Research illuminates the benefits of random practice over blocked practice in motor learning
This is an excerpt from Motor Learning and Performance 5th Edition With Web Study Guide by Richard Schmidt & Tim Lee.
Blocked Versus Random Practice
In many, if not most, real-world settings, the learner's goal is to acquire more than a single skill or task in a limited practice period, sometimes even in a single practice period. Physicians practice different skills related to surgery (such as suturing and knot-tying skills), musicians practice multiple songs at a time, tennis players practice serving and volleying as well as the more usual ground strokes during a single session, and so on. An important question confronting the learner or instructor is how to sequence the practice at these various tasks during the practice session so as to maximize learning. Two variations have powerful effects on learning: blocked and random practice.
Suppose that your student has three tasks (tasks A, B, and C) to learn in a practice session and that these tasks are fundamentally different, such as tennis serves, volleys, and ground strokes. That is, tasks are chosen such that one cannot argue that any of them are in the same class or use the same GMP. A commonsense method of scheduling such tasks would be to practice all trials of one task before shifting to the second, then to finish practice on the second before switching to the third. This is called blocked practice, in which all the trials of a given task (for that day) are completed before moving on to the next task. Blocked practice is typical of some drills in which a skill is repeated over and over, with minimal interruption by other activities. This kind of practice seems to make sense in that it allows the learners to concentrate on one particular task at a time and refine and correct it.
Another practice scheduling variation is called random (interleaved) practice; where the order of task presentation is mixed, or interleaved, across the practice period. Learners rotate among the three sample tasks so that, in the more extreme cases, they never (or rarely) practice the same task on two consecutive attempts. And from a common-sense perspective, the random method, with its high level of trial-to-trial variability, its high level of contextual interference would not seem optimal for learning.
The Shea and Morgan
John Shea and Robyn Morgan (1979) conducted a groundbreaking experiment that revolutionized the way scientists think of the processes involved in practice. Following some of the original ideas of William Battig (1966), Shea and Morgan had subjects practice three different tasks (A, B and C) that involved responding to a stimulus light with a correct series of rapid movements of the hand and arm, with each task having a different predetermined sequence. One group of subjects practiced the tasks in a blocked order, completing all task A practice before moving to task B, which they completed before moving to task C. A second group practiced in a random order; no more than two consecutive trials could occur for any one task. The two groups had the same amount of practice on tasks A, B, and C and had the same amount of total practice—they differed only in the order in which the tasks were presented.
The results are presented in figure 10.10. The goal was to respond to the stimulus and complete the movements as quickly as possible, so lower total times indicate more-skilled performance. Notice that, during acquisition, the blocked condition was far more effective for performance (with shorter times) than the random condition. But recall that differences during acquisition cannot be interpreted as differences in learning; rather, delayed retention (or transfer) tests are needed to evaluate learning (these concepts were presented in chapter 8).
Shea and Morgan tested for learning by conducting retention tests after 10 min and 10 days; these tests were conducted under either randomized or blocked conditions, which produced four subgroups. The following abbreviations indicate the condition in acquisition and the condition in retention, respectively: R-B, R-R, B-R, and B-B. The first character in the pair indicates the condition during acquisition (random or R and blocked or B), and the second member of the pair indicates the performance conditions in retention.
When the retention tests were under random conditions, the group that had random practice in acquisition (R-R, solid blue) greatly outperformed the group with blocked conditions in acquisition (B-R, solid red). When the retention tests were under blocked conditions, again the random condition in acquisition (R-B) outperformed those who had blocked conditions in acquisition (B-B), but these differences were much smaller than for the random retention tests. Clearly, the random conditions in acquisition were always more effective for retention, but this benefit was clearly dependent on the nature of the retention test.
An issue regarding variable practice was alluded to earlier. A very important factor concerns how variable practice is scheduled; this issue now becomes better understood due to the results of Shea and Morgan (1979). Studies in which variable practice was scheduled in a trial-by-trial random order showed rather large advantages compared to constant practice (e.g., Catalano & Kleiner, 1984; also Pigott & Shapiro, 1984). The Shea and Morgan findings suggest that scheduling how variable practice is ordered influences its effectiveness.
Why Random Practice
Is So Effective
The Shea and Morgan findings surprised many scientists in the field by showing that, even though random conditions result in much less skilled performance than blocked conditions in acquisition, random-practice conditions produce more learning. The findings were a large surprise because most conventional viewpoints would suggest that learning should be maximized by those conditions that make learners most proficient during practice—there was no motor learning theorizing that could explain this opposite result. As a result, some interesting new hypotheses were offered to explain the findings.
Shea and Zimny (1983) argued that changing the task on every random-practice trial made the tasks more distinct from each other and more meaningful, resulting in more elaborate memory representations. As revealed in subject interviews after the experiment, random-practice subjects tended to relate the task structure to already learned materials (creating “meaningfulness”), such as discovering that task B had essentially the shape of an upside-down “Z.” Also, they would make distinctions between tasks, such as “Task A is essentially like task C, except that the first part is reversed” (creating “distinctiveness”). The blocked-practice subjects, on the other hand, tended not to make such statements. Instead they talked of running off the performances more or less automatically, without thinking much about it, and blocked practice did not induce the kind of comparative and contrastive efforts in practice that were experienced during random practice. According to this elaboration hypothesis, increased meaningfulness and distinctiveness produce more durable memories for the tasks, and thus increased performance capabilities in tests of retention and transfer.
An alternative hypothesis explains the beneficial effects of random practice somewhat differently. Lee and Magill (1983) suggested that when the learner shifts from task A to task B, the “solution” that was generated (in short-term memory; see chapter 2) for performing task B causes the solution previously generated for task A to be forgotten. When task A is encountered again a few trials later, the learner must generate the solution anew; therefore, performance in practice is relatively poor. Yet this solution-generation
process is assumed to be beneficial for learning (see also Cuddy & Jacoby, 1982). In blocked practice, on the other hand, the performer remembers the solution generated on a given trial and simply applies it to the next trial, which minimizes the number of times the learner must generate new solutions. Therefore, performance during practice in a blocked schedule is very effective because the solution, once generated, is remembered for a series of trials. Yet learning is poor because the learner is not required to generate a “new” solution to the task on every trial. In this way, the key focus of the forgetting hypothesis is the fact that new solutions are required frequently in random practice, but not in blocked practice; thus, the development of the solution for the task is the key feature that facilitates learning. Interestingly, the forgetting hypothesis suggests the somewhat ironic and counterintuitive idea that “forgetting facilitates learning.”
A number of investigations have evaluated and supported both of the hypotheses. For example, in a study by Wright (1991), members of a blocked-practice group were encouraged to make explicit comparisons of the task just practiced with one of the other tasks to be learned—essentially inducing this group to “mentally” practice the tasks with meaningful and distinctive processing. This special blocked-practice group outperformed the other practice groups that had a similar intervention but without the benefits of the explicit comparative and contrastive processing. The results supported the elaboration hypothesis predictions because of the insertion of these specific mental processing activities.
A key prediction of the forgetting hypothesis was that random practice forces more extensive planning operations on each trial compared to blocked practice. A study by Lee and colleagues (1997) attempted to reduce the need for these planning operations by presenting a powerful “model” just before each practice trial. This model was designed so that it would inform subjects how to perform the next trial, and because the model provided extremely strong memory guidance for the upcoming trial, the model was hypothesized to prevent the construction process (because the model provided the solution for the next trial). In the experiment, the presence of the model was combined with random practice. The model, eliminating as it did the subject's requirement to reconstruct the “solution” for the next trial, would interfere with performance in acquisition more or less as blocked practice does. As figure 10.10 shows, the model obliterated the usual benefits of random practice.
In the experiment, the random and blocked conditions are contrasted with this special “random + model” condition. Clearly, the model was beneficial for performance during acquisition (when the model was present), as seen on the left side of figure 10.11 where the “random + model” group was far more skilled than the group that had only random practice. However, in the retention tests, where the model was withdrawn, the random + model group regressed considerably, to the point that this condition led to the most error in the delayed retention test. Providing the powerful model before each practice trial, while it was beneficial for performance when it was present, was disastrous for learning. The model obliterated the beneficial advantages of random practice. These findings support strongly the forgetting hypothesis for the random- versus blocked-practice effect, and they show that random practice is not necessarily the “magic bullet” for effective motor learning.
A number of studies have provided evidence supporting the elaboration hypothesis, and a number have supported the forgetting hypothesis; but no clear “winner” has yet emerged. As a result, it is probably a better idea to consider these hypotheses as complementary, rather than competing, explanations of the random- versus blocked-practice effects. The beneficial effects of random practice over blocked practice appear to be due to several factors:
▶ Random practice forces the learner to become more actively engaged in the learning process by preventing simple repetitions of actions.
▶ Random practice gives the learner more meaningful and distinguishable memories of the various tasks, increasing memory strength and decreasing confusion among tasks.
▶ Random practice causes the learner to forget the short-term solutions (from working memory) to the movement problem after each task change.
▶ Forgetting the short-term solution forces the learner to generate the solution again on the task's next trial, which is beneficial to learning.
Learn more about Motor Learning and Performance 5th Edition.More Excerpts From Motor Learning and Performance 5th Edition With Web Study Guide
Get the latest insights with regular newsletters, plus periodic product information and special insider offers.