When Is Statistical Evidence Superior to Anecdotal Evidence in Supporting Probability Claims?
Under certain conditions, statistical evidence is more persuasive than anecdotal evidence in supporting a claim about the probability that a certain event will occur. In three experiments, it is shown that the type of argument is an important condition in this respect. If the evidence is part of an argument by generalization, statistical evidence is more persuasive compared with anecdotal evidence (Experiments 1 and 2). In the case of argument by analogy, statistical and anecdotal evidences are equally persuasive (Experiments 2 and 3). However, if the case in the anecdotal evidence is dissimilar from the case in the claim, statistical evidence is again more persuasive (Experiment 3). The implications of these results for the concept of argument quality are discussed.