In practice, though, Broader Impact is usually about education, public outreach, and inclusion. A researcher submitting a proposal to NSF would be well-advised to incorporate some aspect of their research into a course module (preferably one that can be easily adopted by other instructors) or a presentation to grade school kids, and include some members of under-represented groups in their labs. These are fine things, things that are often worth doing. HOWEVER (you knew there was a "however" coming) in a country with lots of instructors if every instructor out there is developing course modules and trying to get everyone else to use them, well, that's more modules than we need, and the quality will be variable. Outreach is fine, but some people are better than others, and frankly the occasional dog-and-pony-show at grade schools is probably not the biggest thing that we need if we're serious about improving k-12 science education. Moreover, inclusivity in a research group is a fine thing, but even that is better addressed at the level of admissions committees than individual research groups. (Though I do acknowledge that tying it to funding for individual labs creates a bottom-up pressure that can matter.)
I've pointed out some of the drawbacks here, and other people have documented just how confusing and contradictory the implementation of this criterion can be. I do freely acknowledge its upsides, of course. What's more interesting to me, for the purposes of this blog, is not the "on the one hand...on the other hand..." stuff, but the history of it:
From 1981 to 1997, NSF guidelines identified four criteria for the evaluation of proposals:One could note that from the 1990's onward we no longer felt that our chief geopolitical problem involved an adversary with world-class nuclear physicists and rocket scientists. In the 1990's our chief geopolitical concern was, um, actually, nobody really knows. The 90's were a weird time. We did "humanitarian" interventions against penny ante-foes and worried about French industrial espionage. Since 2001 our chief geopolitical concern has been people whose arsenal primarily consists of improvised explosives, rifles, and box-cutter knives. Yeah, yeah, Iran and North Korea, but Iran is more of a diplomatic issue and North Korea's nuclear program is even less sophisticated than that of Mao-era China.
● Research performance competence.
● Intrinsic merit of the research.
● Utility or relevance of the research.
● Effect of the research on the infrastructure of science and engineering.
Since 1997, however, NSF has used two criteria for the review of grant proposals: one focuses on the “intellectual merit” of a proposed activity, while a second asks for evaluation of the “broader impacts” of the research.
Anyway, in this era where we no longer worry about adversaries with world-class nuclear physicists and rocket scientists we are quite comfortable trying to bring democracy to science. One constant theme of this blog is the tension between academic excellence and democratic values. One can quite easily resolve those tensions by viewing the academically successful as simply having a place in society but not viewing academia as the path to prominence in society. It means that you'll have to make place for the middle class (and especially the lower-middle class) on their own terms, in an economy that needs them. Alternately, one can engage in self-deception and deny any tension between academic excellence and democratic values. Broader Impact is, in some sense, NSF's attempt to do that, and de Tocqueville would no doubt recognize it as such.