I think from a practical standpoint, the science community has fallen into the hole the educational industry has: Relying soley on references on papers. If you have sufficient number of references, your paper gets a pass regardless on your data. This means you can easily piece together a paper that contradicts itself but still get published if you pet enough egos in the process.
I've found from my observations toward universities, even right down to high school level that it is taught that unless you reference a "known name" your work is crap. Many students (and even professional scientists) are not allowed to question any works produced by these "known names" unless they are a direct peer. This means even if one can with substancial data prove both Albert Einstinen and Stephen Hawking totally wrong in every respect with emprical evidence, that they would not be allowed (and in fact be shunned/banned from scientific groups) because it goes against "the norm". Take note of many revolutionary inventors in how hard it was to get even their practical empirical experiments acknowledged because major science panels did not want them disrupting what was the common belief to be true established by various notible names.
At this point and how referencing is regarded now, referencing is the scientific and acedemic community's way of enforcing a status quo. It is no longer about proving yourself or your data, it is about providing another column to another work to ensure that that other work is made "more unquestionable". Only if you're high enough on the food chain do you get columns supporting your work.
Referencing needs to get back to being a secondary measure of solidity where one's own data along with the mode of how an experiment was performed is the first, where if one wants to challenge, one can say "The mode of experimentation is faulty because..." and to intrinsicaly challenge the work on its own merits, THEN to use the references to back that up. Not references first then experimentation second.