NATURAL SAMPLING AND BASE-RATE NEGLECT

Using an analogue of the layer-and-engineer item (Kahneman & Tversky, 1973), we compared conditions in which base rates were either presente d as percentages (A), or frequencies (B), to conditions in which the n atural sampling process was described additionally (C) or was directly experienced (D). We expected the likelihood of base-rate utilization to increase as the presentation approaches the process of natural samp ling. Accordingly, results showed that the contingency of judgments on base rates systematically increased across conditions A to D. (C) 199 8 John Wiley & Sons, Ltd.