[ CogSci Summaries home | UP | email ]

Krampe, R. Th., & Ericsson, K. A. (1996). Maintaining excellence: Deliberate practice and elite performance in younger and older pianists. Journal of Experimental Psychology: General, 125, 331-359.

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

Krampe & Ericsson (1996)


Krampe and Ericsson investigate the role of deliberate practice in the maintenance of skill in music. Specifically, they challenge the general factor notion of age-related skill decline. That is, "we question the common assumption that age-related declines in speeded performance is inevitable and report evidence for the role of continued practice in older professional pianists’ maintenance of expert-level performance" (p. 331).


The General Factor Account


Salthouse and others have argued that age-related declines in a wide range of cognitive functioning is attributable to general slowing, that is, to reduced efficiency of basic cognitive and perceptual operations. This hypothesis reduces age-related declines toa single, common cause. In terms of skilled performance, the argument is that basic cognitive processes are building blocks for more complex forms of cognition. Reduced efficiency of these processes therefore results in decline in the more complex skills.


Salthouse et al. advanced a variation of this account in their study of architects. In brief, they found comparable rates of age-related decline in performance on spatial visualization tasks for architects and non-architects, even though the latter outperformed the former at all ages. Salthouse et al. argued that this reflected initial individual differences in spatial visualization ability, but that experience did not attenuate age-related declines in the sample of architects. They called this the preserved differentiation perspective.


Deliberate Practice


Ericsson and colleagues make distinguish among activities relevant to improvement—deliberate practice activities—and types of experience not relevant to performance improvements. With respect to skill and aging, Krampe and Ericsson make a distinction between acquisition and maintenance. This distinction is important because acquiring a skill may require more practice than maintaining one. Krampe and Ericsson argue that, contrary to the general factor account of aging and its emphasis on inevitable decline, "regular efforts are necessary to maintain expert . . . performance" and physiological and anatomical adaptations are the result of intense training. In short, "a genuine understanding of older experts’ performance requires a more detailed description of past and concurrent deliberate practice activities to distinguish decrements due to reduced practice levels from the inevitable decrements due to aging" (p. 333).


General Factor vs. Selective Preservation


The criterion tasks were 1) performance of a technically simple piece of music and 2) a finger tapping task. The predictor tasks were measures of 1) perceptual processing speed and 2) music-specific processing speed. Krampe and Ericsson claim that "the general factor account predicts that age-related decline should be similar for the skill-related performance and measures of general processing speed . . . " (p. 334). This interpretation is confusing, because older adults’ superior performance could be based on their greater knowledge of music. That is, general processing speed does not necessarily underlie skilled performance. Indeed, Salthouse (1991) raises the possibility that expertise might be conceptualized as the circumvention of human processing limitations. That is, "A . . . possible interpretation of how experts seem to escape normal limitations on human information processing is that expertise may develop through the acquisition of mechanisms or processes specific to particular tasks or activities" (p. 291).


Perhaps, then, Krampe and Ericsson’s prediction is based on the assumption that skill in music is based on aspects of cognition whose relations with age are mediated through perceptual speed. This would explain the suggestion that age-relations with both fluid and crystallized measures can be accounted for by individual differences in perceptual speed. Curiously, however, age relations with crystallized measures are usually positive. Perhaps with this in mind, Krampe and Ericsson state that "Even if the rate of age-related decline in music-related performance differs for amateurs and experts, the general factor account would still predict an age-related reduction in general capacity and some amount of associated decrease in music-related performance for both amateurs and experts" (p. 334). The most important prediction concerning the general factor account is that age-performance relations should not be influenced by experience.


Krampe and Ericsson contrast the extended general factor account with the selective maintenance account, which "asserts that superior performance is the result of domain-specific mechanisms that individuals have acquired and must actively maintain through deliberate practice" (p. 334). The two major predictions of this account are 1) comparable rates of age-related decline in general processing speed for amateurs and experts and 2) reduced age-related decline in music-specific processing speed for experts. Most important, the degree of reduction will be a function of amount of recent deliberate practice, or maintenance practice.


Study 1


The interesting results from study 1 can be summarized as follows. First, rates of age-related decline in general processing speed (i.e., digit symbol and choice CRT) were similar for expert and amateur pianists. Second, there was an interaction between expertise and a contrast of single-finger vs. alternate-finger tapping. Third, age effects on the music-related speed tasks were smaller for experts than for amateurs (I think). Moreover, a three-way interaction among age, expertise, and type of task (music-related or general) was significant. Age-related reductions on the music-related tasks were reduced for experts, but not for amateurs, and age differences were similar for experts and amateurs on the general processing speed task. Finally, the age x expertise interaction was apparently not significant for the musical interpretation task.


The next series of analyses focused on deliberate practice estimates. Three deliberate practice measures were computed with the distinction between acquisition and maintenance in mind: 1) practice accumulated until 20; 2) practice during 20-year period of most intense deliberate practice, and 3) practice during last 10 years. Briefly, three models were compared. The first model included skill level, age, and the skill level x age interaction. The second model included these three variables and four practice variables: practice until age 20, practice during peak 20 years, practice during last 10 years, and current practice. The final model included only the practice variables.


The results relevant to the selective maintenance interpretation can be summarized as follows. First, in the combined model, the practice factors had unique effects on performance on the complex conditions. The effects were small, however. Second, the magnitude of the effects of the skill variable were reduced when the practice variables were added. Third, the practice model accounted for less overall variance than the other models. Fourth, practice during the last 10 years was the strongest predictor among the practice variables, and accounted for a reliable amount of variance for several tasks. Fifth, for experts but not for amateurs, the residual in the practice model was unrelated to age. Finally, age-effects in the amateur group, but not in the expert group, persisted after controlling for the practice variables. The musical interpretation task was not included in this analysis, and overall, effects on the musical interpretation task were generally small.


Comments and Questions


Krampe and Ericssson suggest that deliberate practice is the key to maintenance of skilled performance in adulthood. Some results seem supportive, but, unfortunately, Krampe and Ericsson overload the reader with statistics. They do not make a strong case for the argument. It should also be noted that the age x expertise interaction was non-significant for the musical interpretation task, and that the practice variables did not have effects. The most interesting finding "is that the pattern of age-related declines across speeded music-related performance and psychometric measures of general processing speed differed for amateurs and experts" (p. 353).



Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)
Last modified: Tue Mar 7 15:39:05 EST 2000