Repetition and Learning
REPETITION AND LEARNING
Sayings such as "Practice makes perfect" illustrate the well-known fact that repetition improves learning. This was discussed by numerous ancient and medieval thinkers and was demonstrated empirically by Hermann Ebbinghaus, the first researcher to carry out a prolonged series of experiments on human memory. In a classic 1885 book, Ebbinghaus showed that retention of information improves as a function of the number of times the information has been studied. Since the time of Ebbinghaus, countless investigators have used repetition to study learning and memory.
Although experimenters typically find a consistent relationship between repetition and learning, numerous authors (Guthrie, 1935) have pointed out that this does not necessarily mean that the learning process itself has to be either gradual or continuous. Most learning situations contain a number of smaller facets or subproblems that must be mastered before learning is complete. It is possible that each of these subproblems is mastered suddenly, perhaps through insight. However, the subproblems are learned at different times, with more and more of them mastered as the number of trials increases. This analysis proposes that a gradual improvement in learning as a result of repetition may reflect the accumulation of subproblems that have been mastered in a sudden fashion. Distinguishing between a truly continuous learning process and the accumulation of small, sudden insights is difficult. A common assumption is that learning may be either gradual or sudden, depending on the background of the learner and the nature of the information to be learned. For example, Harry Harlow (1949) showed that learning to novel situations may occur slowly and continuously but may appear in sudden flashes of insight when the organism has had experience in a number of similar situations. Thus, although the amount of learning may appear to grow gradually and continuously as a result of repetition, determination of whether subcomponents of the task are learned gradually or suddenly is more difficult and requires careful analysis.
Although the total amount learned increases as a function of repetition, the amount learned on each trial will not be constant. Repetition effects exhibit negative acceleration: The most learning occurs in the first exposure to a stimulus or situation, and the amount learned in each subsequent exposure continually declines until further improvement is too small to be detected. The rate of learning is negatively related to the amount already learned. Hintzman and Curran (1995) have shown that people can register the occurrence of a repeated stimulus while failing to learn more about its specific details. First impressions of a repeated stimulus are particularly important, as people may show little evidence for having noticed subtle changes that are introduced to a stimulus after its first presentation (DiGirolamo and Hintzman, 1997).
Why Does Repetition Improve Learning?
Anderson and Schooler (1991) have pointed out that the sensitivity of learning to repetition is evidence for its efficiency and adaptiveness because the frequency with which information has been used in the past is a very good predictor of whether it will be needed in the future. Still, although repetition has been intensively studied, the mechanisms underlying its effects are still poorly understood. Moreover, there is no reason to believe that a single explanation could apply to all situations where repetition facilitates learning.
Of particular interest to many researchers has been the effect of repeated study on human memory, and the two dominant explanations of these repetition effects were both discussed by Ward (1893). One class of explanations (called a functional approach by Ward but more commonly known as strength theory in twenty-first-century scientific circles) claims that there is a single location in memory storage that corresponds to an event. Every time the event is repeated, that location (known as the memory trace) increases in effectiveness or strength. It is also assumed that stronger traces are easier to retrieve from memory than are weaker traces. Repetition thus improves learning by increasing the strength of a single memory trace. A second class of explanation for the effects of repetition on memory was called an atomistic approach by Ward but is now known as multiple-trace theory. This approach assumes that every occurrence of an event is a unique episode. Every time an event occurs, a separate, independent memory trace is formed. This trace contains information about the time and situation in which that occurrence happened. The more times an event occurs, the more traces of that event are placed in memory. According to this multiple-trace theory, repetition improves learning because finding at least one trace of an event becomes easier when there are more traces of that event in memory.
A fundamental difference between these two accounts concerns the representation of the individual occurrences of a repeated item. The strength theory claims that each occurrence of an event strengthens a single memory trace. Since each occurrence has the same effect, the specific details of individual occurrences are lost. In contrast, the multiple-trace theory claims that every occurrence produces its own trace. The individuality of specific occurrences is maintained.
Experiments distinguishing between these two accounts have often required participants to remember a list of words. A word on the list may occur once or a varying number of times. After seeing the list, participants are shown the list items again and asked to make a judgment regarding how often each item occurred on the list. Even when they do not expect to be tested on the frequencies of the items, people are typically able to perform this task with considerable (but not perfect) accuracy. However, strength theory and multiple-trace theory make different proposals as to how participants are able to make judgments about the frequency of occurrence of list items. Strength theory claims that participants retrieve the memory trace corresponding to a test item and evaluate that trace's strength. They then use the strength to make a judgment of frequency. For example, if a memory trace is very strong, participants will guess that the item occurred many times on the list. If a memory trace is weak, they may decide that the item occurred once (or possibly not at all) on the list. In contrast, the multiple-trace theory claims that participants make judgments of frequency by retrieving as many traces as possible of that item occurring in the context of the list. They then base their judgments on a count of the traces they found.
Numerous experiments have investigated whether a frequency judgment is based on a single trace or on the retrieval of many different traces. For example, Hintzman and Block (1971) showed participants two lists of words, five minutes apart. Some words occurred on both lists. Each word occurred zero, two, or five times on List 1 and zero, two, or five times on List 2. After seeing both lists, participants were asked to estimate frequency of occurrence separately for each list. They were quite accurate at this task; their estimates were chiefly influenced by the frequency of the item on the list being judged and were influenced little by the item's frequency on the other list. Such a finding is difficult for a strength theory to explain: If judgments of frequency were based simply on the overall strength of the trace of the word, people would not be able to make separate estimates for the frequency of an item on two lists. However, a multiple-trace theory would predict this finding because frequency judgments are seen as being based on a count of individual traces, each carrying information about its time of formation.
Subsequent studies have found further evidence in favor of a multiple-trace theory. For example, when some words are presented visually and others auditorily, participants are able to give separate frequency judgments for each kind of presentation. Also, they are able to judge how often a word followed another word on a list. Such findings suggest that the individual identities of the occurrences of a repeated event are maintained in memory, as assumed by the multiple-trace theory (Greene, 1992).
Studies such as these suggest that a multiple-trace theory is necessary to account for the effects of repetition on memory. They do not show that such a theory is sufficient to account for all the effects of repetition. The question of whether repetition has other effects in addition to the creation of multiple memory traces has not been resolved, although there is some evidence that repeated events are remembered better than would be expected on the basis of memory for specific presentations (Watkins and LeCompte, 1991). Moreover, there is little evidence that would allow one to determine whether the multiple-trace approach can be applied to all of the situations in which learning is improved by repetition.
When Is Repetition Ineffective in Increasing Learning?
Although the emphasis in this entry has necessarily been on the mechanisms through which repetition improves learning, one should not assume that repetition alone is always sufficient. For example, consider a common coin, such as the American penny. Although people have seen such coins countless times, as Nickerson and Adams (1979) showed, people can have quite poor memory for the details of a penny. They are often unable to remember exactly where such features as the date and the words "In God We Trust" are located. There is no need for people to attend to these features of a penny because pennies can easily be distinguished from other coins on the basis of their size and color. This suggests that attention to an event may be necessary before repetition of that event leads to noticeable improvements in memory. The generality of this claim has been established by studies demonstrating poor memory for other currencies, for the details of telephone dials, and for the messages of common advertisements.
Additional examples of ineffective repetition have come from experiments on rote rehearsal (Glenberg, Smith, and Green, 1977; Rundus, 1977). In these studies, participants read repeated words aloud over and over. An unexpected memory test on the words is later given. Memory performance is usually only slightly affected by the number of times that a person read each word aloud. On the other hand, if people are encouraged to carry out more active, effortful processing on the words, memory improves dramatically as study time is increased.
One situation in which repetition impairs memory is when people have to recall a short series of digits or letters in order. Recall is impaired if one of the items is repeated in the series. This phenomenon, known as the Ranschburg effect, was introduced into the modern psychological literature by Crowder and Melton (1965). Critical to understanding this negative effect of repetition is the fact that people have to remember that an item was repeated and the locations of each occurrence in the series. The Ranschburg effect occurs because recall of the first occurrence of the repeated item inhibits accurate recall of the second occurrence (Greene, 2001).
Thus, repetition need not lead to improved learning. Rather, repetition leads to increased opportunities for learning to occur. Whether learning takes place will depend on the type of information that has to be remembered and the amount and nature of processing that a person carries out.
See also:DISTRIBUTED PRACTICE EFFECTS
Anderson, J. R., and Schooler, L. J. (1991). Reflections of the environment in memory. Psychological Science 2, 396-408.
Crowder, R. G., and Melton, A. W. (1965). The Ranschburg phenomenon: Failures of immediate recall correlated with repetition of elements within a stimulus. Psychonomic Science 2, 295-296.
DiGirolamo, G. J., and Hintzman, D. L. (1997). First impressions are lasting impressions: A primacy effect in memory for repetitions. Psychonomic Bulletin & Review 4, 121-124.
Ebbinghaus, H. (1885; reprint 1964). Memory: A contribution to experimental psychology. New York: Dover.
Glenberg, A., Smith, S. M., and Green, C. (1977). Type 1 rehearsal: Maintenance and more. Journal of Verbal Learning and Verbal Behavior 16, 339-359.
Greene, R. L. (1992). Human memory: Paradigms and paradoxes. Hillsdale, NJ: Erlbaum.
—— (2001). Repetition effects in immediate memory in the absence of repetition. In H. L. Roediger III, J. S. Nairne, I. Neath, and A. M. Surprenant, eds., The nature of remembering: Essays in honor of Robert G. Crowder. Washington, DC: American Psychological Association.
Guthrie, E. R. (1935). The psychology of learning. New York: Harper.
Harlow, H. F. (1949). The formation of learning sets. Psychological Review 56, 51-65.
Hintzman, D. L., and Block, R. A. (1971). Repetition and memory: Evidence for a multiple-trace hypothesis. Journal of Experimental Psychology 88, 297-306.
Hintzman, D. L., and Curran, T. (1995). When encoding fails: Instructions, feedback, and registration without learning. Memory & Cognition 23, 213-226.
Nickerson, R. S., and Adams, M. J. (1979). Long-term memory for a common object. Cognitive Psychology 11, 287-307.
Rundus, D. (1977). Maintenance rehearsal and single-level processing. Journal of Verbal Learning and Verbal Behavior 16, 665-681.
Ward, J. (1893). Assimilation and association. Mind 2, 347-362.
Watkins, M. J., and LeCompte, D. C. (1991). The inadequacy of recall as a basis for frequency knowledge. Journal of Experimental Psychology: Learning, Memory, and Cognition 17, 1,161-1,176.