[ CogSci Summaries home | UP | email ]

Salthouse, T. A. (1989). Aging and skilled performance. In A. M. Colley & J. R. Beech (Eds.), Acquisition and performance of cognitive skills (pp. 247-263). Chichester, UK: John Wiley & Sons.

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

The chapter is organized around three issues: 1) age-related decline in basic cognitive abilities, 2) effects of age on acquiring new skills, and 3) effects of age on attainable levels of proficiency.


Age and Elementary Cognition


A tenet of the information processing perspective is that cognitive functioning is predicated on a few basic processes and that these processes serve as building blocks for more complex forms of cognition. It is well established that younger adults outperform older adults on a wide range of cognitive tests, including perceptual comparison speed, verbal memory, spatial memory, etc. Salthouse therefore concludes that "it seems reasonable to expect that older age would be a disadvantage in most complex skills" (p. 249).


Efficiency of Skill Acquisition


Salthouse observes that there is little evidence relevant to the question of whether younger adults are more efficient in skill acquisition than older adults. The argument that the absence of evidence for age-related increases in cognition, vis a vis documented positive age-experience relations, cannot be interpreted as evidence for age-related declines in acquisition efficiency for a number of reasons. First, range of experience might become restricted with age. Second, if acquisition efficiency remains stable, information would be lost and gained and the same rate. What is the evidence?


Studies of skill acquisition are typically short-term, and are often not age comparative. A study by Salthouse and Somberg (1982) is one of the few age-comparative studies that involved extensive practice. After initial trials, they found no evidence for age x practice interactions for a reaction time task. Asymptotic levels of performance were different, but asymptotic levels do not reflect skill acquisition efficiency. Skill acquisition efficiency is operationalized as the slope of the practice function. Steeper slopes indicate more efficient acquisition. Ceiling effects for three other tasks limited conclusions. Nevertheless, Smith, Kliegl, and Baltes showed that younger adults required fewer trials to memorize a list of words using the Method of Loci than older adults. In short, the results on age and the efficiency of skill acquisition are mixed.


Mechanisms of Skill Maintenance


Salthouse describes different strategies for understanding skill maintenance. In the Molar Equivalence-Molecular Decomposition strategy, young and old subjects are matched on overall level of performance, and then age differences in component processes are examined. The absence of age differences at the molecular level may be attributable to positive effects of experience or to sampling problems in which the older subjects are more select or in which the younger subjects are less select. The converse finding—age differences at the component level—also lends itself to different interpretations. First, people may compensate for declines in one component through gains in another component. This would be shown by different regression weights for people in different age groups. And second, something analogous to compilation (Anderson, 1982) might occur in which overall task performance is no longer predicated on the efficiency of components. This alternative predicts weaker relations between overall performance and subtask performance as a function of practice.


Salthouse reviews research on three activities—chess, bridge, and typing—in which the Molar Equivalence-Molecular Decomposition strategy has been applied. Charness applied the approach in a series of studies on bridge. In one study with a near-zero age-skill correlation, for a memory task, he found positive relations with skill but negative relations with age. However, it is unclear whether molar skill level was maintained through a compensatory or compilation process. It might be interesting to look use a concurrence ratio methodology along with the Molar Equivalence-Molecular Decomposition approach. However, in another study, he found that although time to make a bid was negatively related to skill (shorter latencies for bids, higher skill), there was no relation with age. This suggests a maintenance explanation—that is, perhaps older adults maintained the relevant component through experience.


Charness also extended the analytical strategy to chess. In one study in which older and younger adults were again matched on skill level, he showed a positive relationship between performance on a memory task, but a negative relationship between age and performance on this task. Charness also found that older adults were quicker in a choose a move task than younger adults, and suggested that more efficient search in older adults might be the mechanism that offsets age-related decline in processes tapped by the memory task. The memory task reflects memory skills that support that rapid and efficient storage and retrieval of information in LTM. Thus, Charness is suggesting that older adults could circumvent LTM storage and retrieval processes by knowing better where to look. As Salthouse observes, however, verbal protocol analyses suggest that the older adults were considered fewer moves than the younger adults. This could be attributable to 1) a strategy difference in which older adults decided quicker on a move, 2) less extensive search, or 3) more efficient search.


Salthouse (1984) provided more convincing evidence for compensation in a study of transcription typing. In the sample, older and younger subjects were matched for typing skill, and there was a positive relation between typing speed and perceptual motor speed, and a negative relation between this variable and age. Briefly, he showed that older adults compensated for declines in perceptual processing speed by looking further ahead in the to-be-typed text than younger adults. Eye-hand span was determining how many letters in advance of the currently typed letter could be covered before a performance deficit was observed.


Age and Skill


What can the study of aging tell us about skill acquisition. First, using a molar decomposition-molecular decomposition approach, the finding that individuals differ on molar level performance might suggest that performance of a complex task is more than the sum of its parts and that some coordination is required at the molar level of performance. Second, using a molar equivalence-molecular decomposition approach, the finding of no age differences on component measures might be informative about the effects of experience on cognitive functioning, assuming that selection bias issues are addressed. Finally, a compensation finding might suggest how flexible task performance is.


How can the study of skill acquisition contribute to aging research? Studying skill acquisition throughout the lifespan is informative about the range and nature of abilities that can be preserved in adulthood. Moreover, studying skill acquisition is relevant to understanding the causes of individual differences in age-cognition relations. More specifically, according to one popular idea, age-related changes in cognitive functioning are attributable to lack of current use. If so, then one would expect qualitatively differences between young and old individuals in skill performance and between individuals who lack current practice in that skill. Also, experience in that activity might moderate the relationship between age and cognition.


Comments and Questions


Salthouse provides an excellent review of issues related to age and skill acquisition. The second issue concerns age-related changes in efficiency of skill acquisition—that is, in the "responsivity to training." He points out a number of important methodological issues. First, a fair assessment of the issue requires extended training. Ideally, asymptotic levels of performance should be reached. Second, many studies do not include younger adults. A study of age differences obviously must be comparative. The third issue concerns mechanisms of maintenance. Two analytical approaches are contrasted: molar-equivalence and molecular decomposition. In the former, younger and older adults are matched on molar skill level (e.g., chess rating), and age differences in component level processes are examined. In the latter, younger adults are matched on component level processes, and age differences on overall performance level are examined. The compensation idea is appealing, but Salthouse points out that a compensatory mechanism must be identified for it to be convincing. If age differences at the component level are not observed, the issue of selection biases (select older or less select younger) must be addressed.


Salthouse’s chapter is relevant to expert performance and aging. Do older and younger experts achieve expert performance in the same way? Charness’s research on chess suggests that search in older adults may be more efficient, but alternative explanations are possible. What are the effects of experience on skill? One possibility is that skill maintenance is attributable to preservation of component processing efficiency.


Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)
Last modified: Tue Mar 7 14:52:34 EST 2000