The whole “publish or perish” thing is what drives most scientific fraud. Realistically, most science gives negative results or otherwise fails. But despite the failures being just as useful as successes, nobody is interested in the failures. So everything has to be framed as a success. Often this is possible, but not always. 

If you’re months or years into a research project and find that you’re stuck with essentially nothing even remotely publishable, through no fault of your own, you’re screwed if you’re not a tenured university professor (and these days most people doing the most risky/speculative research aren’t). It can ruin your career entirely. So a lot of BS gets published. 

I have personal experience with this and let me tell you, it is heartbreaking to realize that the research you’ve spent months/years working on is useless. I decided not to BS it and did not try to publish my master’s research. I’ve been unemployed ever since and my chances of going back to school are not the greatest.

Or as they say, “Inconclusive. More funding required.” When the funding dries up, the published content may be refutable, with an argument that, hence the lack of funding to back up the premise makes it refutable. Such is the swindle.

* excluding, of course, the whole corrupt world of corporate scientific findings-for-hire, in which you can pretty much reverse everything I’ve said in this comment.

For actual research you have to distinguish between “cherry picking”  data (e.g. throw out all patients that didn’t respond positively to a drug) v.s. not including say some measurement that doesn’t seem to support the theory or go against it  that you have no understanding of or explanation for.