I am continually amazed by how many people completely blow this section off. In theory, NSF weighs this portion of the grant on equal footing with the science. I know that this doesn't happen in practice, but they do actually care about it. I finally got the last of my grant reviews off my desk for this round and I saw nothing but the bare minimum of effort put into this section, and you know what? I called people on it in my review. Since there has been some recent advice about grant writing around here, I thought I would put together my thoughts on the broader impacts section for those of you writing NSF grants out there (and other agencies might have similar requirements).
Read the guidelines on what NSF is looking for and make an effort to meet their requirements! This may seem really obvious, but almost every grant I read this round did not do this. The criteria are as follows:
•How well does the activity advance discovery and understanding while promoting teaching, training and learning?
•How well does the proposed activity broaden the participation of underrepresented groups (e.g., gender, ethnicity, disability, geographic, etc.)?
•To what extent will it enhance the infrastructure for research and education, such as facilities, instrumentation, networks and partnerships?
•Will the results be disseminated broadly to enhance scientific and technological understanding?
•What may be the benefits of the proposed activity to society?
If your proposal details science that blows my socks off, I'm not going to care if you don't put in much effort to your BI section because it should hold up great science. But the other 98% of us, make a fucking effort. You need to at least touch on most, if not all, of the above points. Some of the points can be addressed quickly - for instance, saying that all sequence data will be deposited in GenBank - but it is a good idea to deal with each one.
Do not use the BI section to talk about how much your science will affect other fields! This is not what NSF means by broad (see above). You should bring this up, but in the intellectual merit section.
Commit more than a few sentences to this section, preferably a page or more. When turning in one's review, there is a separate section to comment on the BI merits. Give your reviewer something more to talk about than a paragraph.
If possible, it is a really good idea to include some money in your budget for your BI goals. I know it seems odd the NSF would want you to add money into a budget, but money = accountability. If you put money for a workshop into the grant and it is left over at the end, they can ask you why you didn't follow through on the BI. If you promise to organize a symposium at some conference (which screams no BI effort, BTW), there is no way for NSF to know whether or not you followed through.
Partner with existing programs at your institution. This makes your life easier because the existing program will likely write part of the BI section and help organize whatever it is that you are proposing. Also, NSF like to see cross-talk between researchers and on-going programs that they have already put money into. Even better is if you can say that you will provide half the money for XXX and have the existing program kick in half. Again, there is a financial commitment from both sides, indicating a willingness to partner.
Make it viable. There is a delicate balance between doing something worth doing and proposing something that will suck up more time than it should. This is where leaning on infrastructure already in place will allow you to get more done for your time "buck".
It takes a bit of creativity and some talking to some of the centers or programs at your institution, but it is really not difficult to come up with a BI section that will make reviewers say "that'll work". So few applications actually put in any effort, that those which do stand out.