Shocking degree of bias in medical expert testimony: Radiologist physicians who testified on behalf of plaintiffs in asbestos lawsuits saw significant abnormalities in nearly 96% of the patients' chest x-rays. However, when the same set of 492 chest x-rays were reviewed by a panel of neutral radiologists, significant abnormalities were described in less than 5%! More infomation is available here at this Nature article and this Rocky Mountain News article.

Since this is my field of practice, namely diagnostic radiology, I find this deeply disturbing. Of course, it would be very interesting to know if the results were due to a few unscrupulous physicians that were disproportionately represented in the ranks of these so-called "experts" because the plaintiffs' lawyers preferentially put them on the witness stand, or if there was a more widespread unconscious bias that affected even (otherwise) good radiologists. In theory, the opposing counsel can challenge the credibility of bad "expert" witnesses and attempt to get the jury to discount their testimony, but I don't know how effective this is in actual practice.