There's an arms race happening in the world of NSF broader impacts these days. Where it was once okay to basically describe the parts of your job that contribute to training (grad and undergrad student mentoring, incorporation of current research into the classroom), the bar for what constitutes a successful broader impacts section of an NSF proposal continues to rise. Everyone is looking for new and creative ways to satisfy this requirement without committing huge amounts of time to the endeavor.
Potnia Theron has a post up about an NSF panel and the discussion there regarding blogging as an acceptable BI activity:
The blogging people who got good scores on this were ones who had blogged for a while and who could demonstrate significant traffic (by statistics). The PZ Myers of this world (whatever you think about his politics and his feminist credentials) do well, and the guy who is just starting a blog about the sadness of endangered species does not.
This reflects my experience as well. At the last panel I was on, reviewers had checked the blogs or websites being held up as BI contributions and found almost every one of them to be unremarkable in content. Many had only just been started, with a single or a few posts. That's not the kind of thing that gets any BI cred. The people who had something established and could give traffic numbers were taken more seriously. And no, posting a bunch of risotto recipes doesn't count.
For those of you considering taking this approach, it would be a good idea to decide what you want as an outcome of blogging and how you might measure success. Are page views a good yard stick? Unique viewers? Clicks to links of papers? What are you going to use to demonstrate that your writing is reaching your audience and not just racking up Google hits because you have the phrase "nude cats" in many of your post titles?
With the right approach, blogging is starting to become an acceptable form of outreach. But just like anything else in the BI section, it needs to be thought out and have measurable outcomes or else it's just another empty promise.