Naked Statistical Evidence of Liability: Is Subjective Probability Enough?

Five studies tested the idea that people are reluctant to make proplaintiff liability decisions when the plaintiffs evidence is based on naked statistical evidence alone. Students (n = 740) and experienced trial judges (M = 111) averaged fewer than 10% affirmative decisions of liability when a case was based on naked statistical evidence but averaged over 65% affirmative decisions based on other forms of evidence even though the mathematical and subjective probabilities were the same for both types of evidence. Numerous hypotheses, including causal relevance, linkage to the specific case, and fairness to the defendant proved inadequate to explain the data. For evidence to affect decisions, the evidence must do more than affect people's perceptions about the probabilities associated with the ultimate fact; people seem to require that suppositions regarding the ultimate fact affect their perceptions of the truth or falsity of the evidence.