Tainted evidence: science in the dock

Comment from Australia

Television teaches us forensic science is always right but increasingly courts are begging to differ, writes Stuart Washington.

When a sledgehammer-wielding robber broke through the front doors of the Willoughby Hotel early one January morning in 2008, his balaclava-clad image was captured by a security camera.

Fewer than 30 minutes later the same robber, still bearing a sledgehammer and still wearing a balaclava, was photographed stealing just over $13,500 at the P.J. Gallagher's Irish Pub in Drummoyne.

Twenty-three months later, a Tempe man, Raymond George Morgan, was found guilty in the District Court of two counts of robbing more than $58,500 in the heists and sent to jail.

The security cameras helped convict Morgan of the crimes when an expert witness linked him to the scene of the crimes through a science used in court known as "body mapping".

Earlier this month Morgan won a retrial because the expert's evidence was rejected by the Court of Criminal Appeal. The court found the body mapping was "simplistic" and did not use one measurement.

In the appeal court's view, the "science" of body mapping used in the case of Morgan was wearing no clothes.

So much for all those toned and svelte CSI forensic investigators running around on television each week.

Morgan's robbery convictions have been overturned at a time when increasing concerns are raised about "junk science" appearing before the courts: unreliable expert evidence leading to contaminated criminal trials.

At its worst, such evidence could result in innocent people being sent to jail. At the very least, such evidence could tilt trials unfairly.

Disturbing examples of the fallibility of expert evidence have played out in two recent Supreme Court appeals involving Gordon Wood, found guilty of murdering his girlfriend Caroline Byrne, and Jeffrey Gilham, found guilty of murdering his parents.

Judgments are reserved in both cases, but the court has made no secret it has dim views on the credibility of expert evidence in both cases

An academic in Canada, Emma Cunliffe, has also raised serious questions about the robustness of expert evidence used to convict the Novocastrian Kathleen Folbigg of killing her four children.

The state of forensic science appears to be so parlous that further examples seem likely.

Gary Edmond, a University of NSW law academic who has researched expert evidence over the past four years, wrote in a paper in the Australian Journal of Forensic Sciences earlier this year: "A good deal of the opinion evidence produced by forensic science and medicine appears to be unreliable or of unknown reliability, and obtained in conditions that make few, if any, sustained attempts to minimise known risks and dangers."

In 2009, the peak body of US scientists, the National Academy of Sciences, released a report that found only DNA testing among the broad range of forensic sciences was sufficiently grounded in science to regularly and accurately identify a suspect. (And even then serious concerns remain about DNA evidence, including the ability to wrongly convict people on tainted or rigged samples).

The flip side is that almost every other field of forensic science - the science used by expert witnesses in court to establish guilt or innocence - does not have the same level of scientific rigour.

The pall cast by the report hangs over areas that have developed a TV-inspired mythos of infallibility: blood spatter examinations, bullet rifling and even fingerprinting.

"What we see on TV is a super science," says Richard Kemp, a University of NSW psychologist who is studying the way juries react to expert evidence. "And that is so at odds with the reality of what is occurring. You want scientists doing science, which on the whole is really dull. It doesn't make for good TV."

The arguments of Edmond and Kemp are familiar to those at the highest levels of the judiciary, with the chief judge at common law in the NSW Supreme Court, Justice Peter McClellan, talking about a future need for structural reforms to help courts deal with the way they assess expert evidence.

"I have written over many years about the issue of the capacity of courts under the conventional structure to decide between the views of scientists where there are issues of controversy," Justice McClellan says.

He says that as the complexity of science grows "courts will have to be alert to ensure that the science which is admitted in the courtroom is good science if not the best science available".

Kemp gave evidence in the case about Morgan, raising questions about the science that helped secure the conviction.

Evidence against Morgan was based on body mapping in which distorted and poor quality images on security cameras from both hotels were compared to images of Morgan's body, head and face shape.

The comparison, made by a University of Adelaide biological anthropologist and anatomist, Maciej Henneberg, was able to find "a high level of anatomical similarity".

Henneberg's comparisons were able to find similarities including a broad "orthognathic" nose, an elongated "dolichocephalic" head when seen from above and a "straight" posture.

The comparison was made remarkable by the fact the robber was dressed head-to-toe in black, including a close-fitting balaclava and a hoodie pulled over his face.

The appeal court sided with defence experts including Kemp about the deficiencies of Henneberg's comparisons and excluded his evidence.

The court found Henneberg's reliability was affected by his failed attempt to compare images with Pauline Hanson after nude photographs purporting to be of Hanson were published in 2009. Henneberg had told the media he was "99.2 per cent sure" the photographs were of Hanson. The photographs were later found to be not of Hanson.

The court also agreed Henneberg's methodology was "simplistic" and found it was concerning there was no research into the validity, reliability and error rate of body mapping.

Edmond is calling for such problems to be addressed by the introduction of a specific reliability test for expert evidence to be heard by the courts. At present there are lower conditions placed on expert evidence under the Evidence Act: it must involve specialised knowledge based on his or her training; and the opinion expressed must be wholly or substantially based on that knowledge.

Kemp expresses some incredulity that reliability is not specifically demanded by law. There are explicit requirements for reliability of evidence in the US, Canada and, shortly, Britain. "When I speak to fellow scientists, they can't believe that such a statute doesn't exist," he says.

The confronting message being pushed by Edmond is being heard within some of the most senior institutions in forensic science in Australia, with some reservations.

James Robertson, who joined the University of Canberra as forensics professor after a 21-year career with the Australian Federal Police, joined Edmond to organise a conference that aired perceived problems about expert evidence earlier this month.

But Robertson, and several other senior forensic experts spoken to by the Herald, denies there is a crisis of confidence in the expert evidence appearing before courts. Robertson says Australia, through accreditation systems and other safeguards, leads the world in terms of a consistent approach to the way forensic experts conduct their work.

Morgan's court appeal victory was bittersweet. He made no contest to two other charges of possessing almost $30,000 in proceeds of crime and remains in jail on those charges. And two bundles of money found in his possession were neatly bundled in the way the Willoughby Hotel bundled its notes.

On technical grounds about directions given by the trial judge, Morgan successfully appealed against a conviction of receiving an Audi stolen from Mosman on the night of the robberies and used as the getaway car.

But the court of appeal referred to evidence that Morgan had possession of the car keys and lock-picking implements. In total there was enough evidence without body mapping that would allow a jury to convict on the robbery charges, meaning he faces a retrial instead of an acquittal.

Edmond points out the justice system is built on a foundation that it is better for 10 guilty to go free than allow one innocent person to be jailed. And he warns junk science may erode that foundation. "We have a system that requires people to have a fair trial," Edmond says.

Original report here




(And don't forget your ration of Wicked Thoughts for today. Now hosted on Wordpress. If you cannot access it, go to the MIRROR SITE, where posts appear as well as on the primary site. I have reposted the archives (past posts) for Wicked Thoughts HERE or HERE or here