There's also the responses of the PI and the PI's mentor. A real scientist doesn't take such questions as "a personal insult" as the now almost completely disgraced PI did (state medical boards that are desperate like North Dakota will still license the PI as a doctor, at last count). The PI's mentor wasn't that sort of unprofessional, but the article says he didn't respond well to this. Note also one of the Duke higher ups, apparently unprompted, introduced the word "misconduct" in all this.
If you're going to be a "life and death doctor" as I put it, for him one of the harshest examples of that, oncology, you're going to be making a lot of decisions that patients lives hinge on. A degree of certainty and decisiveness, what some call doctors' "God Complex", is required. I judge that he was sufficiently certain in his judgement of the research, and its implications for patients being treated based on it at the time. It's that last bit that I believe requires you to go above and beyond in determining if you're correct or not, and then doing what's right for the patients.
And for that matter it's inverse: http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect
You're basically asking the junior developer to question all the senior developers, the lead developers, and the executive managers. There is only so many people a person can say are wrong. It's not possible for him to know everything, for all he knew he was wrong, and the professors were merely unwilling to take the time to explain it (or bring in an outside professor to explain it). That smelled wrong enough for him to get out, but he probably had a nagging thought in the back of his mind that he simply didn't understand.