There have been several recent stories documenting the recurrent incompetence and dishonesty of crime labs throughout the country. [See here, here, and here.] The most recent ABA Journal has a good discussion here. Those crime labs supposedly provide independent scientific evidence supporting an accurate determination of guilt or innocence when, in fact, they most often work in support of the prosecution. As these articles repeatedly point out, crime labs are often hardly “scientific” in their methodology and frequently exert themselves to find empirical support for a conviction; even when such support just isn’t there. Conversely, crime labs have often withheld exculpatory evidence from the defense even if it results in a wrongful conviction.
The alliance between crime labs and the prosecution appears even more suspect in those jurisdictions where crime labs are completely reliant on the fines collected from those they help to convict. In other words, there’s no monetary incentive to be honest. The problem is discussed here at Overlawyered.com and by Radley Balko here; both of which rely on the work of Roger Koppl and Meghan Sacks in the Journal of Criminal Justice Ethics here.