ANCHORING IN THE INTERPRETATION OF PROBABILITY-EXPRESSIONS

Prior research has demonstrated that people's interpretations of nonnu merical probability expressions (e.g. slight chance, probably) are sen sitive to contextual factors such as event base rate. The present stud y examines the contextual role that numerical anchors may play in inte rpretation of these expressions. Participants read a series of vignett es describing fictitious doctor-patient communications in which the do ctor used a nonnumerical probability expression to describe the likeli hood of an illness given certain symptoms. Participants interpreted th is expression (i.e. made a percentage probability estimate of the illn ess likelihood) in the context of a numerical anchor which they were l ed to believe was arbitrarily chosen. Despite the ostensible arbitrari ness of the numerical anchors, interpretations of the probability expr essions were systematically influenced by anchor value. These results are consistent with previous demonstrations of the 'anchoring-and-adju stment' heuristic and offer further evidence of the context sensitivit y of probability expressions.