crisp

 Addendum 2 May 2008:

I have been alerted to a beta version of NIH's CRISP site, now termed CRISP-RePORT (Research Portfolio Online Reporting Tool), that already addresses the concerns below and makes several other improvements! Kudos again to the NIH for making the entry below basically moot. The NIH may be a behemoth bureaucracy, but it really does seem to have the S.C.I.En.C.E. ethic behind many of its initiatives...

Submitted 25 April, 2008 by Noam Y. Harel

Query NIH CRISP

The link above allows you to search through thousands of US government-funded grant proposals from 1972 to today. Searches may be keyword-based or limited to specific Institutes, Study Sections, etc.

A better CRISP through Crowdsourcing 

Kudos to the NIH for allowing public access to its funded grants by organizing CRISP-Computer Retrieval of Information on Scientific Projects. However, even as the largest current repository of (funded) grant proposals, CRISP could be so much better - the interface is not very intuitive or user friendly; the links quite often lead to a '404 Not Found' error; and of course, CRISP limits access to abstracts, only for those proposals that have already been funded.

Fortunately, many of CRISP's flaws could be corrected very easily, without spending extensive computing resources. The Sunshine Project did just that, creating 'CRISP-ER', for Extended Results. CRISP-ER collected and collated NIH grant data from CRISP and other sources, providing a more comprehensive search tool. However, CRISP-ER has recently ceased operations.

If the NIH made CRISP's raw database server openly available, everyone would benefit - interested scientists and lay citizens would be able to create a much more user-friendly tool to search through US government-funded grants; and the US government would more effectively convey to the public where its tax dollars are going. This would correlate exactly with the NIH's stated goal of "... directing programs for the collection, dissemination, and exchange of information in medicine and health..." (see NIH's mission link).

There are three alternative routes toward a better CRISP:

• • NIH invests in-house resources to improve CRISP.

  • NIH shares its raw database of funded proposals publicly, in effect outsourcing (or crowdsourcing) the CRISP-improvement project for free.

  • NIH does nothing - a progressive entity such as S.C.I.En.C.E. has to devote extra effort to extracting and rearranging the CRISP information into a more useful, open resource. This should be possible by initiating an automated recurring CRISP query that would obtain all the information necessary to reformat CRISP from the outside.

Clearly, the second option would maximize the benefit and minimize the effort required on all sides. Hopefully, the NIH sees it the same way - this proposal has been emailed to CRISP as well. In the meantime....

Make your suggestions for a better CRISP search tool

On S.C.I.En.C.E.'s companion Wikispaces site, please voice your opinion on how to make CRISP (and all grant proposal repositories, for that matter) better. All suggestions will be forwarded to the NIH.

For starters:

• • Use a simpler search algorithm for basic searches - easily accomplished using free Google search tools.

  • Use a more reliable server to host the database, reducing the frequency of '404 Not Found' errors.

  • Don't limit the number of records listed (CRISP's default is 250) - just list all hits, broken down into pages of 10 hits each.

AGAIN, PLEASE ADD YOUR SUGGESTIONS AND YOUR COMMENTS ON OTHERS' SUGGESTIONS AT THE S.C.I.En.C.E. COMPANION WIKISPACES SITE. THANKS!