Hypothesis Testing in the Courtroom
Hypothesis Testing in the Courtroom, in Alan E. Gelfand ed., Hypothesis Testing in the Courtroom, 1987, Orlando, FL: Academic Press, Inc., pp. 331–356
12 Pages Posted: 26 Oct 2018
Date Written: 1986
This paper explains why it is impossible to set a significance level for a statistical hypothesis test that matches the legal burden of persuasion without considering jurors' prior probabilities. In particular, it shows why satisfying a test with equal probabilities of false rejections and false acceptances does not necessarily correspond to the more-probable-than-not standard. It discusses alternative procedures to enable statistical experts to assist courts and jurors to infer the true value of a statistical parameter. It suggests that confidence intervals or likelihoods are preferable to assertions of statistical significance.
Keywords: statistical significance, p-value, likelihood, likelihood ratio, Bayes, burden of persuasion, burden of proof, probative value, hypothesis testing, error probabilities, prior probabiity
JEL Classification: C11, C12
Suggested Citation: Suggested Citation