Bosman, E. A., & Charness, N. (1996). Age-related differences in skilled performance and skill acquisition. In F. Blanchard-Fields & T. M. Hess (Eds.), Perspectives on cognitive change in adulthood and aging (pp. 428-453). New York: McGraw-Hill.

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

Bosman & Charness (1996)

It is well-established that younger adults outperform older adults on many measures of cognitive functioning. Yet it is also clear that older adults can maintain high levels of performance for some tasks throughout adulthood. For example, McEvoy and Cascio found no evidence for a negative relation between age and job performance, even though measures of cognition negatively related to age are positively related to work performance. How do older adults achieve levels of performance comparable to younger adults? Bosman and Charness contrast four perspectives.

First, older adults change the criteria for successful performance. For example, older adults may give up driving at night because of vision impairment. As Bosman and Charness explain, "within their domain of expertise, older adults have learned to identify the conditions under which their performance is adversely affected by age-related declines and to selectively avoid these conditions" (p. 431). This perspective is called accommodation. Second, older adults may develop compensatory mechanisms or strategies for offsetting age-related declines. A loss in one relevant aspect of functioning is compensated for by a gain in another aspect. Third, from a maintenance, or disuse, perspective, age-related declines in skill components are prevented through sustained practice. Thus, unlike with compensation, a decline never occurs. Finally, the encapsulation hypothesis is that a skill is more than the sum of its parts—"once established, it may become independent of . . . component processes" (p. 432).

The molar-equivalence, molecular-decomposition technique is a common strategy for investigating compensation. Older and younger adults are matched on overall skill-level, and then age comparisons on component processes are made. Bosman and Charness point out a number of difficulties with this approach. First, matching older and younger adults on skill may result in a sample consisting of select older adults but less select younger adults in terms of overall level of cognitive functioning. This can be addressed by demonstrating normal age differences in the sample. Second, use of this strategy preclude looking at age-skill relations.

In a sample of younger and older adults matched for chess rating and performance on a choose-a-move task, Charness found a positive relationship between chess skill and memory for chess configurations, but a negative relationship between age and this measure. If the process captured by the memory task captures a genuine component of chess skill, then this suggests that the older adults may have used a compensatory mechanism to compensate for the memory deficit. Salthouse found that older adults looked farther ahead in the to-be-typed text than comparably skilled younger adults. Age-related declines in perceptual-motor measures, which were positively related to typing speed, suggest that looking farther ahead was a compensatory mechanism.

The fundamental prediction of the maintenance hypothesis is that skill preservation is attributable to sustained experience. Conversely, age-related skill declines are attributable to disuse. Typically, the goal of investigations of maintenance is to show that a particular type of experience (e.g., occupational) is related to preservation of certain cognitive abilities. Bosman and Charness explain that "the underlying cognitive ability assessed must be shown to be an important determinant of performance in the skill domain of interest" (p. 438). Salthouse et al. found comparable rates of age-related decline in spatial visualization ability for architects and non-architects. Significantly, the architects indicated that the spatial visualization abilities assessed were strongly related to their everyday activities as architects.

Are age-related differences reduced or eliminated with practice? An important methodological point is that older and younger adults must be given enough practice to reach asymptotic performance levels. This is demonstrated by showing that additional practice does not lead to additional improvements in performance.

Charness and Campbell devised a clever approach for investigating age differences in the knowledge compilation process. The task was to mentally square two-digit numbers using a particular algorithm. Measures of the time to apply the algorithm (total time) were obtained. In addition, measures of time to apply each of the components of the algorithm were obtained independently. The ratio of sum of component times to total time served as an index of the cost associated with applying the algorithm. No cost would result in a ratio of 1. Overhead associated with applying the algorithm would be mean that total time would exceed the sum of component times, therefore resulting in a ratio of less than 1. A ratio of greater than 1 would indicate that the components tasks are performed quicker in the context of applying the algorithm than alone. Charness and Campbell proposed that changes in the ratio reflect knowledge compilation. Age differences in ratio changes were examined to see whether older and young adults differ in knowledge compilation. The interesting finding was that age-differences in the ratios persisted throughout practice. "The implication is that the older adults were less efficient at implementing the squaring algorithm, even after considerable practice" (p. 443).

Age-differences in skill acquisition have also been examined in visual and memory search paradigms. Briefly, learning only occurs when target-distractor mapping is consistent across trials. In CM, stimuli that are targets on one trial will be targets on all trials, and stimuli that are distractors on one trial will be distractors on all trials. In VM, stimuli that are targets on one trial may either be targets or distractors on subsequent trials. Bosman and Charness explain that CM allows for the development of production rules for the automatic detection of targets. Consistency is thus necessary for production rules to form. And how is consistency communicated?

Consistency (or inconsistency) is communicated through feedback. Therefore, in domains like clinical assessment in which feedback is often not provided, knowledge compilation never occurs. To illustrate how CM/VM studies with laboratory tasks are relevant to learning more complex skills, consider the following example. John is interested in learning how to identify ducks. Now assume that a duck with a green head on one pond is a mallard, and a duck with a green head on another pond is also a mallard. In order to make the connection between green head and mallard, John needs feedback—say a guide to identifying ducks. Each time he looks he spots the green headed duck, he matches with the green-headed duck in the guide labeled "mallard." The consistency of the green-head mallard relationship is communicated through feedback, and in the absence of feedback will remain unknown to John. Through repeated pairings of the green-head mallard label, a production rule forms: IF green head THEN mallard. Now consider a situation in which ducks with green heads are sometimes mallards, but at other times are woodducks or green-winged teals. This is an example of variable mapping, and feedback will not help John learn because a green-headed duck on one trial might not be a mallard on another trial. Production rules for the automatic identification cannot form under this type of mapping.

This is a sidetrack. What do visual and memory search studies suggest about age-differences in skill acquisition? Salthouse and Somberg looked at the effects of extended training (50 hours) on performance of psychomotor tasks in younger and older adults. In brief, they found that, for an RT task, older adults improved more rapidly initially but that age differences were stable thereafter. Fisk and Rogers found similar results using a memory search task under consistent mapping conditions. Somewhat different results are found with visual search, but I’ll have to come back to that later. Lindenberger, Kliegl, and Baltes found that even after extensive practice younger adults outperformed older adults on a memory task that involved using the Method of Loci.

To what extent do previously acquired skills enhance learning of new skills? Lindenberger compared training effects with the Method of Loci in a sample of graphic designers and unselected adults. The younger graphic designers and younger unselected adults did not differ and outperformed both groups of older adults. However, the older graphic designers outperformed the older unselected adults. Bosman and Charness state, "The implication is that the older graphic designers’ professional experience in generating visual images enhanced their ability to acquire skill using the Method of Loci" (p. 447).

A number of questions with respect to age-related differences in skill acquisition. For example, some studies show that practice reduces age-related differences, while others show the differences persist. Bosman explain that "It may be the case that one of the factors influencing the effect of practice is the age sensitivity of the underlying processes" (p. 449). If the underlying processes are not age sensitive, then practice may reduce or eliminate differences. However, if they are age sensitive, practice may have little effect. The more general question is: What are the mechanisms underlying age differences, and does practice operate on the same processes that produced the initial age differences? A final interesting remark concerns the possibility that aging has different effects depending on when the skill was acquired. That is, "skill acquired while young may be more robust with regard to the effects of aging, while skills acquired when older are not" (p. 450).