misleading testimony. They were computing their statistics incorrectly - in a way that was biased against the accused in many cases."

In some cases, Thompson found simple errors where documentation said Sample A matched Sample B, for instance, which was untrue. There were cases where Thompson found inconsistencies between the lab report and what was said in court.

"Along the way I encountered this case of Josiah Sutton, where not only did the lab work look sloppy and bad, but it looked like they'd misinterpreted the results in a way that they'd reported something as incriminating when it was exculpatory."

In addition to improper or incomplete lab work, the work wasn't being reviewed by anyone, and in many cases the results weren't available to the defense until trial, at which time it was sometimes declared too late for the defense to get an independent analysis of blood or DNA evidence - if there was any blood or DNA left.

Like most labs, the HPD Crime Lab could barely keep up with its workload because staff were underpaid, lacked training and weren't properly supervised.

In the Sutton case, a lab analyst used "poor technique," and had insufficient training, according to the Bromwich reports, www.hpdlabinvestigation.org, one of the independent looks into the lab. The analyst produced "ambiguous results, reflecting complex mixtures." The report said the lab's murky work was exacerbated by its practice of not accurately explaining just how inconclusive the results were.

During the Sutton trial, for example, the lab analyst testified that the DNA found on the victim was an exact match to Sutton's, when in actuality 1 in 16 black men would be a similar match, according to Thompson. Further, the analyst used all four vaginal swabs, not a necessary or advisable practice, limiting the possibility for retesting.

Similarly, in a 2001 kidnapping and sexual assault case, the crime lab analyzed semen samples taken from the victim's cheeks. The analyst used all the samples, eliminating the opportunity for an independent review by the defense. The report and testimony didn't acknowledge that the samples were "very difficult forensic examples," according to the Bromwich reports. The lab came up with clear-cut results and created the impression that the defendant was the only person who could have contributed to the DNA sample taken from the cheek swab, which was incorrect.

Although most of the errors can be attributed to sloppiness, incompetence and lack of training, Thompson found the lab almost always erred on the prosecution's side. Lab analysts cut corners by not using control samples to make sure there was no cross-contamination between two samples - a common problem in crime labs. "When I started looking at this, that was the first thing I saw," Thompson said. "Where are the controls? The answer is they didn't have any."

No Safeguards

And there was nobody to review or catch the problems - no safeguards in the system.

"To the extent that labs are doing bad lab work, there's a whole series of points at which that should be caught," Thompson said. "The criminal justice system in Houston, and much of Texas, doesn't really function effectively to screen and evaluate scientific evidence, or probably many other kinds of evidence."

Thompson said there were times when well funded defense lawyers would file discovery motions and go to court to get access to lab materials and hire their own lab experts to test the materials. In many cases they found problems. "So it wasn't that nobody knew the lab was doing bad work, because the lab had been shown to be doing bad work over the last 10 to 15 years. Once the well funded private attorneys would win their cases, things would go back to normal - the rest of the lawyers

Jim McKay, Justice and Public Safety Editor  |  Justice and Public Safety Editor