Effects of Legal Instructions on Behavioral and Neural Mechanisms of Decision-Making

Abstract

A fundamental question in decision-making is how instructions can be used to influence behavior. Nowhere is this question more important than in the legal system, in which the fate and lives of individuals are in the balance. In the United States, jurors are frequently instructed to apply poorly defined and varied decision thresholds (burdens of proof), as well as instructions to disregard inadmissible evidence. For civil trials, a preponderance of evidence (PoE) is set as the decision threshold, while criminal trials require the more stringent threshold of beyond a reasonable doubt (BaRD). It is unclear, however, how decision threshold and disregard instructions are applied by laypeople and how they compare to participants’ intuitive decisions. To evaluate burden of proof instructions, here we applied quantitative, psychometric analyses to assess and compare decision thresholds across instruction-type in both legal and non-legal contexts. We found a consistent pattern across all legal scenarios, non-legal scenarios, and psychophysical tasks: As expected, naive individuals interpreted PoE less stringently than BaRD, but surprisingly, they interpreted PoE more conservatively than their own intuitive belief (IB) and the legally prescribed threshold. Decision thresholds were also more stringent for legal than any non-legal contexts. Follow-up studies addressed why laypeople's PoE standard is interpreted more stringently than expected, and why adjudication in the legal domain is consistently more conservative than any other domains of decision-making tested. We also examined how domain expertise affected this pattern of results. Furthermore, to understand the application of disregard instructions we built on an existing neuroimaging paradigm to evaluate the brain mechanisms involved in assessing biasing character evidence and admissibility instructions using fMRI. We found that an instruction allowing individuals to consider evidence increased punishment independent of character evidence, and identified several prefrontal regions associated with individual differences in the application of this instruction. Taken together, our findings suggest that legal instructions can influence decisions in ways not intended or accounted for by the law. This may call for a re-assessment of the intended outcomes of legal standards – particularly PoE – as well as the potentially biasing effect of admissibility rulings.