Chat with Jack, live from 11.30am. Leave your comment below for Jack to respond to.

Prosecutors and detectives complain that modern jurors are often highly unimpressed by professional briefs if they do not contain forensic evidence as a routine feature. Why was not the burglary crime scene checked with fluorescent light? What about trace particles from the defendant’s shoe? Did anyone check the last movements of the deceased by interrogating mobile phone records, closed circuit cameras on public streets?

Scientific evidence has the cachet of seeming to be factual, objective – incapable of lying. By contrast, even citizens with a genuine respect for police accept that some detectives may once have ‘‘verballed’’ accused criminals by pretending that they have confessed, planted evidence, and colluded to provide corroboration. They are more comfortable with ‘‘real’’ evidence.

That scientific or ‘‘forensic’’ detectives often appear, at least on TV, to be detached from ordinary detectives or policemen adds to the air of impartiality, detachment and care.

Everyone understands, of course, that evidence can be misinterpreted. Often, indeed, the very plot of a CSI episode is about honest misinterpretation – at least until the diligent investigator recognises another possibility that more accurately fits with the facts.

In the annals, real and fictional, of science applied to detection, there is no reputation standing higher than that of the United States Federal Bureau of Investigation, or FBI. Even when, sometimes, the reputation of the FBI suffered because of J. Edgar Hoover and his obsessions, the bureau was known for its development of scientific data bases, expertise and professionalism in the analysis of evidence – of fingerprint, blood, materials, ballistics, explosives, and thousands of specialisations in typewriter faces, paper, tyres, tools, paint, metallurgy and, later, electron microscopy, nanotechnology, biometric identification systems, and the use of computers to access zillions of files. In many fields of forensic science, most of the knowledge – even of bizarre matters such as the rate of decay of human bodies in different conditions, rates of growth of aerobic and anaerobic bugs, patterns of exit wounds from bullets – was developed by the FBI. And taught by its scientists to detectives and forensic scientists all about the world.

It was also a world of a good deal of shared expertise, so that discoveries made, say, by French, or Italian or British forensic scientists were shared, discussed and criticised in an open scientific environment. Even with Australians, who regularly go to worship at Quantico.

Thanks to The Washington Post, that reputation is taking a pummelling this week, with a belated and unwilling FBI disclosure of a 1997 inquest into the methods of the centre from 1986. The Inspector-General of the US Department of Justice began its nine-year investigation after a whistleblower complained that sloppy work by examiners produced unreliable evidence in criminal trials. While the report was in some respects critical of the whistleblower, it found that many of his complaints about unprofessional and bad science by the FBI were justified.

Typically, the examiners were not so much fabricating results as making far more of them than they could. Examiners often jumped to conclusions, operated well out of their fields of expertise, and gave evidence specifically tailored to lead towards the (generally incriminatory) conclusion being sought. They were very definitely biased towards prosecution cases, and tended to ignore material that cast doubt on evidence that was often indicative as best. There was a propensity to make speculative jumps unjustified by either the science or the findings, producing definite results that were, simply, unsound and unjustified.

Records were often dodgy, done in arrears, and tended to elide any contradictory or ambiguous results.

Here’s a sample finding, about FBI evidence in each of the World Trade Centre bombings and, the later Oklamhoma City bombing.

‘‘We concluded ... that [expert witness David] Williams based some of his conclusions not on a valid scientific analysis but on speculation from the evidence associated with the defendants. Just as ... in the World Trade Centre case, [Williams] offered an opinion about the velocity of detonation (VOD) of the main charge that was unjustified.

‘‘His statement about the VOD of an ammonium nitrate fuel oil (ANFO) explosive – the explosive allegedly used – was incomplete. His categorical identification of the main charge as ANFO was inappropriate based on the scientific evidence available to him. Here Williams did not draw a valid scientific conclusion but ... speculated from the fact that one of the defendants bought ANFO components.

‘‘His estimate of the weight of the main charge was too specific, and again was based in part on the improper, non-scientific ground of what a defendant had allegedly purchased. In other respects as well, his work was flawed and lacked a scientific foundation. The errors he made were all tilted in such a way as to incriminate the defendants.

‘‘We concluded that Williams failed to present an objective, unbiased and competent report.’’

The Inspector General recommended major changes, some of which occurred. But the report itself was not published. And, so far as specific findings through definite doubts on the safety of hundreds of convictions, the FBI told only prosecutors, who, in more than a half of the cases, somehow forgot to pass the evidence on to defence counsel.

The investigation reviewed only a limited number of cases and focused on the work of only a few scientists. But there were indications that problems were far more widespread and could affect thousands of cases in federal, state and local courts, the Washington Post report, by Spencer Hsu, says.

‘‘... Hundreds of defendants remain in prison or on parole for offences that might merit exoneration, a retrial or retesting of evidence using DNA because FBI hair and fibre experts may have misidentified them as suspects. In one Texas case, Benjamin Herbert Boyle was executed in 1996, more than a year after the Justice Department began its review. Boyle would not have been eligible for the death penalty without the FBI’s flawed work, according to a prosecutor’s memo.’’

Hsu details a number other cases where there is now profound doubt about forensic evidence establishing guilt. He quotes Justice Department lawyers saying ‘‘they met their legal and constitutional obligations when they learned of specific errors, that they alerted prosecutors, and that they were not required to inform defendants directly’’.

These are not cases where police and forensic examiners have necessarily decided to frame the innocent. They are, however, cases of manifest human frailty, where witnesses have let their sense of certainties overwhelm their capacity, competence and judgment. A witness in a white coat is, all too often, an advocate not a servant of the truth.

Does this mean it’s all nonsense? By no means. But it does suggest that some forensic evidence warrants a higher level of judicial and police scepticism. The public needs more quality assurance, accountability and regular external audit than it is currently getting. There is ample evidence, here as in the US, that prosecutors, police and forensic scientists are not doing enough to warn of the limitations of their science.

Jack Waterford is The Canberra Times' Editor-at-Large