PUBLISHED:August 16, 2021

Garrett’s Autopsy of a Crime Lab illuminates the flaws in forensic science


“Autopsy of a Crime Lab” analyzes how systemic failures in forensics compromise evidence and have called into question untold numbers of convictions.

Autopsy of a Crime Lab

Nearly 20 years ago, as a civil rights fellow at the famed law firm Cochran, Neufeld & Scheck (now NSB Civil Rights), Brandon Garrett was struck by the role that flawed forensic evidence played in many of his clients’ wrongful convictions.

That experience is one reason why Garrett, now the L. Neil Williams, Jr. Professor of Law at Duke, has made preventing wrongful convictions through research on accuracy and transparency in forensic science a priority for the Wilson Center for Science and Justice, which he directs. While DNA evidence has helped exonerate more than 350 people since it was first used in 1989, more than half were convicted on the basis of fingerprint, bite mark, blood spatter, or other forensic analysis presented at trial, according to a database Garrett maintains.

In his latest book, Autopsy of a Crime Lab: Exposing the Flaws in Forensics (UC Press, March 2021), Garrett conducts a critical review of how forensics are used in criminal cases, detailing the many ways that evidence relied upon by courts and juries can be – and frequently is – compromised on its way from the crime scene to the lab to the courtroom. They include an overreliance on forensic analysis methods that lack scientific validity or have been disproved altogether; the lack of federal regulation of crime labs; and the allowing of trial testimony by “expert” witnesses with unvetted proficiency, undisclosed bias or conflicts of interest, and claims of “100 percent certainty” in results from methods with troubling error rates not disclosed to jurors. In most criminal cases, this evidence is presented by prosecution witnesses and goes unchallenged by under-resourced defense teams – if the case goes to trial at all.

The book illuminates the failure of the forensic profession to self-regulate, and the toll that widespread problems at crime labs and with forensic disciplines themselves have taken on the integrity of the criminal legal system. Garrett also proposes a comprehensive reform agenda that includes adequate funding, independence from law enforcement, and the establishment of quality control measures at the nation’s virtually unregulated network of crime labs.

“Stubborn resistance to criticism and hostility to scientific research is the very antithesis of good science,” Garrett writes. “What leading scientists keep telling us is that, apart from the DNA area, no forensic techniques have undergone sufficiently rigorous testing. Error rates have been unknown ... Simply put, we need to bring good science to forensics.”

At a Wilson Center event earlier this year, Jennifer Mnookin, dean of UCLA School of Law and founding co-director of its Program on Understanding Law, Science and Evidence, praised the book’s “synthetic” quality.

“It reviews the entire range of issues around crime labs and some of the challenges around them. It really gives us the whole soup-to-nuts range of engagements,” Mnookin said.

“And in so doing, it tells a really powerful story of how this is a systemic set of problems. It’s not just an issue in one area but, from beginning to end, forensic science doesn’t appear to be operating with a serious focus on making sure we’re doing all that is reasonably possible to ensure both accuracy and validity, and to ensure transparency about what we know and don’t know.”

Origins of crime labs are in law enforcement, not science

In Autopsy of a Crime Lab, Garrett traces the origins of crime labs and shows how their beginnings inform the way many still operate today. Crime labs didn’t emerge from the scientific community but as “cop shops” within law enforcement, staffed by police officers.

According to the book, when J. Edgar Hoover established the FBI’s Technical Crime Laboratory in 1932, only a few of the big-city police departments dealing with a wave of Prohibition-era gangster activity had evidence-processing divisions. Today there are more than 400 publicly funded crime labs in the U.S., the majority of which are still part of law enforcement agencies.

“When you get your funding from police you see yourself as an arm of law enforcement to help solve crimes,” Garrett points out.

The “war on drugs,” a term coined during the Nixon administration, spurred an expansion of crime labs, and today drug-testing is still their bread and butter, accounting for about half the work of a typical lab. The other half involves linking evidence to individuals through firearm and fingerprint analysis, and disciplines like hair and handwriting comparison. DNA analysis, despite its high profile, makes up only 4% to 5% of most labs’ work, yet three-quarters of the $200 million in federal grants available to crime labs is earmarked for it.

For all the scientific gloss on TV shows like CSI, there are no federal regulations holding crime labs to scientific or quality standards. In fact, clinical laboratories diagnosing minor illnesses like strep throat are more scrutinized – by many orders of magnitude – than crime labs processing evidence that could help put a person on death row, Garrett notes. Clinical laboratories are regulated under a system administered by the Food and Drug Administration, the Centers for Disease Control and Prevention, and the Centers for Medicare and Medicaid Services, while no federal agency holds crime labs to account.

“It just shows how much we care about people charged with crimes versus people asking for a medical test,” Garrett says.

Lack of oversight has led to scores of failures at crime labs across the nation, covering the gamut of forensic disciplines. In Massachusetts, misconduct by two analysts at state drug labs dating back to 2003 has affected close to 100,000 cases, many of which have been vacated. More recently, Washington, D.C.’s Department of Forensic Sciences’ accreditation was suspended over a ballistics error in a murder case and the emergence of broader issues at the lab. And tainting, falsifying reports, improper handling, and other misconduct involving testing at crime labs from New Jersey and Florida to California and Oregon have resulted in the reopening and vacating of tens of thousands of convictions.

“I have documented over 130 crime lab scandals, involving errors or audits of multiple cases, at labs across the country,” Garrett writes. “Hardly a month goes by that I do not find more labs to add to the list.”

Along with the growing tally of costs to government related to investigation, retrial, and remediation, Garrett illustrates the human costs, from fines and incarceration to death, of wrongful convictions based on flawed forensics.

Autopsy of a Crime Lab opens with the high-profile case of Brandon Mayfield, an Oregon attorney whose fingerprints were erroneously matched by three experienced FBI examiners to evidence from the 2004 Madrid terrorist bombing – despite Mayfield’s never having been to Spain. While the federal government issued a formal apology and $2 million settlement, and the FBI made changes to the way it performs fingerprint comparisons, Mayfield has suffered extended emotional and reputational damage from two and a half years under suspicion for terrorism.

Garrett uses examples from his earlier book, Convicting the Innocent: Where Criminal Prosecutions Go Wrong (Harvard University Press, 2012), which analyzes the first 250 people exonerated by DNA, to illustrate the many ways forensics can go wrong and the suffering caused to its victims. Among them are Keith Harward, who spent 33 years in prison partly on the basis of bite mark identification, a widely discredited technique with a high error rate, before being exonerated by DNA. Harward spoke about his case, the lingering effects of his long incarceration, and his struggle to rebuild his life during an appearance at Duke Law in March 2019.

“I’ll never be free. This is something I have to live with the rest of my life,” Harward said, choking up and drawing tears from audience members. “I get emotional about every little thing. Because unlike most people, I see things and witness things that y’all pass right by.”

The reform agenda: a way forward

Autopsy of a Crime Lab also documents how the scientific community and concerned government leaders have worked to champion forensic reforms. In 2009, the National Academy of Science issued a landmark report, Strengthening Forensic Science in the United States: A Path Forward, that cast light on the unscientific methods and foundations of many forensic disciplines and called for major reforms and substantial research to validate forensic evidence and techniques. The report cited Garrett’s article “Judging Innocence” 108 Columbia Law Review 55-142 (2008), an empirical study of the evidence presented at the trials of DNA exonerees.

More recently, a 2016 report from the President’s Council of Advisors on Science and Technology (PCAST), focusing on problems in forensic techniques including firearms, fingerprints, and bite mark comparisons, concluded that the use of unscientific forensics should be discontinued until reliability studies were completed.

Scientific criticism of the forensics profession has been met with resistance and outright hostility from its practitioners, but progress has been made since the creation in 2015 of the Center for Statistics and Applications in Forensic Evidence (CSAFE), funded by the National Institute of Standards and Technology, which sponsors data-driven research on forensic evidence at universities and works to build relationships between scientists and practitioners.

Garrett is a member of the CSAFE leadership team, and the projects it has sponsored at the Wilson Center have focused on how forensic evidence is used in courts, including mock trials and training for lawyers and research on how jurors evaluate forensic evidence and error rate information. In 2019 Duke Law hosted a CSAFE conference, “Getting Forensics Right,” for statisticians, lawyers, and forensic practitioners.

The growing body of scientific research on forensic research and high-profile lab failures are also giving momentum to the reform movement. After multiple problems at the Houston Police Department Crime Laboratory, the city severed its forensics divisions from law enforcement and created the Houston Forensic Science Center with an independent laboratory, comprehensive quality measures, and a citizen oversight board based on recommendations in the 2009 NAS report. It is now considered a national model. In a May op-ed in Slate, Garrett and the Houston lab’s current president and CEO called for the allocation of more federal funding to the nation’s crime labs, severing them from law enforcement, and instituting quality control measures including blind proficiency testing as part of a broader policing reform agenda.

“Many of the key reforms that Brandon outlines in Autopsy of a Crime Lab are part of my own wish list, and many of them would bring forensics in line with modern scientific practice, which is long overdue,” said Ed Cheng, the Hess Chair in Law at Vanderbilt Law School, at the Wilson Center event, calling it “a timely and important book.”

“I too would like to see traditional match determinations replaced by probabilities and error rates. I’d like to see the further use of blind testing, where analysts don’t know what the desired result is, and in fact whether or not the sample is even real or a test sample. And I’d also like to see a lot of the quality controls and standards that we’ve come to expect of our medical laboratories replicated in crime laboratories.”

In the book Garrett also calls for reforms in the way forensic evidence is presented in court. Jurors have long assumed infallibility of evidence, he says, both from experts long having been allowed to declare “100 percent certainty” and “zero error” in their testimony, and in popular culture, where TV shows have presented misleading depictions of how labs actually work and the reliability of forensic techniques.

In “Forensics and Fallibility: Comparing the Views of Lawyers and Jurors,” 119 West Virginia Law Review 621-637 (2016), Garrett and Gregory Mitchell found that most jurors bring to the courtroom strong confidence in the reliability of forensic evidence. Robust defense is needed to counter these perceptions, he says, yet in the vast majority of criminal cases defense counsel doesn’t question the prosecution’s forensic experts or call its own experts, meaning error rates go undisclosed and the expert’s own proficiency at interpreting the evidence goes unchallenged. In Garrett’s 2013 paper “Invalid Forensic Science Testimony and Wrongful Convictions,” 95 Virginia Law Review 1, he and co-author Peter J. Neufeld found that the prosecution called forensic analysts at trial in 156 of the 232 post-conviction DNA exonerations they studied, and the expert presented invalid testimony in 60% of those cases. The defense retained an expert in only 19 cases. More funding for defense experts is urgently needed to counter prosecution experts, Garrett says.

Judges bear responsibility too, for failing to disclose to jurors the limits and reliability of forensic evidence and allowing in disproved forensic evidence under precedent rather than applying the Daubert Standard to scientific testimony. Judges must “rethink their role as gatekeepers” in reviewing the reliability of evidence and expert witnesses allowed in court, he concludes.

The legal community also should work to improve scientific literacy among lawyers by offering more forensics education, both as continuing education and, as the 2009 NAS paper recommended, in the JD curriculum. In Garrett’s recent paper “Forensic Science in Legal Education,” Duke Law School Public Law & Legal Theory Series No. 2021-22, he and co-authors Glinda Cooper of Yeshiva University and Wilson Center research assistant Quinn Becker T’22 surveyed 192 U.S. law schools and found only 42 courses offered in forensic sciences, many offered only intermittently and as an introductory course without a focus on statistics or quantitative or scientific methods.

Garrett is working to increase the number of forensic science-related offerings at Duke Law to empower students interested in the criminal bar. One such course is Amicus Lab, taught in Fall 2019 by Garrett and Robinson O. Everett Professor of Law Nita Farahany, in which students write and submit briefs in appellate cases on issues regarding scientific methods. In Autopsy of a Crime Lab, Garrett describes the case of Joe Bryan, a Texas man who served 30 years for murder based on blood spatter analysis, a scientifically unsupported forensic method that was performed by an inexperienced police officer from evidence at a contaminated crime scene. In April 2019 Bryan was denied parole for the seventh time.

But thanks in part to Amicus Lab students, that wasn’t the final word.

“Afterwards, my law students and I filed a brief in the Texas Court of Criminal Appeals in support of Bryan’s post-conviction motion and detailing the problems with the evidence,” Garrett recounts.

“The judges dismissed the motion without any explanation. In March 2020, however, Texas officials finally granted Bryan parole, and he was released.”