 I was on a grant NSF panel recently that decided on which grants were going to be funded and given the low levels of funding at NSF I was surprised by how many really good proposals there were. About 40 of the 138 proposals in this panel in my opinion really should have been funded and would have been funded not that many years ago. However, of those 40 truly superb proposals, only about 12 ended up getting funded. And I was surprised, you would even say shocked, to find out that it takes great science to get into that top 40, but which of those top 40 actually got funded turned out to be determined in part by data management plans, postdoctoral mentoring plans, and the broader impacts of the research. So things that are not directly related to the research project itself ended up being an important determinant of which grants were funded.