The problem is somewhat recursive. Ideally, politics wouldn't get in the way of accumulating the data. Ideally, unverified data would stay out of our public/ins

The problem is somewhat recursive. Ideally, politics wouldn't get in the way of accumulating the data. Ideally, unverified data would stay out of our public/institutional/organizational decision-making processes until the dust has settled (studies reproduced, methodologies fully critiqued, synthesized into frameworks that square large amounts of high-quality data, frameworks withstand decades of counter-proposals and attempts at falsification...). But it doesn't really work like that. If the results sound salacious, they'll get picked up in the science/popular press. Someone else might destroy it later, but no one's going to update every PDF, abstract and bit of news coverage to note it was destroyed. Even if the destroyal is so salacious that everyone trying to use the result as a buttress for their favorite ideological/policy positions sees it, they have to have the intellectual integrity to stop trying to use it that way. And people they're using it on have to be curious enough and capable of finding out that it's been destroyed. (To be clear; I don't think this means we should avoid researching uncomfortable/controversial topics.)