USING BURDENS OF PROOF TO ALLOCATE THE RISK OF ERROR WHEN ASSESSING DEVELOPMENTAL MATURITY OF YOUTHFUL OFFENDERS.

• Author: Faigman, David L.
• Position: Imagining the Future of Law and Neuroscience

TABLE OF CONTENTS

Introduction. 1292 I. The Group to Individual Problem 1293 II. Sentencing Youthful Offenders Under the Eighth Amendment. 1297 III. Burdens of Proof and Developmental Maturity . 1305 IV. Employing Burdens of Proof for Sentencing Youthful Offenders. 1310 V. The Sentencing Decision. 1314 Conclusion. 1315 INTRODUCTION

In Roper v. Simmons, the United States Supreme Court held that the Eighth Amendment prohibited executing youthful offenders who committed capital crimes before the age of eighteen due to the inherent immaturity of adolescents relative to adults. (1) Although the Court recognized that this categorical rule was both over- and under-inclusive because some offenders younger than eighteen are considered developmentally mature and some over that age are not, the difficulty lay in identifying which were which. (2) This is a classic example of a core challenge that lies at the intersection of law and science. Whereas scientists primarily research phenomena at the population or group level, courts typically seek to answer whether a particular case is an instance of some phenomenon. (3) This is referred to as the "group to individual" (G2i) problem, an issue of statistical inference that plagues most uses of scientific research in court. (4) Although the problem appears in virtually all situations in which scientific research is used to inform individual courtroom decisions, there is no one approach that can be used across legal contexts to manage it. (5) In this Article, we consider an approach to meeting the G2i problem in the context of sentencing youthful offenders in which the constitutional rule was set at the group level--age of offender at time of offense--but its application inevitably must be done at the individual level. We propose using burdens of proof, a classic mechanism for balancing the cost of making errors under conditions of uncertainty, to best manage the G2i problem in the sentencing of youthful offenders. (6)

We begin in Part I with a brief background on the limitations posed by the group to individual problem and an introduction to how it applies to sentencing youthful offenders. Parts II, III, and IV focus exclusively on the law and science surrounding the issue of assessing developmental maturity in different age cohorts. Part V returns to the question of sentencing in light of the fear of "permanent incorrigibility."

1. THE GROUP TO INDIVIDUAL PROBLEM

   The law has largely failed to recognize the fundamental disconnect between the usual level of study in science and what courts typically expect when they employ scientific evidence to decide cases. Although courts decide cases based on the individual or situation at hand, scientists primarily conduct studies at the group level.' Unfortunately, group observations rarely apply universally to their individual members, meaning that group- or populationlevel findings may only provide weak support for individual determinations. (8) For example, scientists may have considerable confidence that cross-racial identifications are less accurate than same-race identifications or that the chemical Benzene causes leukemia, but this research cannot show with confidence that a particular identification is accurate or that Benzene caused a particular person's leukemia.

   In the eyewitness accuracy example, researchers study accuracy rates between groups of eyewitnesses. A particularly robust finding in this literature comes from comparisons of same-race identifications to cross-race identifications. Typically in this research, one group of subjects is given a scenario that requires same-race identifications and another group different-race identifications. (9) This research finds that, on average, same-race identifications are more accurate than different-race identifications. (10) However, nothing in the research literature indicates that an eyewitness researcher could reliably determine whether a particular cross-race identification was accurate. (11) As a consequence of this inherent limitation of the research, eyewitness experts ordinarily are permitted to testify only to the general phenomenon and do not offer a clinical opinion regarding an individual case. (12)

   Eyewitness identification experts generally conform their testimony to how most scientists would approach application of their findings. Scientists ordinarily describe their research probabilistically and apply it to some group or population. (13) Any application of the findings to an individual is usually done probabilistically as well. (14) In forecasting whether it will rain this afternoon, for instance, meteorologists estimate the likelihood it will rain based on models built from group data and speak to individual cases using likelihood estimates. (15) Similarly, when doctors provide informed consent regarding whether a drug causes side effects, the information is provided statistically, not as a statement of scientific certainty. In other words, scientists ordinarily retain their group perspective even when describing individual cases.

   In the example of Benzene, however, courts expect a different approach to the scientific evidence, despite the similar limitations of the underlying science. In the area of medical causation, typically arising in cases involving toxic torts or medical malpractice, courts insist on expert testimony regarding the individual case. (16) The issue of "general causation," that is, can Benzene cause leukemia, remains a threshold issue that must be supported by valid research. (17) However, courts also require proof of "specific causation" in such cases--in other words, sufficient proof that Benzene caused the plaintiff's leukemia. (18)

   Over the years, courts and the experts that appear before them in cases involving disputes over medical causation have jerry-rigged a method to overcome the disconnect between what science can confidently do and what the law demands of it. This method is referred to as "differential etiology," and essentially calls on clinical judgment to "rule-in" the plaintiff's claimed cause of the injury and "rule-out" all alternative causes. (19) This process of logical deduction is largely ill-defined, (20) yet inevitably leads to statements from experts that the plaintiff's condition was caused by a particular substance "to a reasonable degree of [scientific] certainty." (21) Such conclusions appear to be little more than oft-repeated mantras that have little basis in either science or law. (22)

   Hence, although science almost invariably operates at the group level, its use in court depends on the legal framework in which it is being fitted. Returning to the above examples, research on eyewitness identification is relevant and admissible at the group level because it educates the fact-finder regarding factors that might interfere with accuracy. (23) Research on Benzene is also relevant and admissible at the group level, but courts have deemed it necessary that experts opine on the individual case as well, even if such opinions have little empirical support. (24)

   The lesson here is that, despite the limitations inherent in the G2i problem, the law determines the necessary level at which expert testimony must be presented--group or group and individual. (25) Although different legal or scientific contexts might demand different kinds of presentation, the goal for the law should be to resolve the G2i problem in ways that are conducive to obtaining the most valid scientific opinions possible while also serving the practical demands involved in trying individual cases. Arguably, this is what has occurred in both the eyewitness and Benzene examples, albeit in different ways. In the former, individualized expert opinions are not necessary for the fact-finders to do their jobs. (26) The insight that eyewitnesses are not as accurate under certain specific conditions as perhaps generally supposed is helpful to the fact-finder. (27) In the latter, fact-finders would be lost without expert assistance in applying the general science to an individual case, even if its reliability is uncertain. (28) Most fact-finders would have considerable difficulty reasoning from general toxicological and epidemiological evidence to determine causation in a specific case. Experts are presumed to be needed, and they are thought to be capable of making this inferential deduction. (29)

   The situation of sentencing youthful offenders is something of a blend of these two examples. First, as with all applied science, the G2i problem is inherent in the issue of developmental maturity. (30) Second, similar to eyewitness identification, group data about the relative maturity of different age cohorts is relevant for judges and fact-finders in considering appropriate penalties. (31) Third, similar to the Benzene example, sentencing of a youthful offender is inherently an individualized determination, one in which fact-finders might benefit from guidance regarding how to apply the general research literature to individual cases. (32) Finally, as the next Part details, unlike many interactions between law and science, the Supreme Court has explicitly recognized the presence of the G2i issue in sentencing youthful offenders, though the Court has yet to identify an appropriate method to manage it.

2. SENTENCING YOUTHFUL OFFENDERS UNDER THE EIGHTH AMENDMENT

   In a series of cases, the Supreme Court has limited courts' ability to impose extreme sentences on juvenile and youthful offenders. The Court has repeatedly recognized that children are different from adults, so subjecting them to the same extreme punishments is inconsistent with Eighth Amendment principles. (33) In Roper u. Simmons, the Court held that executing minors under the death penalty is "cruel and unusual" punishment prohibited by the Constitution. (34) In Graham v. Florida, the Court held that the same Eighth Amendment clause prohibited a juvenile offender from being sentenced to life...