Overconfidence in ignorant experts

Abstract

When asked true-false questions that were too difficult to be answered correctly at better than a chance frequency, subjects often guessed in preference to admitting ignorance and often expressed certainty in the correctness of their answers. These tendencies were strongly correlated with the degree of expertise the subject believed himself to have in the area in which the question was asked. Under certain presumably ego-threatening conditions, subjects contradicted by answering "false" appreciably more often than they agreed by answering "true, " although half the questions were, in fact, true and half were false.