tag:blogger.com,1999:blog-8890204.post427748167370258964..comments2024-03-10T05:26:42.148-04:00Comments on My Biased Coin: Service and the NSFMichael Mitzenmacherhttp://www.blogger.com/profile/06738274256402616703noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-8890204.post-29618324573671279402010-03-08T04:25:42.718-05:002010-03-08T04:25:42.718-05:00You can telecommute to NSF meetings. I did it las...You can telecommute to NSF meetings. I did it last year. They put a speakerphone in the middle of the table and I called in with skype. It was reasonably successful, although I had to often ask people there to speak more loudly.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-8890204.post-68283107579794778022007-11-26T02:17:00.000-05:002007-11-26T02:17:00.000-05:00Your thoughts on serving on an NSF panel are inter...Your thoughts on serving on an NSF panel are interesting. I served on one last year for the first time, having finished graduate school several years ago. My motivation was essentially to gain insight into the whole process of how proposals are selected. <BR/><BR/>I found that the discussion there actually did drastically change the course of what would have happened to a significant fraction (1/4) of the proposals. In most of these cases, I doubt that the same result would have occurred if each reviewer just posted their comments online and read those from others (in which case I think people would be unlikely to alter their perspectives much).Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-8890204.post-46186484528827028912007-11-10T14:38:00.000-05:002007-11-10T14:38:00.000-05:00This post raises an issue with "service" in genera...This post raises an issue with "service" in general. For example, there is little motivation for senior academics to serve on program committees, and even less for them to do journal reviewing (as opposed to being editors). <BR/><BR/>Tenure and promotion committees do pay lip service to "service", but it is interpreted very broadly and is, in general, much less important than good research. <BR/><BR/>In some cases, such as editorships and a few program committees, the visibility provided by a service position is a partial reward. However, NSF panels and reviewing definitely do not have visibility benefits. <BR/><BR/>How can we create an incentive to encourage low-visibility service activities? <BR/><BR/>This is a bit like creating an incentive for people to be nice. Studies (and experience) tell us that humans are nicest when somebody is watching, since we use kindness mainly as cement for social structures (families, communities, etc). This doesn't work at the NSF, though, if the program manager changes every three years (why create a bond with someone who is leaving?). And it doesn't work with journal reviewing if nobody except the editor knows you're doing it. <BR/><BR/>As the post said, requiring one panel membership per grant for people with NSF grants would be problematic. A hard constraint like that would drag the program manager into creating a panel schedule that uses everyone exactly once, while simultaneously maximizing the quality of reviewing in each sub-area... sounds NP-complete. <BR/><BR/>But you could imagine a looser coupling: the NSF could keep a record of invitations made to serve on panels vs invitations accepted, or simply panels served on per dollar/year/unit of grant money. The program director could then be allowed to explicitly take this institutional memory into account when deciding which grants to recommend. Of course, this system is open to abuses of different kinds, but one could imagine guidelines that would keep things reasonable ("if you've *never* served on a panel, and you have X or more grants..."). <BR/><BR/>In the crypto community, there is an informal version of this that is used to help program chairs put together their committees. John Black at Colorado maintains a list of how many papers different authors have at CRYPTO/EUROCRYPT authors, together with how many times each author has served on a PC. Program chairs can look for underused researchers (and avoid abusing the nice ones who can't say no!). Perhaps the NSF could use something similar internally. <BR/><BR/>All this may be far too naive to work in reality. I have never served on an NSF panel myself (I just took up a faculty position) so my ignorance about the process is, um, significant.<BR/><BR/>-Adam Smith, Penn State<BR/><BR/>(NB: I don't know if John created his list with this in mind, but I know from talking to program chairs that it gets used this way, at least occasionally).Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-8890204.post-49172235314433327532007-11-09T11:15:00.000-05:002007-11-09T11:15:00.000-05:00One small item to remember: Make sure that you ar...One small item to remember: Make sure that you aren't getting paid on a grant during the days when you serve on an NSF panel. That counts as double-dipping.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-8890204.post-32959294246177660362007-11-08T16:13:00.000-05:002007-11-08T16:13:00.000-05:00I think there are several even better and more "pr...I think there are several even better and more "pragmatic" reasons the serve on the panel:<BR/><BR/>1) you see how decisions are made, which can help you in your own grant applications. Not necessarily to "game" the system (although this too), but just to see what gets funded and what does not.<BR/><BR/>2) you can try to make sure that some morons who know how to game the system don't get the money (instead of stronger, but more self-conscious researchers).<BR/><BR/>3) most importantly, and related to 2), theory program is not the only program open to theory people. When you serve on other panels (such as cynertrust, nets, GENI, etc.), you want to make sure that theory gets proper representation and respect. Ultimately, this is what will improve the funding situation in theory.Anonymousnoreply@blogger.com