Tim Lee, Ph.D., McMaster University

Research in the field of motor learning is often motivated by theory, with seemingly little relevance to practitioners who want to see the bottom line about how new findings could benefit their learners. As well, the simple tasks and contrived laboratory conditions used in experiments also limit the potential for application of the findings. Fortunately, however, this seems to no longer be the case, as exemplified by some recent advances in research that has examined practice conditions and methods of providing feedback. This research seems to have a clear bottom line.

In one area of study, researchers have examined performance and learning under conditions where different skills are practiced within the practice session; the focus being on the order by which the skills are practiced. One condition that is common in many teaching, coaching, and rehabilitation settings is to drill practice, where performance attempts on one skill are concentrated for a period of time, only moving on to a different skill when some criterion of success has been attained. An example in golf would be practice at a driving range by hitting many balls with one club before changing clubs. In contrast to drill practice, researchers have assessed learning under random conditions where practice attempts in one skill are interspersed with attempts at other skills. In the golf example, this would involve changing clubs after each attempt.

It is probably not surprising at all that researchers have found drill practice to be a faster and more economical method of acquiring a certain level of skill within the practice session. The rather counterintuitive finding of this research, however, has been that performance following random practice almost always exceeds drill practice when skill is measured after a rest period (in a "retention" test) or in a simulated competition (a "transfer" test). These findings, which were initially reported for laboratory tasks under contrived conditions, have since been replicated many times, using sport, industrial, and rehabilitation tasks, and using subjects of different age, gender, and health characteristics. Quite simply, drill type practice is good for facilitating short-term performance gains, but random practice is a better method for learning.

A related set of findings has also been reported by researchers who have investigated different methods of providing advice to learners about errors that they have made (called "augmented feedback"). The general purpose in these studies has been to determine how feedback can be provided in such a way that its usefulness can be maximized for learning. Different methods have been used, and generally the research has found that providing feedback often and as soon after performance as possible will show the fastest gains in skill. However, similar to the random/drill studies, tests of retention and transfer show opposite effects. Providing feedback relatively infrequently and only when needed the most, and providing it after the athlete has done some thinking about the problem first tends to have the strongest impact on learning. Feedback is a powerful method of helping athletes to learn but seems to have a negative impact on learning if used in such a way that the athlete comes to rely on the feedback-provider too much.

Both sets of research findings seem rather counterintuitive at first, but are they really? Both findings suggest that making performance gains "easy" do not lead to the types of long-term retention and adaptability of the skills to new situations that practice should promote. In contrast, the emphasis that is placed on the learner to problem-solve in random practice and when feedback is withheld seems to have its maximum benefit in competition - exactly when we want the value of practice to emerge. As well, the skills used in competition need to be accessed as the situation arises ("randomly") and the athlete must problem-solve during a match without feedback from the coach. So, in a way, these training methods are much closer to the actual competition situation than are the methods that produce the best performance in practice.

There are two obvious difficulties in applying these methods to everyday learning environments. First, we often coach in the same ways that we were taught -- and the use of drill practice and frequent feedback have a long history as recommended training methods. Second, performance gains in practice are incorrectly assumed to automatically result in gains in competition, and the "artificial" barriers to performance gain that result from random practice and infrequent feedback may be troubling to the learner. Fortunately, neither of these difficulties present a major limitation -- they just require a shift in perspective on thinking about what is best for performance in practice versus performance in competition.