Learning from failure: Errorful generation improves memory for items, not associations

Abstract

Potts and Shanks (2014) recently reported that making mistakes improved the encoding of novel information compared with simply studying. This benefit of generating errors is counterintuitive, since it resulted in less study time and more opportunity for proactive interference. Five experiments examined the effect of generating errors versus studying on item recognition, cued recall, associative recognition, two-alternative forced choice and multiple-choice performance. Following Potts and Shanks (2014), participants first attempted to learn the English definitions of either very rare English words or Euskara nouns. During encoding, participants either guessed the definition (and almost always made an error) before the correct definition was revealed, or simply studied the words for an equivalent period. Experiments 1-4 used rare English words. In these experiments, generating errors led to better subsequent recognition of both the cues and targets compared with studying (Experiments 1 and 3). Tests of cued recall and associative recognition, by contrast, revealed no significant benefit of generating errors over studying (Experiments 1-3). Generating errors during encoding also improved performance on a two-alternative forced choice test when the correct target was presented with a novel foil, but not when the familiarity of the target and the foil was matched (Experiment 4). In Experiment 5, a different set of materials-Euskara nouns-and a different (intermixed) encoding procedure was adopted. Here, guessing improved target recognition (performance was improved on a multiple-choice test with unfamiliar foils), but impaired cued recall performance. These results suggest that, when learning word pairs that do not have a pre-existing semantic association, generating errors strengthens the cues and targets in isolation, but does not strengthen the cue-target associations.

Weighing Up Exercises on Phrasal Verbs: Retrieval Versus Trial‐and‐Error Practices

“…A series of experiments by Seabrooke et al. ()—all approximate replications of Potts and Shanks ()—indeed suggest that the benefits of error‐prone meaning guessing before learning the correct word–meaning correspondences are less likely to emerge when multiple‐choice posttest items include foils that the participants also encountered during the learning procedure (and which consequently stand an equally good chance of being recognized). Moreover, in a cued recall test (where the meaning of the target words was presented, and the participants were asked to produce the target words), it was the study‐only (i.e., no meaning guessing) condition that fared the best.…”

“…The participants in both trial-and-error conditions, but especially those who had been asked to freely guess the word meanings, performed better in a posttest than participants in a comparison condition who had been given the correct word meanings from the start and had been asked to study them. In a recent publication, however, Seabrooke et al (2019) have pointed out that the posttests used in Potts & Shanks (2014) did not necessarily measure whether the participants recalled the association of the new words and their meanings. These posttests used a multiple-choice format where the foils (i.e., the incorrect options) were mostly new items, not included in the learning phase.…”

“…Moreover, in a cued recall test (where the meaning of the target words was presented, and the participants were asked to produce the target words), it was the study‐only (i.e., no meaning guessing) condition that fared the best. Seabrooke et al.’s () findings thus cast doubt over the efficacy of an error‐prone trial‐and‐error procedure for learning associations. This is highly relevant to the subject of the present article, because learning a phrasal verb requires establishing a dual association: (a) between the phrasal verb and its meaning, and (b) between the verb and its particle.…”

“…In a recent publication, however, Seabrooke et al. () have pointed out that the posttests used in Potts & Shanks () did not necessarily measure whether the participants recalled the association of the new words and their meanings. These posttests used a multiple‐choice format where the foils (i.e., the incorrect options) were mostly new items, not included in the learning phase.…”

A dimensional summation account of polymorphous category learning

“…Although the issue of whether making errors is beneficial or detrimental to learning remains controversial (Ashby, Maddox, & Bohil, 2002;Edmunds, Milton, & Wills, 2015;Potts & Shanks, 2014;Seabrooke, Hollins, Kent, Wills, & Mitchell, 2019), it is undeniably the case that, in the current experiments, single feature pretraining is not well matched to the polymorphous control condition in terms of the number of classification errors participants produce. It is therefore possible that the SFPT advantage comes from this difference in error frequency, rather than from the provision of deterministic feature-category associations assumed to be the cause by the dimensional summation hypothesis.…”