Regressive Bias

What is regressive bias?

In judgment and memory research, regressive bias describes estimates pulled toward less extreme values: low quantities may be overestimated and high quantities underestimated.

Evidence and Explanation

Fiedler and Unkelbach’s 2014 review explains how uncertainty and extremity can produce regressive judgments. Its examples include estimates of observed frequencies and evaluations of performance.

A person might estimate a rarely presented item as more frequent than it was and a frequently presented item as less frequent. This illustrates compression of the differences between them.

Regressive bias vs regression to the mean

Regressive judgment is not the same as neglecting regression to the mean, where someone expects an extreme outcome to repeat unchanged. Such an insufficiently adjusted forecast is a non-regressive prediction. The two can point in different directions.

Assessments should consider uncertainty, measurement, and the judgment task when explaining regressive estimates.