FancyMyBlood:
Yes, i have a negative bias against JM - the same negative bias i would have against any scientist who engages in the kind of misrepresentation of their work that JM and Ruscetti have admitted to and defended.
'Sloppy' doesn't necessarily mean bad science - although it is harder to do good science if one is sloppy with one's work - and one doesn't have to be sloppy to be wrong. Silverman did good work on RNase L and innate immunity, but it seems very likely that he was wrong about a retroviral association with prostate cancer. Being wrong doesn't make it bad science either. Silverman impressed me with the PCR contamination work. He stood up and said, 'hey, this stuff doest make sense,' did the careful work to show his earlier result was wrong, and why it was wrong, and retracted the incorrect result. THAT is good science.
He also documented the stunningly paranoid precautions he took to avoid contamination. Mikowits has blamed Silverman for the contamination, claiming it happened after she sent him the samples - I don't believe it.
Silverman's work showed that the contamination was widespread in patient samples, but none in healthy controls. The charitable explanation for this is sloppiness - that the patient and control samples were handled so differently that patient samples got contaminated while controls somehow were safe from contamination. The less charitable interpretation is intentional spiking, but there isn't evidence to support that strong a claim.
Against this backdrop of patient-specific contamination, there is the fact that JM was able to reliably get PCR positives from patients, and only low levels form controls as long as she knew which samples were patients and which were controls. When she was sent blinded samples, she still found positives -but they were equally split between patient and control. Again, the charitable interpretation is that the patient samples were handled differently from control, in a way that allowed contamination of patient samples. And that when she got blinded samples, she handled some of them the way she handled patient samples, and some the way she handled controls, so the contamination got into both sets.
Again, I'm being charitable.
But what she didn't do is what Silverman did - say 'hey, no on else is confirming our work, perhaps I should blind samples and check myself, and figure out what the issue is." It didn't get found out until blind samples were pushed on her from outside.
THAT is a problem, and that is the difference between how I perceive Silverman and Mikowits.