Practice Variability

Practice Variability

Introduction

Learning is defined as 'a set of processes associated with practice or experience leading to relatively permanent changes in the capability of responding' (Schmidt, 1988, p. 346). When learning is .evaluated in a retention test of what has been practiced with no knowledge of results (Salmoni et al., 1984) in which the subjects are asked to perform the task(s) they had previously practiced, two factors have been identified as being the main mediators of how much is learned.

First, there is the amount of practice trials performed by the learner prior to the retention test. Generally, it is accepted that the more practice trials were performed, the more an individual learned about the practiced skill (see Newell, 1991). Second, knowledge of results (KR) has been recognized, with the exception of practice itself, as the single most important determinant of learning (Bilodeau, 1966; see also Schmidt, 1988, 1991 for recent discussions). If optimal KR and an optimal KR schedule (see Winstein and Schmidt, 1990) can be provided in a particular setting, one can propose that the more practice trials are realized the more an individual will learn about that skill. Over the last decade there has been renewed interest in determining the best practice schedule for learning a motor skill or a set of motor skills (see Magi11 and Hall, 1990, and Van Rossum, 1990 for recent reviews). One question that has been asked when different skills, or variations of a particular skill, have to be mastered at approximately the same time, is whether it is better to practice them using a blocked or a random schedule. In a blocked schedule, each skill is practiced separately on a series of successive trials.

For example, in a conventional educational setting, if there is sufficient time in a particular session to realize 54 practice trials, the pupils will practice skill A for the first 18 trials. Then, they will practice skill B for the next 18 trials and skill C for the last 18 trials of the practice session. In contrast, under a random schedule, the pupils will vary from trial to trial, and from skill A, to skill B, to skill C without knowing in advance which skill they will practice on the next trial.

Discussion

Paradoxically, since the Shea, J.B. and Morgan (1979) study, investigators have repeatedly shown that, although the performance in acquisition (i.e., during practice with KR) is better when practiced under a blocked rather than a random schedule, learning (as measured in a retention test) is enhanced when practice occurred under a random schedule. Magi11 and Hall (1990) have recently proposed that this effect, known as the contextual interference (CI) effect, might be larger when the different movements to be learned are controlled by different motor programs rather than by the same motor program. This could be so because in the latter case, the subject only needs to modify a specific parameter of the required program instead of fully restructuring it between each trial, as ...