The 2009 Report of the National Academy of Sciences (NAS) on the state of forensic science in the American criminal justice system has fundamentally altered the landscape for scientific evidence in the criminal process, and is now setting the terms for the future of forensic science reform and practice. But the accomplishments of the Report must not obscure the vast terrain that remains untouched by the path of reform that it charts. This Article aims to illuminate a critical and currently neglected feature of that territory: namely, the manner in which police and prosecutors, as upstream users of forensic science, select priorities, initiate investigations, collect and submit evidence, choose investigative techniques, and charge and plead cases in ways that have critical and systematic, though poorly understood, influences on the accuracy of forensic analysis and the integrity of its application in criminal cases. By broadening our understanding of how forensic science is created and used in criminal cases-by adopting a systemic perspective-the Article points to a raftof yet unaddressed issues concerning the meaning of scientific integrity and reliability in the context of investigative decisions that are by and large committed to the discretion of decidedly unscientific actors. Critically, the Article demonstrates that systemic dynamics affecting upstream use of forensic science might well undermine the reliability-enhancing goals of the reforms advocated by the National Academy Report. As the NAS Report begins to set the agenda for active conversations around legislative and executive action to reform forensic science, it is critical to consider these questions. Moreover, the Article suggests that the embrace of science as a unique evidentiary contributor within the criminal justice system problematizes some of the bedrock assumptions of American criminal procedure that have, to date, prevented more robust doctrinal intervention in the investigative stages and decisions that the Article explores.
Over the course of the last half century, science has moved from the periphery to a place of prominence in the investigation and prosecution of crime. Analysis of physical evidence, particularly with recent scientific and technological advances in the arena of DNA, has been embraced as advancing the fundamental epistemic goals of the criminal justice system by enhancing society's ability to connect the guilty with their misdeeds and, even more powerfully, enabling exculpation of the innocent. As the reliability of more traditional investigative tools such as eyewitness identifications and confessions has been increasingly scrutinized, the comparative accuracy of scientific evidence has been hailed.1 Yet at the same time, news headlines continually reveal laboratory- or analyst-level breakdowns, and many of the hundreds of exonerations seen in recent decades-through DNA testing or otherwise-have exposed error or outright fraud committed under the guise of "scientific" opinion.2 A significant body of critical academic commentary on the forensic science field strongly suggests structural rather than individual causes of these ills: the surprisingly thin research base for many forensic methodologies;3 systematic compromises to the quality of crime laboratory output due to under-resourcing and the undue influence of police and prosecutorial agendas on scientific analysis;4 poor access to, and use of, expert resources by defense counsel;5 and lax scrutiny of scientific evidence by courts.6
Thus, forensic evidence is both special and mundane. It is special in its potential to identify and exclude with a degree of reliability that sets it apart from more traditional forms of proof in criminal investigations (eyewitness identification, confessions, informants, and the like). But it is also, like all evidence produced by humans in the crucible of the criminal justice system, susceptible to error, bias, manipulation, rationing, and other dynamics that compromise its reliability both in theory and in practice. …