Research School Network: The hypercorrection effect, spaced practice and remote learning

What is the hypercorrection effect?

Have you ever been so certain about a fact that when you found out you were wrong the shock caused you to remember the right answer forever? As a science teacher I see this quite a lot. Fr example, living on the coast a lot of students think the sky is blue because it reflects the sea. The hypercorrection effect was the name given to this phenomenon by Janet Metcalfe and Brady Butterfield in 2001[1]. For the cog sci nerds out there, it appears that hypercorrection is distinct from metacognitive ability, so all of us do it even if we are unable to verbalised our level of confidence[2]. There is however an indication that young adults display a more pronounced hypercorrection effect than older adults[3].

Interesting for teachers is the effect of spaced retrieval on the newly corrected answer

Work by Andrew Butler and the team at Duke university appear to show that the hypercorrection effect persists for about a week, but longer than this and the ​‘high-confidence error’ (misconception) will persist.[4]

What does this mean for teachers?

This is all very fascinating but what does it mean to teachers on a day-to-day basis?

I think one of the key things this work tells us is the importance of the Teach Like A Champion (TLAC) technique called ​‘Right is right’. Often students confidently give answers which are decent but lacking in a particular key detail. This is then often ​‘rounded up’ by the teacher. When the teacher repeats the answer, they pepper in the missing information, so the class hears a perfect answer. The problem is that the student thinks that is the same as their original answer. This robs them of the hypercorrection effect. They now think that their answer was correct and will hold onto their ​‘high-confidence’ almost right answer as fact.

The second implication for teaching I think this demonstrates is the importance of revisiting these corrected errors often. We all know that knowledge fades if not revisited regularly and it appears the hypercorrection effect does not override this foundational concept of memory. As teachers we could revisit this work a number of ways. We could use a variation of another TLAC technique ​‘No Opt Out’. Here, when we encounter a student’s wrong answer we move to another student and when we receive the right answer we return to the original student and get them to repeat it. We could also use our regular retrieval practice routine to regularly check student’s ability to recall the information. Asking a student who got it wrong previously to check they are now getting it right. Another way we can do this is to use technique I like to call ​‘This answer is sponsored by..’ In this case the student who has the high-confidence error becomes the classes official sponsor of that answer. For example, let us assume it’s a definition. Every time the word is used in class the student is asked to define it. They get multiple regular chances to recall it and it is a little bit of fun when used with warmth.

What about teaching remotely?

Remote teaching is hard. I never realised just how much I gathered data from circulating a class until this year. The good thing about the hypercorrection effect is that it occurs weather the student wants it to or not. Providing they are paying attention to the feedback; it does not matter if they tell you their answer or it is just in their head. So, if you are teaching remotely in the next few weeks take heart in the fact that the hypercorrection effect will be on your side even if you can’t check every student answer.

By Adam Robbins, Durrington Research School ELE

References

[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. doi:10.1037/0278 – 7393.27.6.1491.

[2] Williams, David M; Bergström, Zara; Grainger, Catherine (2016−12−15). ​“Metacognitive monitoring and the hypercorrection effect in autism and the general population: Relation to autism(-like) traits and mindreading” (PDF). Autism. 22 (3): 259 – 270. doi:10.1177/1362361316680178. ISSN 1362 – 3613. PMID 29671645.

[3] Eich, Teal S.; Stern, Yaakov; Metcalfe, Janet (2013). ​“The hypercorrection effect in younger and older adults”. Aging, Neuropsychology, and Cognition. 20 (5): 511 – 521. doi:10.1080/13825585.2012.754399. PMC 3604148. PMID 23241028.

[4] Butler, Andrew C.; Fazio, Lisa K.; Marsh, Elizabeth J. (2011−12−01). ​“The hypercorrection effect persists over a week, but high-confidence errors return”. Psychonomic Bulletin & Review. 18 (6): 1238 – 1244. doi:10.3758/s13423-011‑0173‑y. ISSN 1531 – 5320. PMID 21989771.