reduced working memory resources by knowledge argument?: 1) positive relations between task performance and working memory, and negative relations between age and working memory (decline) and 2) positive relations between both age and experience and age and knowledge, and 3) positive relations between knowledge and age.


The authors point to the need to examine the effects of different variables (e.g., task difficulty) on age-experience relations: "Future studies should independently vary task difficulty and domain relevance, which will require a more precise metric of task relevance" (p. 147). This latter issue in an interesting one. Salthouse has argued that a main effect of experience can serve as an operational definition of domain-relevance. Does this imply that a task showing a larger experience effect is more domain-relevant than a task showing a smaller experience effect? Domain relevance does not seem like an all or none thingóarenít there degrees? Another interesting issue concerns the role of task difficulty. Morrow et al. suggest that use of knowledge to compensate may be greatest when working memory demands are highest. That is, experts resort to knowledge, but non-experts canít. There may, however, be situations when working memory demands are so great that younger experts can capitalize on resources while older experts cannot. Under such circumstances, elimination of age differences at the expert level would not be expected. However, to the extent that experts could still rely on knowledge, older experts would probably outperform older non-experts (but maybe not younger non-experts.) It would be interesting to introduce a secondary task methodology to manipulate working memory demands.) In short, the influence of task difficulty and domain relevance on age-expertise relations should be investigated.


Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)
Last modified: Tue Mar 7 15:15:06 EST 2000