including annie dukin's in boston. they raping from downright fraud to a broader array of difficulties. sometimes it's not intentional fraud, sometimes it's not dry-labbing or, you know, pretending to do tests you haven't done at all, but rather something more inadvertent, which is the kind of bias that can come from feeling like you are part of the side, part of the team - that you are - you are attached to the prosecution, and you want to get the bad guy. that can lead honest analysts to make mistakes. >> this happens with some frequency when you look at the d.n.a. exonerations. 55% say there's some sort of error, improper use of forensic science that leads to the cases. >> that's right. faulty forensic science, whether it's outright fraud for something more limited, but equally disturbing, has been a significant contributor to a few of these exonerations. that's right. >>> what happens, why is it? you mention a team approach, a mentality. what would solve that? >> it is. one of the real dangers is a mentality. in mo