23 karat gold, or is silver good enough?

Aug 09 2014 Published by under Uncategorized

There was a twits exchange in response to a bit that DM had in a comment someone else wrote about a post he wrote about what makes success in grant writing.

Original DM:

@pottytheron: from @drugmonkeyblog Perfecting that one gold application gets in the way of submitting more of the silver ones. So it is an error.

Reply from @MTomasson: @pottytheron @drugmonkeyblog Silver grant applications don't get near fundable scores.

some more back and forth.... and the source of the title of this post:

@MTomasson: No. But we can both be right..different fields. In mine, 23 karat gold is an also ran. @drugmonkeyblog @pottytheron

Part of the problem is what do we mean by 24K gold? My interp of DM's foolproof method (or full proof if you're imbibing tonight), is that there is a diminishing return to putting effort from going to 23 to 24 K (which is still not 100 % gold, I should point out). I did a post on this once upon a time.

When Michael Tomasson says that "23K grants don't get funded", I am not convinced. There are lots of 24K that don't get funded either. (according to the PI's). What the study section thinks is largely a mystery, except of course, for the pink sheets.

And that's the important point. What YOU, the PI think is GOLD is not necessarily what the STUDY SECTION thinks is gold.

I have put in grants that are not just 24K gold, but embellished with the finest gems and prime numbers. They get short shrift. Some of the ones that are a bit less than sterling, however, have gotten funded. Trust me, I am old enough to have a statistical sample.

In a true cost-benefit analysis, any PI needs to ask: is the time to take this from 95% to 99% worth the cost in units of my effort, which I might alternatively produce another 92% proposal that in fact, the study section will see as the most incredible thing ever written (for reasons that escape me).

So in answer to Michael I say: neither you nor I know what the study section thinks is 24K.

What constitutes one's best is not a univariate measure. It is a complex function of many compromises. Do you wait to do that extra set of experiments for prelim data? Will that data make or break the proposal? Do you include this collaborator or that? This one will hand me a letter tomorrow, but wants 10% salary, that one is taking longer, but won't want so much, and I am damn near the top of the budget. Include this aim or that one? And of course the ultimate decision: for an R01 you've got 12 pages, what to leave in, what to take out?

My take on DM's post (and I've had enough of Scotland's finest to not be able to cut and past the link on the relatively small device aka phone. Damn, I've written this on my phone. Probably full of typos. Once again, Bette Middler to the rescue Fuck 'em if they cant take a joke). Anywho... back to DM's post. His advice about persistence is very good. It is a long game. But it is a long multivariate game.

No responses yet

Leave a Reply