An experimental exploration of reasonable doubt

Abstract

The “beyond a reasonable doubt” standard is a constitutional requirement in U.S. criminal cases, but jury instructions on “reasonable doubt” vary across jurisdictions. We use a controlled experiment to analyze the relationship between the definition of reasonable doubt and juror decisions. In our novel (pre-registered) experiment, we vary the definition of reasonable doubt between subjects and elicit the level of evidence required for subjects to convict a defendant. We analyze juror decisions under two state definitions that are markedly different (Wisconsin and West Virginia) and analyze juror decisions when reasonable doubt is not explicitly defined. We find similar behavior in each treatment. We ran three additional treatments to determine why behavior does not seem to vary across definitions. Our data is consistent with subjects having pre-conceived notions of reasonable doubt that are not affected by jury instructions.