Research Bites: Corrective Feedback – A Meta-Analysis


Li, S. (2010). The effectiveness of corrective feedback in SLA: A meta-analysis. Language Learning (60)2, 309-365. [$link]

Twitter Summary

Study of corrective feedback: effective for SLA; explicit for short term; implicit (recasts) for long  


As a former grammar Nazi language prescriptivist, I have become a bit leery about correcting people, especially students. For the general public, my mother excluded, I tend to ignore errors in speech or otherwise embrace them as a linguistic peek inside the backgrounds, experiences, and heads of my interlocutors. For non-native speakers, I have always been careful about when and where to correct. The classic accuracy/fluency dichotomy in which one corrects during accuracy-based activities only and lets anything flow during fluency-based activities has been about the only training and advice I received on giving oral corrective feedback (written feedback is a whole separate thing). Like most dichotomies, things are not always black and white: students not only need correction at many points other than focus-on-form(s) activities, but they also want it.

During the past decade or so, error correction has become a more prominent subject in language education. Moving from an art that says when and where to use it (based mostly on intuition), it has become more of a science, qualitatively and qualitatively measured for its effect and effectiveness. There are a lot of studies on corrective feedback that look at a lot a different features. Luckily, a meta-analysis (an analysis of the individual analyses) was completed by Shaofeng Li in 2010 that examined 22 articles and 11 dissertations on corrective feedback. Among the many factors he looked at, below is the most relevant information he found. Following that, there is a brief chart of the different types of corrective feedback (taken from Lyster, Saito, and Sato, 2013). Below that is a video of someone demonstrating the different correction strategies.

  • Corrective feedback has a “medium effect” on language acquisition for both the short and long term. A medium effect size (here, Fixed Effects: 0.61; Random Effects: 0.64) represents the magnitude of the effect on the included populations.
  • Explicit correction had a larger immediate effect than metalinguistic feedback and recasts. However, the effect of recasts may be retained for a longer amount of time.
    • In the studies that examined explicit correction, the tests used to examine the treatments were extremely similar to the treatements themselves and therefore may have contributed to the overall positive effect.
  • Likewise, explicit feedback seemed stronger in the short-term, with implicit correct stronger in the long-term (i.e. on delayed posttests).
    • Li writes that implicit feedback may be “more enduring.” He speculates that implicit feedback may lead to implicit knowledge, and thus perhaps automaticity. However, this needs more study.
  • Computer-based and face-to-face feedback had no significant difference in terms of effect.
  • Shorter treatments (50 mins or less) had a larger effect than longer treatments.
    • These short treatments were mostly laboratory based.
    • I might add that this seems obvious given the information is still fresh in participant’s heads.
  • Feedback from native-speakers or computers was more effective than teacher feedback.
    • All studies with native-speakers and computers were done in a lab, as opposed to the classroom, and lab studies tended to have higher effects in general. This variable needs more study.



(Check the description on YouTube for the names of the strategies shown)
Among the numerous things that Li analyzed (such as publication type or research context), I tried to choose the ideas that held the most relevance for classroom teaching. I recommend checking out his entire article, which, at over 50 pages, is actually a pretty easy read.

All this being said, moving from practical knowledge to practical classroom action is not so easy. Not only do we need to know how to provide corrective feedback, but when to use it, how to allocate classroom time to it, how best to provide it to individuals or groups without leaving the rest of the class without a meaningful task, and how not to overburden ourselves with correcting everything or everyone every time.

So, to you dear reader, I ask: how do you handle corrective feedback? What are your strategies for classroom implementation of this important L2 facet?

One thought on “Research Bites: Corrective Feedback – A Meta-Analysis

Leave a Reply