What the research says

The Hypercorrection Effect: Confident Mistakes Get Fixed Best

You would expect a mistake someone was sure of to be the hardest to shift. Memory research found the opposite. What the hypercorrection effect is, and how to use it in class.

By the Giga Quizzes team4 min read
Sharpened yellow pencils standing round a pink eraser
Photo: Tracey Parish, Unsplash, Unsplash License, cropped.
In this post
  1. 1The study that named it
  2. 2People know which answer is right
  3. 3What a later review added
  4. 4What these findings do not say
  5. 5Using it in a lesson

Picture two students who both get the same question wrong. One guessed and knew it. The other was certain. Which of them is more likely to give the right answer next time, once they have been told it? Most people, teachers included, would bet on the guesser. The certain student, after all, has a wrong idea firmly in place. Memory research says the opposite, and the finding has a name: the hypercorrection effect.

The study that named it

In 2001 the psychologists Brady Butterfield and Janet Metcalfe set out to test a prediction that followed from standard models of memory. The idea was that the more confident a person is in an answer, the stronger that answer is in memory, and the more it should get in the way of learning a different one. So, they expected, errors made with high confidence should be the hardest to correct.

Participants answered questions and rated how confident they were in each answer. They were then shown the correct answers, and later took a retest. The researchers could then compare what happened to errors made with low confidence and errors made with high confidence.

The result went against the authors’ own expectations. The errors people had been most sure of were the ones they were most likely to fix on the retest. The more certain someone had been of the wrong answer, the better they learned the right one.

People know which answer is right

The study had a second, quieter finding. On the retest, some people found that both answers came to mind: the correct one they had been shown, and their original wrong one. In those cases, they nearly always chose the correct answer.

The authors read this as a sign of a refined metacognitive ability, people’s skill at judging their own knowledge. The old wrong answer was still there in memory, but people could tell it apart from the new right one. A confident error, once corrected, did not leave a lasting tug of war behind it.

What a later review added

In 2017 Janet Metcalfe reviewed the wider research on learning from errors for the Annual Review of Psychology. Her starting point was a habit she saw in American classrooms: avoiding errors during learning as if they were dangerous. The laboratory evidence, she argued, suggests that this may be counterproductive, at least for most students.

  • Making errors and then getting corrective feedback is beneficial to learning, not harmful.
  • The benefit is strongest for high-confidence errors, which are corrected more readily than low-confidence ones.
  • The feedback matters. Corrective feedback is crucial, including going through the reasoning that led to the mistake.
  • Errors help teachers too. They show what students actually think.
  • Tolerating errors encourages exploration. Students who are not afraid of being wrong engage more actively.

Her conclusion is a useful one for anyone who runs a classroom quiz: if the goal is to perform well when it matters, it may be worth letting students make and correct errors in low-stakes situations, rather than avoiding errors at all costs.

What these findings do not say

Hypercorrection is about errors that are followed by the correct answer. A confident mistake that nobody ever corrects is just a confident mistake, and it can last. The evidence also comes mainly from laboratory studies, and the effect fits questions with a clear right answer better than matters of judgement. And the review’s caveat about neurologically typical students is worth keeping in mind; it does not claim the same pattern holds for every learner.

Using it in a lesson

The practical lesson is to make confident errors visible and then correct them well. A quiz is one of the easiest ways to do that, because a whole class commits to an answer before anyone is told what is right.

Three ways to put hypercorrection to work

  • Ask for confidence. After a tricky question, add a quick poll: “How sure were you?” The students who were sure and wrong are the ones primed to learn most from what comes next.
  • Use questions that tempt a confident mistake. A question where the most common wrong answer feels obviously right is more valuable than one nobody can guess. Science, history and geography are full of them.
  • Make the correction count. Show the right answer, then explain why the wrong one looked right. Metcalfe’s review points to going through the reasoning, not only the answer.

Then bring the question back. A week later, ask the same thing in a slightly different form. If hypercorrection did its job, the students who were confidently wrong will now be confidently right.

One more thing follows from the research, and it is about tone. Students who expect to be laughed at for a wrong answer learn to hedge, and a hedged guess does not produce the same effect as a committed answer. Treating wrong answers in a quiz as normal, even interesting, keeps students committing. The post on teaching in the reveal has more on the moment straight after the answer is shown, and pre-testing shows how to use deliberate wrong answers before a topic is even taught.

Sources

  1. Butterfield, B. & Metcalfe, J. (2001). Errors committed with high confidence are hypercorrected. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(6), 1491–1494.
  2. Metcalfe, J. (2017). Learning from errors. Annual Review of Psychology, 68, 465–489.

More from the blog

All posts →