Photograph: Victoria Janicke/thisisreno

Cops in Lab Coats: Curbing Wrongful Convictions through Independent Forensic Laboratories
Sandra Guerra Thompson, Carolina Academic Press, $37 (paper)

Inside the Cell: The Dark Side of Forensic DNA
Erin E. Murphy, Nation Books, $27.99 (cloth)

Unfair: The New Science of Criminal Injustice
Adam Benforado, Broadway Books, $17 (paper)

“Scientific evidence really nails this man to the wall,” the Harris County, Texas, prosecutor said at the trial of George Rodriguez. The alternative suspect identified by the defense, a man named Isidro Yanez, “could not have committed the offense.” It was beyond doubt. The jury convicted Rodriguez, and he was sentenced to sixty years in prison.

Two decades later the faulty scientific evidence in his case would help bring down the entire Houston police crime laboratory and call into question the very model of the modern crime lab: a relatively new institution that has fundamentally changed how science is used in American courtrooms. Heroic depictions of crime labs in mass media, and in ever-popular forensic dramas from CSI to Crime Scene to Bones, attract millions of viewers, but, as Rodriguez’s story shows, reality is far less appealing.

That story began in 1987, when a fourteen-year-old victim reported being abducted and raped by two Latino men. Detectives focused on two brothers who lived at a house described by the victim, as well as on Yanez, whom the brothers knew and the victim identified. The victim later changed her identification to George Rodriguez, an acquaintance of the brothers, based on “the way he stood” in a suggestive one-on-one “show up” procedure, whereby police show a witness only one suspect. Police also showed her photos of Rodriguez two additional times, and an initial non-identification turned into a hesitant and then a more confident identification.

Still, Rodriguez had a solid alibi. His boss testified that he was at work for the entire day of the assault. But then the Houston Police Department Crime Lab delivered a powerful rejoinder.

One of their analysts used a microscope to compare Rodriguez’s hair with a single strand located on the victim’s panties and found them “consistent.” The analyst also tested swabs from a rape kit for the familiar A, B, and O blood types and concluded that Rodriguez could have left the stains. “One would predict his genetics would show up as a donor in a sexual assault or intercourse,” the analyst said—though DNA technology was then in its infancy, and no test could confirm or refute the conjecture. The analyst also said Yanez was definitively excluded by the blood analysis, but this was flat-out wrong. The analyst should have told the jury that, while Rodriguez could be included, so could just about any male in the population, and Yanez certainly could not be excluded. Rodriguez’s hapless lawyer had no idea there was a problem with the science and didn’t challenge the crime lab findings.

Rodriguez would spend eighteen years in prison before DNA tests performed on the “consistent” strand of hair led to his exoneration in 2005. Why did it go so wrong? In her new book Cops in Lab Coats: Curbing Wrongful Convictions through Independent Forensic Laboratories, legal scholar Sandra Guerra Thompson examines the case and finds a conflict of interest: this was a police-run crime lab, where analysts were working hand in hand with investigating officers. This is how crime labs operate.

• • •

Sir Arthur Conan Doyle’s Sherlock Holmes fictions drew readers to the scientific investigation of criminality, but the reality of forensic technology and practice caught on much later. The modern crime lab has its roots in J. Edgar Hoover’s FBI, which first set up rudimentary forensic facilities in 1932. The new laboratory was at first housed in a single room. Its employees studied and implemented the latest practices in fingerprint comparison, handwriting analysis, and ballistics and famously yielded results early on in the Charles Lindbergh kidnapping case.

Today there are more than four hundred public crime labs in the United States employing, as of 2009, 13,000 full-time personnel. Labs do everything from drug and DNA testing to fingerprint analysis, fiber comparisons, ballistics, and, increasingly, digital forensics—unlocking and examining smartphones, for example.

Independence can help keep crime labs honest, but honesty does little on behalf of justice if the lab techniques themselves are defective.

Like most aspects of U.S. criminal justice, the forensics system is fragmented into local units. Many labs are small, with fewer than twenty staff members, and some police departments run their own labs. Other labs serve regions and entire states. None are regulated in the way a lab at a hospital or university would be; there is only a loose system of voluntary professional accreditation alongside federal regulations for federally funded DNA labs. Many labs have been the subject of scandals and audits in recent years, as Thompson details in a frightening appendix, listing scores of examples of drug thefts by lab employees, lying about results, planting evidence, concealing errors, and failing to test evidence. As these cases and the unraveling of the Houston lab demonstrate, the American institution of the crime lab may now be more than eighty years old, but it is still very much a work in progress.

Problems at the Houston lab first surfaced in 2002, when journalists began reporting on errors in several criminal cases, including errors contained in DNA tests. Soon the DNA unit was shut down and hundreds of its cases were retested. Then the scientific errors in the Rodriguez case came to light, and DNA tests confirmed his innocence. The analyst who had worked on Rodriguez’s case had gone on to supervise the entire serology and DNA unit at the lab. There were tens of thousands of hair and serology cases to reexamine.

Now the crisis engulfed the entire lab. The city spent millions of dollars auditing it. Michael Bromwich, who had led an inquiry into the FBI’s handling of hair-comparison cases in the 1990s, was selected to lead the process. Bromwich’s team investigated several years of the lab’s work and discovered that analysts were sometimes faking tests—a malfeasance so common that it has a nickname: drylabbing. In some cases, defense attorneys tried to order DNA tests but were falsely told by the lab that there was nothing to test. Scientific fraud was just one symptom of a culture of inadequate supervision, poor quality control, bad procedures, and shoddy work. The auditors found hundreds of cases to reopen.

In the wake of the audit, the lab was closed down and, in 2012, reborn as a new entity independent of the police—the Houston Forensic Science Center. This was a radical idea: a crime lab that answered to scientists rather than cops. The vast majority of crime labs are run by law enforcement and share law enforcement budgets. Most labs will only test evidence if officers request it. But, according to Bromwhich, several years of overhaul in staffing, technology, quality control, and standards have left the Houston lab “a very different place.”

• • •

Independence can help keep crime labs honest, but honesty does little on behalf of justice if the lab techniques themselves are defective. Hence the importance of DNA testing, the new “gold standard,” which has burnished forensic science’s reputation and multiplied lab budgets. But do not blindly trust even DNA, says law professor Erin Murphy. Her book Inside the Cell: The Dark Side of Forensic DNA cautions that, in the era of the cheek swab, the challenges facing crime labs are, if anything, even greater.

Crime labs are now linked together in a national system of DNA databanks run by the FBI, reflecting a remarkable outgrowth of the FBI’s traditional role at the center of American forensic science. But, as Murphy describes, a Hoover-era culture of secrecy still surrounds genetic collection. We do not fully understand the statistics used in DNA database analysis, Murphy says, in part because there is a steadily increasing chance of “adventitious” or coincidental matches as the databases expand. Indeed, the FBI recently admitted basic math errors in the calculation of statistics used in tens of thousands of criminal cases. Although the problem is a national one, Texas is the only state so far conducting a full-scale audit and reopening untold thousands of cases due to that and other recently uncovered errors in DNA interpretation.

Importantly, DNA labs do not have to disclose the error rates in their work, inculcating “the myth that DNA testing is infallible.” History, however, proves otherwise. Murphy describes another Houston lab case that was reversed after outside scrutiny—that of Josiah Sutton, who was exonerated by DNA tests but only after having been convicted based on an erroneous DNA result.

Meanwhile, suspects and defendants lack rights to search DNA databases in order to prove their innocence. After all, law enforcement “owns” the DNA information. Murphy describes the case of Joseph Buffey, a West Virginia man who pleaded guilty to rape and robbery in 2002 but for years was not told that DNA tests performed at the time cleared him. Then, when the results came to light, he was refused access to a DNA database search. Finally prosecutors relented, and he obtained DNA tests matching a different person in the database. The West Virginia Supreme Court then reversed his conviction in an important ruling for the seemingly obvious proposition that people are entitled to be shown scientific evidence of their innocence before they decide whether to plead guilty.

As hard-fought and protracted cases such as Buffey’s suggest, the Constitution offers little protection for criminal suspects and defendants who seek to question or obtain forensic evidence. In many jurisdictions police and prosecutors need not provide defendants and their lawyers detailed documentation of evidence. While the Constitution does obligate police and prosecutors to turn over evidence that might shed light on a defendant’s innocence, the usual rules of criminal procedure seem to vanish into thin air when they collide with the modern crime lab. Lab analysts somehow are not required to turn over their findings for defense scrutiny.

Constitutional rights to privacy have also been shunted aside. Murphy describes how the Supreme Court has endorsed aggressive swabbing of arrestees. In Maryland v. King (2013), the Court ruled that taking DNA from an arrestee is just a method of “booking” to verify the person’s identity and only minimally invades privacy. “License, registration, and cheek swab, please,” says the friendly arresting officer. But we are only beginning to feel the implications of that ruling, Murphy notes. States are responding by expanding DNA collection from arrestees, even for minor crimes. A vast number of Americans are arrested at some point in their lives—nearly half of all men and a nontrivial percentage of the population as a whole. The result is yet another tool of mass surveillance, gathering genetic information that can be archived and searched indefinitely, whether or not an arrestee is ever convicted of a crime. It was left to the late Justice Antonin Scalia to dissent in the King case and deplore the “genetic panopticon” the Court now endorses.

Murphy suggests that as the FBI-run network of forensic databases expands—and these databases increasingly link not just individuals to DNA but also individuals to biometric information—disproportionate policing of minorities will ensure the whole construct is racially disparate in its effect. And more crimes will be solved, but not many more. Solving more crimes requires collecting more samples from actual crime scenes, which means hiring and training police. Our genetic panopticon is also error-prone. While databases rapidly expand, labs try to do more and more with the DNA they obtain. These efforts include risky “low copy” tests of small amounts of material insufficient for ordinary testing, using DNA to locate suspects through their family members, and, still more alarming, testing people’s genes for predispositions toward violence and other behaviors and conditions. There is little regulation standing in the way of such uses of collected DNA. Our genetics are now theirs.

Can the system be improved? In his book Unfair: The New Science of Criminal Injustice, legal scholar Adam Benforado offers a more optimistic vision of the role of science in criminal justice, although he believes today’s evidentiary processes are so unreliable that, in retrospect, they will look little more accurate than witch trials or trial by ordeal. Benforado pins his hopes on the latest psychological research into biases that can affect all of us—including forensic analysts, police, lawyers, and judges. He proposes that not just crime labs but all expert witnesses could be independent and funded by the parties as part of normal court costs. Smartphone apps could guide police to do their work more impartially. More broadly, Benforado suggests shrinking the system itself and focusing more on preventing crime than on blaming and punishing.

DNA labs do not have to disclose the error rates in their work, inculcating ‘the myth that DNA testing is infallible.’

The question remains how much care and objectivity can be encouraged in or expected from people mostly bent on punishing the guilty. When the criminal justice system has had money to spend on technology, the powers that be have not spent it on methods to improve impartiality and fairness but rather on vast databanks designed to sweep up genetic evidence from millions of convicts and arrestees and to ramp up capability to identify culprits en masse. If lab errors result in thousands of convicted innocents along the way, then that is just the price we pay. Perhaps a radically new model is needed, along the lines of Houston’s independent lab. But even that reform, however important, does not change the fact that the chief consumers of forensic evidence, police and prosecutors, are not independent: they are bent on getting convictions.

Lawyers and judges may not be best positioned to identify, much less fix, defects in scientific methods. After all, plenty went to law school because they did not have much interest in math or science. But the modern crime lab has remarkably avoided both the standards of the law—including criminal-procedure rights—and the standards of science, such as quality assurance and basic research concerning reliability. No wonder lab-wide scandals brew undetected. The FBI lab has itself been embroiled in one after another. There was the botched fingerprint match that falsely implicated an Oregon lawyer in the 2004 Madrid train bombing. The Bureau has also abandoned the entire discipline of bullet-lead comparison, used in testimony in hundreds of cases, as scientifically unfounded. And there is the ongoing audit of thousands of hair-comparison cases dating back decades, in which FBI analysts overstated evidence and helped to send innocent people to prison. Several state labs are also auditing hair-comparison cases. The Texas Forensic Science Commission recently stated that bite-mark comparison evidence, too, should no longer be used in criminal cases and is auditing those cases. Enacting the kind of prophylactic measures that might enable labs to avoid errors would require dramatic change.

The exciting news is that, to some extent, this is now happening, and not just in Houston. Many observers, including the National Academy of Sciences in a prominent 2009 report, have called for a “research culture” in forensics. Some labs listened and are implementing new quality controls. More funding is being directed to basic research, and new scientific commissions—including the Department of Justice’s National Commission on Forensic Science, established in 2013, and a large group of committees convened by the National Institute of Standards and Technology—are developing standards. These groups have brought together leading scientists, lawyers, judges, and forensic analysts to consider improved scientific standards, new research, and new procedures. They have already produced important guidance on topics such as accreditation of laboratories; pre-trial discovery—that is, defense access to evidentiary information; and standards to avoid vague language in forensic conclusions. The Department of Justice has just announced that it will conduct “stress tests” on a range of forensic disciplines.

While these developments are important, it remains to be seen whether a new culture will take root throughout the system. Consider that the vast majority of cases are settled with plea deals, public defenders lack access to their own forensic experts, and crime labs are overwhelmed with drugs and cheek swabs for processing. Given these structural obstacles, how much improvement can better forensic standards really yield?

Still, if there were ever a time to reform criminal forensics, now would be it. The punitive aspects of American criminal justice are undergoing broad reconsideration as more and more citizens find themselves hard-pressed to justify mass incarceration and the war on drugs. If we dial down incarceration, we should also focus on the quality and not just the quantity of forensics. That is, we should not demand that crime labs mass-process drug and forensic tests, without adequate quality controls, just to keep up with an overcharged arrest-and-conviction machine.

Improving the quality of forensics will require a cultural shift. In surveys, non-lawyers show extremely strong faith in the accuracy and uniqueness of fingerprint, DNA, and other forensic evidence. (Lawyers are more skeptical.) People believe that forensics can unfailingly identify guilty criminals. But while forensic analysts used to tell jurors that they were error-proof, such infallibility is simply impossible. Human error is inevitable, though it can be minimized.

Reform might at last vindicate Americans’ belief in the ability of science and technology to serve justice. For the time being, though, the crime lab, the very part of the criminal system that is supposed to do science, far too often cannot get it right.