The 1990s are associated with keywords like new economy and information society. The increased importance of information technologies underlying these phenomena was due primarily to the transformation of the Internet into a medium of mass communication, but the rising number of 24/7 TV stations and the popularity of cellular phones contributed as well. As one side effect of these developments, people in the early twenty-first century are confronted with an unprecedented flood of statistical information. Statistics on almost any topic are available anywhere at any time, whether on the latest unemployment rates, recent trends in climate change, or the odds ratio that a football match will end in a draw when the away team leads by two goals at halftime. At least to anyone trained in statistics, however, the question presents itself whether people have the skills necessary to deal with the available information critically. After all, citizens, politicians, consumers, and managers still have to decide on their own whether to trust a statistic and the conclusions derived from it.

A glance at the major findings accumulated by the heuristics and biases program during the past decades suffices to substantiate a skeptic's position. People have been shown, among other things, to prefer individuating information to base-rate information when estimating conditional probabilities (Bar-Hillel, 1980; Kahneman & Tversky, 1972), to collapse data over levels of significant third variables (Schaller, 1992a, b), and to confuse the conditional probability of a criterion event (e.g., becoming a drug addict) given a predictor event (e.g., having consumed marijuana) with its inverse (Evans, 1989; Sedlmeier, 1999).