Damage awards for pain and suffering and punitive damages are notoriously unpredictable. Courts provide minimal, if any, guidance to jurors determining these awards, and apply similarly minimal standards in reviewing them. Lawmakers have enacted crude measures, such as damage caps, aimed at curbing award unpredictability, while ignoring less drastic alternatives that involve guiding jurors with information regarding damage awards in comparable cases (“comparable‐case guidance” or “prior‐award information”). The primary objections to the latter approach are based on the argument that, because prior‐award information uses information regarding awards in distinct cases, it introduces the possibility of biasing the award, or distorting the award size, even if prioraward information reduces the variability of awards. This paper responds to these objections. It reports and interprets the results of a large randomized controlled trial designed to test juror behavior in response to prior‐award information and, specifically, to examine the effects of prioraward information on both variability and bias under a range of conditions related to the foregoing objections. We conclude that there is strong evidence that prior‐award information improves the “accuracy” of awards—that it significantly reduces the variability of awards, and that any introduction of bias, or distortion of award size, is minor relative to its beneficial effect on variability. Furthermore, we conclude that there is evidence that jurors respond to prior‐award information as predicted in recent literature, and in line with the “optimal” use of such information; and that prior‐award information may cause jurors to approach award determinations more thoughtfully or analytically.
*Assistant Professor, SMU Dedman School of Law; Fellow, Harvard Institute for Quantitative Social Science.
** Ph.D. Candidate, Department of Statistics, Harvard University. This paper benefited immensely from the guidance of Professors Donald Rubin, Eric Posner, Michael Saks, Richard Lempert, David Rosenberg, Christopher Robertson, and James Greiner, and from support provided by the Institute for Quantitative Social Science and the Center for American Political Studies.