The Study: Do Jurors Rule Based on Standard of Proof?

No, they don’t. Not really, or at least not when it matters most. That is the conclusion of Florida State University behavioral economist Brent Davis in this recent study. The results do conflict with some research showing very small differences in the right directions (for example Glöckner & Engel, 2013). But Davis’ method was unique in several respects. Most importantly, his study represents the first true incentivized simulation of jurors’ decision making under varying standards of proof. What that means is detailed herein the study’s write-up.

Results

Not much difference. The assumption is that raising the standard of proof raises the threshold for a decision. But in the simulation, asking participants to apply a reasonable doubt or a preponderance standard resulted in only a “mild response” which was not statistically significant overall.

Only matters in low consequence situations. The exception applied to situations where the consequences of a win or a loss were low. “There were significant differences in conviction rates between reasonable doubt and preponderance of the evidence,” Davis noted, “however, this only occurs in the low consequences treatment.” That seems counterintuitive, of course, but the author reasons that when consequences are low, the decision makers may be less sensitized to the consequences of conviction, and thus more likely to reconcile a consistency between their beliefs about probability of guilt and the legal standard they’re asked to follow. In any case, when the consequences are higher — as they would be in real legal cases — the application of different standards of proof does not create a reliable difference in decision making.

No clear threshold. The point of articulating a standard is to identify a definite threshold or tipping point in the decision. That threshold, however, was not found in the data when plotting the percentage found guilty against a juror’s zero to one hundred estimate of likelihood of guilt. “Jurors do no have a universal threshold for guilt amongst preponderance of evidence or reasonable doubt,” the author concluded, “but instead have a threshold individual to each case.”