Grant Writing and Computation Linguistics: Part Two

Our previous post discussed the value of using word frequency data as a computational linguistics tool to identify proposal hot buttons and distinctive features.

In this follow-up blog, we discuss a second application of this technological tool.  Specifically, do the exact same analysis but this time do it on you penultimate proposal draft.  The first time you did it, you looked at the proposal guidelines.  Now, repeat the process using your latest working draft. 

The two sets of frequently occurring words should look similar.  They may not be in the same rank order but considerable overlap should exist, especially among the first 15 or so items listed. This way, you can ensure that your proposal reflects the hot buttons and distinctive features identified in the application guidelines.

This spring, we were asked to critique a penultimate draft of a federal grant proposal.  We first analyzed the RFP using this computational linguistic approach and ended up with the hot button list.  Next, we did the same thing with the proposal draft.  The result?  We found a 15% overlap between the two lists.  The message was very clear: the applicant was missing a number of the key concepts identified in the application guidelines.  We lifted up the missing elements and asked the applicant to make changes in the proposal text.  The revised version came back to us a bit latter and it showed a 92% overlap upon our re-analysis.  The work paid off with a six-figure grant award notice and one very happy organization.

Leave a Reply