Truth or Doubt? An Empirical Test of Criminal Jury Instructions

Abstract

The Constitution protects us from criminal conviction unless the government can prove guilt beyond a reasonable doubt. However, after defining reasonable doubt, many trial courts will then instruct jurors that “you are not to search for doubt. You are to search for the truth.”

Defendants have repeatedly challenged such truth-related language, arguing that it lowers the government’s burden of proof to a mere preponderance of the evidence. That is, if the government’s version of events is only slightly more credible than the defendant’s, it follows that, in a search for the truth, jurors would be obligated to convict.

Appellate courts concede that instructing jurors to search for the truth is not proper and may lower the government’s burden. However, these same courts refuse to reverse defendants’ convictions because, the courts claim, in the context of the instruction as a whole, jurors are probably not influenced by the truth-related language.

In this Article we empirically test this judicial claim. We recruited 300 participants to serve as mock jurors. Every juror read the same case summary of a hypothetical criminal trial. Jurors were then randomly assigned to one of three groups, with each group receiving a different jury instruction on the government’s burden of proof.

One of the groups was instructed on reasonable doubt, but then told “you are not to search for doubt. You are to search for the truth.” This group’s conviction rate was nearly double that of the group that received a standard reasonable doubt instruction, and was statistically identical to the group that received no reasonable doubt instruction at all.

This finding contradicts the courts’ conclusion that truth-related language has no impact on juror decision-making, and is strong evidence that such language not only lowers, but actually eviscerates, the government’s burden of proof in criminal cases.

Keywords: Criminal Law, Constitutional Law, Jury Instructions, Burden of Proof, Empirical