Slightly over a year ago, my advisor and her former post-doc realized that there was an error in a published paper. They insisted that the finding was valid, especially since the error was found in the process of replicating the results; I insisted that the statistical significance of the results had changed to the extent that the paper would never have gotten published, and the editor should be informed. This was incredibly stupid of me.
It wasn't stupid to think that the editor should be told; I stand by that bit of ethics. But it was definitely stupid to think that I could get two professors to admit they were wrong. We got a letter to the editor drafted - and then it had to be a brief report, so they could explain all the interesting reasons they were wrong. And then my advisor sat on the draft for five months, until I demanded to at least give the *#@%A thing a due date so it wouldn't always be the lowest priority. And the the due date came - and went. My advisor has a grant deadline. Our co-author has "a number of pressing deadlines". Having a due date doesn't even give them an artificial sense of urgency.
I find myself faced with the options of continuing to drag them, kicking and screaming, into telling the editor they made a mistake, or giving up. I chose to give up. The alternative is just too stressful; I've spent a year trying to convince them to do the right thing, and I don't have anything left to fight with. Their reputations matter more than their science, and there's nothing I can do to change that.
Frankly, I don't think our research is important enough that it matters, and we did replicate the results. My advisor and her former post-doc can survive with their reputations intact, in everyone's eyes but mine - and, after all, why should they care about the opinion of a lowly graduate student?
As for me, I have one more year here. Fighting with my advisor over this will only hurt my career prospects. I just need to get the dissertation done, get the PhD, and get out. If I have to publish my dissertation results - well, at least my advisor is likely to be interested in those. And then I'll never work with her again.