Receiving right/wrong feedback: Consequences for learning

Abstract

Prior work suggests that receiving feedback that one's response was correct or incorrect (right/wrong feedback) does not help learners, as compared to not receiving any feedback at all (Pashler, Cepeda, Wixted, & Rohrer, 2005). In three experiments we examined the generality of this conclusion. Right/wrong feedback did not aid error correction, regardless of whether participants learned facts embedded in prose (Experiment 1) or translations of foreign vocabulary (Experiment 2). While right/wrong feedback did not improve the overall retention of correct answers (Experiments 1 and 2), it facilitated retention of low-confidence correct answers (Experiment 3). Reviewing the original materials was very useful to learners, but this benefit was similar after receiving either right/wrong feedback or no feedback (Experiments 1 and 2). Overall, right/wrong feedback conveys some information to the learner, but is not nearly as useful as being told the correct answer or having the chance to review the to-be-learned materials.

Keywords:

• Feedback
• Learning

Acknowledgements

This work was supported by a collaborative activity award from the James S. McDonnell Foundation. We thank Mark McDaniel and Andrew Butler for useful comments on this research, and Katya Fernandez, Yusha Liu, Amy Hsu, Tyson Wepprich, and Lauren Rosenberg for preparation of stimuli, data collection, and scoring. The second author is now at Columbia University.

Notes

¹Common alternate misspellings were generated and the experiment was created so that the program would display “correct” for these responses.

²The data were also analysed conditional upon a “don't know” response, with results similar to those obtained when errors were analysed. We included only “incorrect” responses in this analysis to parallel the analysis in Pashler et al. (2005).

³The filler task was shorter for no review participants in Experiment 2 than in Experiment 1, as it was expected that participants in Experiment 2 would need less time to review 20 word-pairs than was needed to review 172 sentences in the first study.