Roboticist Srikanth Saripalli makes this interesting point: "If the government has to decide what to fund and what not to fund, they are going to get their ideas and decisions mostly from science fiction rather than what's being published in technical papers."
Because at NIH indeed you are placed on a grant review board because of your techical knowledge of the matter. On top of that, the applications are all supported by citations in technical (and peer-reviewed) papers.
As best I understand funding at DOE and NSF works much the same way; your odds of getting funded are astronomically better if you have good primary literature to support the experiment you propose. Now, if your funding plans revolve around convincing your favorite congress-critter to write in a line (or a full bill) to get you some money, that might work too but it generally isn't the most reliable way to establish a career path.