In this article I discuss the fundamental relation between people's ability to do induction and their beliefs about randomness or noise, and I illustrate the special difficulties that psychologists face when they try to evaluate the rationality of these beliefs. The presentation is divided into four sections. The first describes the traditional experimental approach to evaluating people's conceptions of randomness and summarizes the data that have been taken to support the conclusion that people have a very poor conception of randomness. The second contrasts the relatively narrow conception of randomness that one finds in philosophical and mathematical treatments of the topic. The third outlines some benefits for psychologists to be gained from thinking about induction as a problem in signal detection., and the fourth presents the argument that any adequate evaluation of ordinary people's conceptions of randomness must consider the role that these conceptions play an inductive inference, that is, in distinguishing between random and nonrandom events. Originally appeared in the Journal of Experimental Psychology; Learning, Memory, and Cognition, 1982, 8 (6), 626 - 636
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education