Academic journal article Memory & Cognition

Making Related Errors Facilitates Learning, but Learners Do Not Know It

Academic journal article Memory & Cognition

Making Related Errors Facilitates Learning, but Learners Do Not Know It

Article excerpt

Abstract Producing an error, so long as it is followed by corrective feedback, has been shown to result in better retention of the correct answers than does simply studying the correct answers from the outset. The reasons for this surprising finding, however, have not been investigated. Our hypothesis was that the effect might occur only when the errors produced were related to the targeted correct response. In Experiment 1, participants studied either related or unrelated word pairs, manipulated between participants. Participants either were given the cue and target to study for 5 or 10 s or generated an error in response to the cue for the first 5 s before receiving the correct answer for the final 5 s. When the cues and targets were related, error-generation led to the highest correct retention. However, consistent with the hypothesis, no benefit was derived from generating an error when the cue and target were unrelated. Latent semantic analysis revealed that the errors generated in the related condition were related to the target, whereas they were not related to the target in the unrelated condition. Experiment 2 replicated these findings in a within-participants design. We found, additionally, that people did not know that generating an error enhanced memory, even after they had just completed the task that produced substantial benefits.

Keywords Memory . Errors . Generation . Metacognition . Associative learning

(ProQuest: ... denotes formulae omitted.)

This article addresses the effect of making errors on learning. Should one learn by studying materials without making mistakes or by attempting to produce answers and committing inevitable errors that such attempts entail? When errors are leftuncorrected, they typically remain incorrect (Butler, Karpicke & Roediger, 2008; Fazio, Huelser, Johnson & Marsh, 2010; Metcalfe & Kornell, 2007; Pashler, Cepeda, Wixted & Rohrer, 2005; Pashler, Zarow & Triplett, 2003). However, feedback is highly effective in allowing the learner to correct previously incorrect answers (Butler et al., 2008; Metcalfe, Kornell & Finn 2009; Pashler et al., 2005; Pashler et al., 2003). In this article, only errors followed by corrective feedback were considered. The question here was whether, and under what conditions, committing an error facilitates learning. Although the main focus of this article is the memorial consequences for errorful, as compared with errorless, learning, a related question of interest is the following: Are learners aware of the circumstances in which committing errors can be effective for improving learning? Accurate metacognitive knowledge is important for metacognitive control and strategy selection (Kornell & Son, 2009; Metcalfe & Finn, 2008). If the learner is not aware of the potential efficacy of a learning strategy, he or she might implement suboptimal strategies. Hence, people's metacognitions about the effects of errors may be nearly as important as the effects of the errors themselves.

From a theoretical standpoint, there is reason to believe that even corrected errors might impede learning. An error, in essence, is often thought to be conflicting or competing information with regard to the correct response. As such, it should create an interference situation. In standard proactive interference paradigms, the first pairing of a target (B) with a particular cue (A) results in interference when cue A is later paired with a different response (C) (J. R. Anderson & Reder, 1999; M. C. Anderson & Neely, 1996; Barnes & Underwood, 1959; Loftus, 1979; McGeoch, 1942; Melton & Irwin, 1940; Osgood, 1949; Webb, 1917). Although there are several theories concerning how this interference arises (e.g., J. A. Anderson, 1973; J. R. Anderson & Bower, 1972; Eich, 1982; Gillund & Shiffrin, 1984; Hintzman, 1984; Metcalfe, 1990; Osgood, 1949), there is general agreement that it does occur. Interference from errors might be expected to be even greater than interference theory would normally predict, since interference theory does not take into account whether or not the interfering information is self-produced. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.