I found this paper after a brief reflection on twitter  about how we as humans tend so massively to undervalue the power of imagination in favor of pure inductive/deductive logic.
I think the scientific paper or report has always been fictionalized to some extent. The earliest scientists may have been more overt about it. It's doubtful that a bunch of soldiers ever tried to burn an enemy ship with mirrors, but that's how Archimedes reported his idea. As I understand it, it's also doubtful what experiments Galileo actually performed on falling bodies before reporting his results.
It seems to me that in a typical paper, one should strive to make the work reproducible, to present the ideas in a form where errors shine out if they are present, and generally to be useful to others.
Using the scientific paper to reveal how science is actually done isn't necessarily a good use of paper or readers' eyeballs. We're bad enough writers as it is, and trying to express a "narrative" in every paper may turn the literature into an unreadable and possibly irreproducible mess. There may be better ways of learning the "narrative" of scientific research, namely by just getting in there and doing it, or at least hanging out with scientists for a while.
Abstract, results, methods, conclusions.
Experts can glean a lot about the methods from the results, and what not can be obtained from the methods. The conclusions are usually debatable, and mostly in the abstract.
The significance statements are worthless wastes of time.
I don't think this is true. Do you have any data or even any examples to back this up?
Early career researchers aren't well known, so if they publish a bad paper who cares? No one knows who they are anyways and likely won't remember for their next paper. Well established researchers, are well, well established... And can weather these papers by the bulk of the rest of their work.
I don't think i ever judge the quality of a paper by the quality of the other papers the authors have written. I rarely even look to see what other papers that author has written.