Component Analysis and Conceptions of Skill
Steven W. Keele University of Oregon
The more one practices a skill, the more the skill improves. In fact, for some purposes it is best to think that no real limit exists for the degree of improvement. This is illustrated in Fig. 6.1 from Crossman ( 1959), showing the speed with which operators of a cigar-making machine make cigars. Different operators practiced on the machine for up to 7 years and made up to 10 million cigars. What is interesting is that performance bottomed out only after 2 years of practice. The apparent reason for lack of improvement after 2 years is not the human but the machine itself. Operators with that amount of experience approached very close to the cycle time of the machine, which is indicated by the horizontal dotted line. It would appear that improvement on this simple skill would have continued for a much longer time were it not for machine limitations.
There is an additional point to make about the practice curve in Fig. 6.1. The horizontal and vertical axes are plotted on a logarithmic scale, and this leads to a bit of an illusion. At first glance one might think that the skill improved linearly with practice--that's really not true because both axes are on a log-log scale. A feature of a logarithmic scale is that is compresses large values so, in fact, the amount of improvement declines with increasing practice. This general formulation, a linear decrease in the log of performance with the log of practice, has been labeled DeJong's Law by Crossman.
Now improvement occurs not just within an individual but also across individuals over the course of generations. Ryder, Carr, and Herget ( 1976) examined the improvement of running speed over the years 1930 to 1970 for distances ranging from 100 yards to 15 miles. Over successive years runners have become faster and faster, and the relationship between speed and year shows no evidence