Understanding the Hypercorrection Effect: Why High-Confidence Errors are More Likely to be Corrected

Abstract

The hypercorrection effect refers to the finding that high-confidence errors are more likely to be corrected after feedback than are low-confidence errors (Butterfield & Metcalfe, 2001). In 5 experiments I examine the hypercorrection effect, offer possible explanations for why the effect occurs, and examine the durability of the effect. In Experiment 1, I replicate the hypercorrection effect and show that delaying the feedback does not reduce the effect. In a secondary item analysis I also show that the effect is not caused by "tricky" questions. In Experiments 2 and 3, I show that subjects are more likely to remember the source of the feedback after both high-confidence errors and low-confidence correct responses. This suggests that it is the discrepancy between the subject's expectation and the actual feedback that causes the hypercorrection effect. In Experiment 4 I show that the hypercorrection effect also occurs for episodic false memories showing the diversity of the effect. Finally, in Experiment 5 I examine the durability of the effect. Initial high-confidence errors that are corrected after feedback remain corrected one week later.