Separating the wheat from the chaff: does discriminating between diagnostic and nondiagnostic information eliminate the dilution effect?

The dilution effect refers to the finding that judgments are often unduly influenced by nondiagnostic information, producing regressive judgment. Because the dilution effect is a problem in various domains, strategies to control the impact of nondiagnostic information were explored by drawing on a perceptual and a conversational account of the dilution effect. Three experiments ( n  = 259) demonstrate that explicit instructions to discriminate between diagnostic and nondiagnostic information did not reduce the dilution effect. Rather, consistent with a perceptual explanation but not consistent with a conversational explanation, the dilution effect disappeared only when participants engage in perceptual control, that is, when they actively remove nondiagnostic pieces of information before making a judgment. Copyright © 2004 John Wiley & Sons, Ltd.