Debiasing

Abstract
Once judgmental biases are identified, researchers start trying to eliminate them using one of two strategies. The first accepts the existence of the bias and concentrates on devising schemes, such as training programs, that will reduce it. The second claims that the bias is not very robust and important and concentrates on devising experimental situations in which it will not appear. Together, these strategies circumscribe the universe of debiasing procedures. The present paper attempts to describe the specific kinds of procedures that can (and have) emerged from these strategies. It then applies this list of 25 procedures to a reanalysis of debiasing studies relating to two judgmental biases, hindsight bias and overconfidence. The ten published articles on hindsight bias have attempted 14 of the 25 procedures and found none of them effective except restructuring the task so as to highlight reasons why a reported event might not have happened. One of another of the 40 papers reviewed dealing with overconfidence has tried most of the possible debiasing procedures. Here, too, most proved ineffective except for task restructuring that highlights the existence of reasons why an answer that appears right might be wrong and training with individualized feedback. Although similar reanalyses need to be done on debiasing attempts applied to other biases here, the picture that emerges here allows some speculations about the robustness of judgmental biases and the most profitable ways to study them.

This publication has 0 references indexed in Scilit: