80% crap? No. Science papers either generate more questions that are followed up by colleagues and competitors, or are crap nobody is influenced by, nobody cites, and nobody reads. Mistakes get caught--the second paper I wrote as a graduate student tore a well established professor a new asshole over a mistake he had published. I didn't care that I was a nobody and kowtow to the bigshot, I saw a mistake made on what was my thesis project and went for the jugular. I replicated the results, added more data, and aside from minor edits from my advisor I wrote the paper all on my own. I didn't do anything special and I'd expect the average grad student to do the same. We're trained to take the rank of somebody with a grain of salt including all the way up to Nobel laureates. A lab I was in collaborated with one a few years before he earned the prize. We (the grad students and postdocs--the bottom rungs of the scientific ladder) thought of him as somebody who was very bright, very aggressive, and who was almost always right but definitely did not dot the i's and cross the t's. If he made a mistake (rare) or went too far too fast (often) we made a note of it, knowing in our supporting role we'd have to run extra experiments to check it out. None of this is special, it's everyday workaday science from a nobody in the trenches.
Contrast that with an attempt at conspiracy. A nobody, one who's already gotten used to not trusting everything a Nobel laureate or National Academy member says, is going to tease that apparent mistake apart, find another, and another, and then smell blood and scientific glory in equal measures. And that nobody is pretty much everybody in science.