Inside Story

Genetic injustices

DNA evidence has exonerated nearly 300 prisoners in the United States, but an Australian case highlights its potential to mislead

Jeremy Gans 7 June 2012 4875 words

Sudanese refugee Farah Jama (left) with his lawyer Kimani Adil Boden after his release from jail in Melbourne, where he had been imprisoned for sixteen months after being wrongly convicted on the basis of DNA evidence. Aaran Francis/Newspix

Convicting the Innocent: Where Criminal Prosecutions Go Wrong
By Brandon L. Garrett
Harvard University Press | $59.95

Genetic Justice: DNA Data Banks, Criminal Investigations, and Civil Liberties
By Sheldon Krimsky and Tania Simoncelli
Columbia University Press | $35.95


Why would a law professor write a book about several hundred horrific murders and rapes? It’s not because Brandon Garrett is trying to cash in on the popularity and guilty pleasures of “true crime.” Rather, he wants to use these stories to reveal another sort of truth.

A centrepiece of the book’s earliest case is a presumed confession to a horrific murder in Garrett’s home state of Virginia. Asked by two detectives what he used to tie up his victim, David Vasquez first suggested rope, then a belt, then a clothes line — answers his supposed interrogators rejected in turn. Finally, they simply told him he used a cord from some Venetian blinds:

Vasquez: Ah, it’s the same as rope?

Detective 2: Yeah.

Detective 1: Okay, now tell us how it went, David – tell us how you did it.

Vasquez: She told me to grab the knife and, and, stab her, that’s all.

Detective 2: (voice raised) David, no, David.

Vasquez: If it did happen, and I did it, and my fingerprints were on it…

Detective 2: (slamming his hand on the table and yelling) You hung her!

Vasquez: What?

Detective 2: You hung her!

Vasquez: Okay, so I hung her.

The cops must have thought Vasquez was toying with them, or stupid, or simply crazy. So must the jury, which convicted him of murder. But, like all 250 people found guilty of horrible crimes in Garrett’s book, the United States criminal justice system now accepts a quite different story: that Vasquez was wholly innocent.

Garrett’s genre of “false crimes” has a worthy lineage. Convicting the Innocent shares its title with a book written by another American law professor, Edwin Borchard, eighty years ago. That was an era when justice officials proudly declared that wrongful convictions in criminal courts were “a physical impossibility.” Borchard’s response was to detail sixty-five tales of innocent people found guilty in the preceding decades in American and (in a few cases) English courts.

Nowadays, the task of convincing Americans that their courts can make the worst of mistakes requires just one, repeated, tale: exoneration by DNA evidence, the common thread running through the 250 cases examined by Garrett. Vasquez’s exoneration came four years after his trial, when pubic hairs from his file were genetically linked to Timothy Spencer, the “Southside Strangler,” who had been convicted of three similar strangulation murders in Virginia in 1987. Vasquez was the first of nearly 290 prisoners (including seventeen facing the death penalty) exonerated so far by DNA, many decades after their alleged crimes. Spencer was the first person executed on the basis of DNA evidence.

Because no thinking American now needs to be convinced that their criminal justice system can and does make terrible mistakes, Garrett sets himself the different task of finding common threads in this extraordinary set of cases, examining available trial transcripts, news reports and court judgments to uncover “systemic” causes of error.

The first half of Convicting the Innocent is devoted to the flawed evidence America’s criminal justice system used to wrongly convict those 250 people. Three-quarters of the cases, including Vasquez’s, involved eyewitness identifications and forensic evidence. Garrett’s study of available transcripts suggested that almost 90 per cent of the eyewitness identifications were transparently marred by procedures that “suggested” identifications to witnesses and by discrepancies between the witness’s initial description and testimony in court. More than 60 per cent of the forensic evidence, meanwhile, involved claims that exceeded or contradicted even the shonky standards of dubious techniques such as hair comparison. Rounding out the categories of flawed evidence are the suspects’ alleged confessions to police and jailhouse informants; almost all were bolstered by claims, evidently false, that they must be accurate because they included details (like the type of cords used to tie a victim’s hands) that only the real culprit could have known.

Flawed evidence alone isn’t enough to explain these wrongful convictions. After all, the criminal justice system is meant to separate the wheat from the chaff. So, in the second half of his book, Garrett turns his attention to flawed court processes, including superficial trials during which the case for innocence was barely mentioned, appeal courts that turned a blind eye to the dangers, and authorities who actively resisted efforts to uncover or act on exculpatory DNA evidence.

Because DNA exonerations tend to occur in particular sorts of cases, Garrett is quick to caution against generalising from the cases he examined. The 250 exonerations were “limited to a small set of mostly rape convictions, mostly from the 1980s, in which the evidence happened to be preserved.” According to the compilers of a new US National Registry of Exonerations, the DNA-based cases make up just one quarter of the miscarriages of justice uncovered there since 1989. The remaining three-quarters include people cleared after key witnesses changed their stories, the real culprit confessed or a court or the executive belatedly acknowledged that the state’s evidence had never been adequate. One possible contender for inclusion is Carlos DeLuna, who, according to yet another law professor’s recently published work, was executed for a murder that was probably committed by a different Carlos (who looked like DeLuna, had been fingered by him as the true culprit from the outset, had committed a string of similar offences and admitted to committing the murder).

Australia’s most famous modern miscarriages of justice, including the cases of Lindy Chamberlain, Andrew Mallard and Farah Jama, would fall outside a similar Australian study to Garrett’s because none were uncovered by DNA evidence. Indeed, Australia has nothing like America’s string of belated DNA-based exonerations. Our best known DNA-revealed error – corrected after post-trial tests of the linen of the accused rapist Frank Button – was labelled a “black day for justice” by the Queensland Court of Appeal, even though (unlike every case in Garrett’s book) it was uncovered before the defendant’s first appeal.

But it is surely noteworthy that contemporary Australian courts still use all four categories of flawed evidence that Garrett found dominated the trials that were overturned in the United States. An example in the first category occurred when Queensland police breached their own procedures by asking three bank robbery witnesses to go to a courthouse and keep an eye out for the robber; all three picked Brunetta Festa, who was being committed for the robbery that day and was the only woman under forty present. The trial judge’s admission of this patently dangerous identification evidence was upheld by the High Court in 2001. The evidence used in 2004 to convict Bradley Murdoch of the notorious murder of Peter Falconio had two kinds of flaw. Not only did it include Joanne Lees’s eyewitness identification (made after she saw his photo in a British paper with a caption revealing that his DNA was linked to the case) but it also featured a Sydney academic’s “body mapping” technique (never explained in court) that “matched” Murdoch to a grainy truck stop video (and has since been rejected by NSW courts as inadmissible for that purpose.)

Three years later, at Glenn McNeill’s 2007 trial for the murder of Janelle Patton on Norfolk Island, the prosecution told the jury that McNeill’s confession to the Federal Police included “information that was not in the public domain.” Not only is such a claim ludicrous in a close-knit island community, but it’s also the case that all the “information” was detailed in a book on the case by ABC journalist Tim Latham published five months before the police first spoke to McNeill. And Australian prosecutors continue to rely on prison informants – the fourth of Garrett’s flaws – in difficult cases, including last year’s trial of Peter Dupas for the murder of Mersina Halvagis, which was partly based on Dupas’s alleged confession to disgraced Melbourne lawyer Andrew Fraser while the pair were gardening at Port Phillip Prison.

Some of the systemic problems identified by Garrett in the United States – decentralised and nakedly politicised courts and prosecutors, hopelessly inexperienced attorneys assigned to defend capital murder cases, and the nation’s tortured and cursory appeals process – do not resonate so strongly in Australia. And some of the reforms he advocates, such as the recording of police interrogations, have been legal requirements throughout Australia for decades.

But the contrast with America isn’t always flattering. The US system for post-conviction reviews by courts, which Garrett excoriates for its slow pace and procedural tangles, is still superior to the systems operating in much of Australia, where defendants who have exhausted their appeals must plea for mercy to politicians if they want to obtain DNA evidence for fresh testing or rely on the results to overturn their convictions.

Nor have either of Australia’s two most populous states followed Britain and some American states in creating independent commissions to comprehensively investigate claims of innocence. The failure of political will was most apparent in 2003, when the Carr government hurriedly suspended a modest “Innocence Panel” it established a year earlier simply because one of the people convicted of murdering Janine Balding asked the panel to arrange DNA testing of some of the evidence against him. It seems that our new foreign minister, like many a hick American prosecutor, thought that the interests of victims of notorious crimes necessitated non-negotiable “closure.” Garrett would no doubt direct Carr’s attention to the Florida prosecutors who spent a decade opposing Frank Lee Smith’s requests for DNA testing in a child murder case. While the prosecutors presumably thought they were protecting Smith’s victims, the person who they were really protecting was the real killer, Eddie Lee Mosley, who committed sixty rapes and seventeen murders, including some during Smith’s lost decade in prison.


THE several hundred Americans who now owe it their freedom are only a tiny part of the story of forensic DNA. The very things that allowed DNA evidence to convince a sceptical American public of the innocence of a fraction of their prisoners – its foundation in mundane lab work and peer-reviewed science, its independence of the biases and memories of eyewitnesses, its longevity and robustness – have also driven its dramatic take-up by police worldwide as a means of convincing courts of the guilt of countless accused. This part of the DNA tale has been told in the many “true crime” books that show how dogged investigators have used emerging technology to zero in on criminals who would otherwise have gone to their graves with their villainy undiscovered.

In Genetic Justice Sheldon Krimsky and Tania Simoncelli offer a different take. Their foreword commences with the story of Lily Haskell, who was arrested at a demonstration against the Iraq war in 2009 and whose DNA profile is now one of more than eight million permanently stored by US law enforcers. This is despite the fact that she has never been charged with, let alone convicted of, any crime, much less a horrible one. The book reviews developments in the United States and (more briefly) five other countries, including Australia, to argue that the once rare use of DNA profiling to catch (and sometimes exonerate) violent criminals “has given way to a massive and ever-expanding system of collecting and permanently retaining DNA for ongoing investigation and use.”

The easy part of Krimsky and Simoncelli’s case is their argument that supposedly strict government regulation of the investigative use of DNA is anything but. Take Australia, where a Model Forensic Procedures Bill was developed at the turn of the century, creating an elaborate regime of orders for “intimate” and “non-intimate” forensic sampling and a complex DNA profile database whose use was regulated by a table of “matching” rules. The Bill was so packed with poorly drafted protocols creating sham protections for defendants (such as a rule requiring everyone who has their hair plucked to be handed one of their own hairs for independent testing) that no Australian jurisdiction adopted it in its entirety. In any case, the sampling regime proved to be irrelevant in practice because nearly all DNA was obtained through people “consensually” swabbing their own mouths; and the matching table has been gradually amended so that virtually all matches are now allowed.

Although the move to allow uncharged people’s DNA profiles to be held in the database is relatively recent in the United States, such holdings have been permitted in Australia from the outset, with any samples taken to test a person’s link to one crime automatically placed on the full database for matching against all unsolved crimes. In England, a significant fraction of the DNA database (the world’s first) is now made up of non-offenders who were only briefly of interest to investigators, such as suspects whose DNA samples were taken to investigate (and often dispel) suspicion and people who volunteered their samples as part of mass DNA screenings aimed at excluding entire sub-populations from a police inquiry.

In fact, few nations prevent the police from asking for DNA samples from anyone they wish or from gathering samples (from cigarettes, cans of Coke and the like) abandoned by people who refuse a request or where the police have insufficient grounds to obtain an order. But perhaps you think the police haven’t got your DNA profile because they never ordered or asked you to provide one (and you don’t smoke or drink Coke)? Think again. Anyone who has been at a crime scene (including many victims of crime) may have their DNA profile unknowingly stored on the database, ready for matching against all unsolved crimes. And, for every profile they’ve obtained, by hook or by (convicted or suspected or even exonerated) crook, the police can also explore the link between any unsolved crime and everyone who genetically “resembles” that profile (that is, all of the close blood relatives of the profile’s “donor”).

Having pricked the balloon of rhetoric about close regulation of genetic forensics, the two American Civil Liberties Union associates (one a research fellow, the other the organisation’s Science Adviser) embark on a far tougher sell: that this mother of all genetic sweeps is a bad thing. Not surprisingly, their first argument concerns forensic DNA’s impact on our privacy – not just of our individual DNA but also our bodies, our health, our relationships and our movements. Krimsky and Simoncelli contrast the high levels of protection afforded to genetic privacy in medical and research contexts with the “different playing field” that applies in forensics. They bemoan the focus by the police and the courts on the relatively trivial bodily intrusion of DNA sampling rather than on the informational value of DNA, and predict that public anxiety about genetic privacy will eventually force a reckoning. But that argument ignores the quite different contrast that can readily be drawn between the informational intrusion of DNA profiling (even in its most extreme forms) and the everyday indignities imposed on criminal suspects by police questioning, or body searches, or entry into homes, or tapping of phones, or searching of government records, which are part and parcel of modern investigations.

The manhunt for the killer of backpacker Peter Falconio is a good example. The Northern Territory police relied on tip-offs, a blurry service station photo and Joanne Lees’s even blurrier memories to generate a shortlist of thousands of “persons of interest.” Because of the legal and political complexities of cross-border investigations, few were forced to provide DNA samples and only a few others were asked politely, despite the police obtaining a DNA profile from a spot of blood on Lees’s shirt. Instead, the investigators worked through every man on the list, questioning them personally, investigating their alibis and searching phone and banking records to determine their movements. It was only information from a fellow drug dealer that led the police to refocus on Bradley Murdoch, who’d politely refused to provide his DNA early in the investigation. Scared of spooking him, they sought DNA from his brother and later used a sample repurposed from an earlier investigation into rape charges that Murdoch claimed were trumped up and that a jury later dismissed. Those who argue against contemporary DNA profiling face the twin burdens of advocating a return to old-school policing methods such as these and barring the police from some of the unorthodox methods that eventually netted a killer.

Ultimately, Krimsky and Simoncelli don’t advocate a retreat from genetic forensics. Instead, they ask whether more of a good thing is necessarily better. They question the extent of the benefits of DNA profiling, pointing out that the evidence offered by administrators about the utility of their database – the proud recitals of “hits” – tells little about outcomes the public actually cares about, such as the number of matches of interest (as opposed to belated confirmations of already known links), the number of convictions that follow (as opposed to charges dropped because the match proves to be uninteresting) and the number of crimes prevented (a matter about which much is claimed but little can be proved).

Krimsky and Simoncelli also question whether the well-publicised successes in forensic DNA’s heartland – the matching of convicted serious criminals to unsolved sexual or violent crimes – will continue as the emphasis moves to arrestees, minor criminals, and property or drug offences. It is one thing to apply cutting-edge science to Joanne Lees’s t-shirt, but another altogether to gather, analyse and identify all of the samples found in every stolen car or drug lab, each yielding dozens of (typically spurious) investigative leads and generating lengthy backlogs at overburdened, demoralised labs. Rather than revolutionising criminal investigation, the expansion of DNA databasing may simply replicate the failings of regular criminal justice, including the social and racial disparities in its application. In short, a lot of contemporary and future DNA databasing may simply be a very expensive way of rounding up the usual suspects.

Alas, this clever argument will find no purchase among voters and politicians who are quite comfortable with the unprincipled face of criminal justice. Many of the reform “principles” set out in the final chapter of Genetic Justice (mostly requiring court involvement in any envelope-pushing steps, such as mass screenings and familial searches) were recommended in 2003 by Australia’s Law Reform Commission but remain unimplemented. Indeed, the trend here is to oust courts from the system, with parliaments in South Australia and Victoria repealing various court-enforced rules after judges had the temerity to apply them against the police. In the United States, broad DNA databasing laws have been approved by popular referendums and the Ninth Circuit Court of Appeals recently rejected a constitutional challenge by Lily Haskell, the briefly arrested protester.

Indeed, Krimsky and Simoncelli’s complaints about the inequities and inefficiencies of contemporary DNA databasing suggest a quite different way forward: putting absolutely everyone’s DNA profile on an investigative database. Their chapter titled “The Illusory Appeal of a Universal DNA Data Bank” actually sets out a compelling case for such a step, which would remove racial and social disparities and considerably reduce the problem of unidentified profiles and spurious matches. The authors’ main counter-arguments – the logistics of establishing and securing such a depository and the failure to tackle the inequities in the rest of the criminal justice system – are far from overwhelming.

“How many offenders might we tolerate escaping in order to avoid an innocent person being wrongfully condemned?” This familiar question is at the heart of Convicting the Innocent, yet it is also the closing plea of Genetic Justice, whose authors invite us to factor in the plight of innocents like Haskell who have been swept up in the worldwide DNA dragnet. But a case that features in both books suggests a different answer to the one they favour.

In Garrett’s book, Darryl Hunt is just one of hundreds of men convicted of a horrific murder on the basis of flawed identification procedures and supposed confessions. Even after DNA profiling proved that the semen on the victim wasn’t his, US courts refused to clear him until the true killer, Willard Brown, confessed to acting alone. The details of Brown’s capture are recounted in Krimsky and Simoncelli’s chapter on familial screening, in which they reveal that police had run DNA from the semen through the North Carolinian database and had obtained a partial match with Anthony, Willard’s brother. Willard’s confession came after the police gathered DNA from a cigarette he discarded and matched it to the semen on Hunt’s supposed victim.

There’s no doubt that the world’s Haskells vastly outnumber its Hunts, but who would baulk at “condemning” any number of innocent people to having their DNA profiles placed on a database if that allowed just one more innocent man to “escape” the true hell of wrongful imprisonment?


THERE’S another story about innocence and DNA that neither book mentions because it’s too recent and too distant from America. It’s the story of a Melbourne teenager, Farah Jama, who recently spent a year in jail for a rape that he didn’t commit (and, indeed, that never happened at all). At Jama’s trial in 2008, there was no witness identification, no junk science and no confessions to the police or anyone else. Nor was his wrongful conviction due to lying complainants, dodgy lawyers, sleepy judges or bad juries. Instead, there was just one unsanitised medical trolley at a rape crisis centre that allowed a single intact spermatozoon (and some fragments) to move from one woman’s hair to a swab taken from a second woman who had passed out at a nightclub. The presence of a male DNA profile on a vaginal swab was enough to wrongly convince everyone (the police, the prosecution, the defence, Jama and, of course, the woman who’d passed out) that there’d been a horrific rape. And a match to Jama’s DNA profile (who, like Haskell, was in the database despite being cleared of the crime that led to his sample being there) was enough to convince a judge and jury that an innocent teenager was a monster with an alibi from a family of perjurers.

Jama’s case highlights two huge problems with DNA profiling that can lead to disasters. The first is that DNA is messy, fiddly stuff. The cells that contain our DNA move from our bodies to other people’s bodies, to surfaces or just through the air. As a result, crime scenes (and victims’ bodies) are typically dotted with various people’s DNA, all mushed together. Labs must do their best to isolate a specific profile through educated guesswork and processes of elimination, subjective choices that complicate the estimates of how rare that profile is in the community. Because DNA, like any complex molecule, is fragile, the profiles in a sample may be only partial ones, albeit still mixed up with others. And because DNA can’t be seen and must instead be “amplified” by the same mechanism that builds life itself, there’s no way to tell whether the bit of DNA that matches a person on a database was in the sample to start with, or whether it just floated there or was introduced at one of the many choke points in the forensic system where evidence from different sources brushes together – places such as a busy crime scene, an evidence bag, a police station, a rape crisis centre or (most disturbingly) a DNA lab.

Messiness is nothing new to crime investigation. What makes it disastrous in the case of forensic DNA relates to its second problem: that most people who use DNA know next to nothing about it. In the courtrooms that are its ultimate consumer, DNA testimony is universally recognised as a punishingly dull affair. Inevitably, everyone snoozes while a lab technician with no background in talking to lay people drones on about alleles and electrophoresis in front of a series of PowerPoint slides. In response, defence counsel resort to scattergun questions that worked once-upon-a-trial or sometimes wheel in experts of their own to go head-to-head in the boredom stakes with warnings about “allelic ladders” and “peak height ratio protocols.”

What about the lab technicians themselves, you might ask. Well, these are people who do hundreds of analyses a week following set procedures, are employed by the police, have learnt to fear defence lawyer tricks and, as a result, often seem to have tunnel vision. When a nervous police officer in the Jama case asked the lab about the possibility of contamination, a technician duly made a perfunctory check of a couple of lab records and gave the all-clear without so much as considering the possibility of a contamination before the swab reached the lab.

And the jury? Actually, Jama’s jury was terrific. They asked the question no one else did: why had the teenager’s DNA been on the database in the first place? The true answer – because Jama had been briefly accused of rape by the woman who attended the rape crisis centre with Jama’s semen in her hair the evening before the woman who passed out in the nightclub – would have cracked the case. But Australia’s rules of evidence prevent jurors from learning about (apparently) unrelated allegations, so the judge hurriedly told them that their question was irrelevant. It was a serendipity – the chance coincidence that the same doctor treated both women at the rape crisis centre – that led to the belated recognition that Jama was innocent.


GARRETT, Krimsky and Simoncelli are all alive to the potential failings of forensic DNA. Three of the 250 wrongful convictions described in Convicting the Innocent were partly due to sloppy DNA analysis revealed by more accurate or advanced DNA profiling, and a chapter of Genetic Justice is devoted to the many ways that DNA analysis can go awry. But both books underplay the implications of these flaws. Krimsky and Simoncelli argue that the science’s fallibility must be factored in when considering “whether justice is advanced by DNA,” but ignore it when considering the potential for genetic privacy to be invaded by individuals or governments. More dubiously, they claim that the risks of error only affect “inclusions” (matches, typically probative of guilt) and not “exclusions” (non-matches, typically probative of innocence), even though contamination can obviously affect either. Indeed, they had already noted the ease with which an offender could leave a trail of artificial DNA (or just some chump’s cigarette butts) at a crime scene as a genetic alibi.

By contrast, Garrett properly concedes that “it is possible that an erroneous DNA test could lead to the exoneration of a guilty person.” But he nevertheless insists that the close scrutiny given to the 250 cases in his study allows everyone to be “quite confident that these convicts are actually innocent.” He, of all people, should know better than to make such claims. For instance, neither Willard Brown’s DNA on a victim’s body nor his confession to acting alone conclusively prove that Darryl Hunt had no hand in her killing.

Neither book’s final chapter, setting out the authors’ recipes for a safer, fairer criminal justice system, contains any recommendations aimed at avoiding convictions that are founded on DNA errors. Justice Frank Vincent, who inquired into Farah Jama’s wrongful conviction and recommended a series of changed evidence-gathering processes, saw it as obvious that no one should ever be convicted on the basis of DNA evidence alone, not least because of the inevitable, albeit often small possibility, that someone else might match any DNA profile.

Vincent’s proposition was tested soon after when Ben Forbes appealed to Australia’s High Court against his conviction for a rape near a Canberra bike path. The prosecution case against Forbes rested solely on a match between his DNA profile and mixed samples from the victim’s bra and jeans, with experts testifying that the chance of a random person having that profile was less than one in a million. While that figure clearly allowed some possibility of other matches in the region, the High Court, a booster of forensic evidence from way back, dismissed Forbes’s appeal on the ground that the prosecution had agreed to keep a more damning (but wholly untested) DNA statistic from Forbes’s jury in return for their experts being allowed to (meaninglessly) label their own evidence against Forbes “extremely strong.” It’s the kind of ruling that makes you wonder whether Justice sometimes peeks from under her blindfold. The seven High Court judges would not have needed to look far. Forbes’s prior conviction for sexual misconduct on a Canberra pathway, inadmissible at his trial, was fully detailed in the court file.

And that’s always the seduction. Yes, we know the risks and we believe in justice, but sometimes we just know the truth too. We know Ben Forbes is a rapist. We know Frank Button and Farah Jama aren’t. We know who Janine Balding’s killers are. And Peter Falconio’s and Janelle Patton’s and Mersina Halvagis’s too. And precisely because we know all that, we also know that there will be more books called Convicting the Innocent written in the decades to come. •