Occasional commenter microfool has pointed to a very interesting powerpoint slideset in a comment over at MWE&G. The comment referred specifically to an observation from writedit that NHLBI is planning to phase out special, more lenient funding line considerations for those New Investigators who are not Early Stage Investigators* in 2010. The discussion then landed on the distinction between ESI and New Investigators in NIH-speak and microfool brought some data to the discussion.
You can see hints of this unintended consequence on slide 11 of this slide deck where we see nice little blips in ages of First Time Investigators at ages 60, 66, 72, and 87.
Let's take a look...
My read of these data seems to falsify my prior assertion that the New Investigators being funded are disproportionately the technically-qualifying, established investigators who did not happen to have prior NIH funds.
Funny, I noticed this trend after about two rounds of study section. Namely that the New Investigator pool included many quite-senior investigators who had enjoyed quite substantial support from nonNIH sources in their careers. Surprise, surprise, proposals from these "New Investigators" did better than did those of recently transitioned New Investigators. I still find it mind-boggling what properties of the NIH grant review / award system seem obvious to applicants and study section reviewers and yet is a "surprise" to the people who actually hold the raw data that would allow them to perform definitive analyses.
Slide 11 [source]
This graph certainly questions the generality of my subjective impression. Of course, this depends on a couple of assumptions. First, that the "First Time Investigator" designation on this particular slide set from the NIH is equal to the "New Investigator" designation of application checkbox fame. This latter was only started around 1999 (see slide 12 of the powerpoint file) so there may be a decent reason that the NIH uses another term when talking about data that predate this checkbox designation. Second, the data are grouped by age and we must therefore do some highly subjective speculating about how old investigators might be before receiving some sort of substantial funding. It would be very strange to me, for example, if a 60 year old investigator were competitive for NIH R01/equivalent research funding without some sort of prior major research support. Likewise it would not be typical for a 35 yr old investigator who is competitive for NIH funding to have had a major research award from another source.
So I could still be right, there is a considerable amount of area under the curve after about 40 years of age. And although it is not typical by any means, yes there are some lucky individuals who manage to secure substantial research funding in their mid 30s. Nevertheless, these data might also reflect the historical funding picture for what we might think of as genuinely n00b PIs as being slightly less dismal than I have been assuming.
I enjoyed a few more of the graphs in the ppt file.
Slide 12: the demonstration of the R03 and R21 mechs being used to replace the R29 as the "starter award" for n00bs- despite the termination of the R29 specifically to try to dismantle the "starter award" trend.
Slide 17: Interesting to review the success rate changes for established investigators over the past three down-cycles (early 80s, late 80s-early 90s, present one starting about 2004). Also interesting to consider that those that were experienced as of the early 80s were in a mindset that about 38% success rate for applications kinda sucked. Puts their present-day attitudes about what is wrong with grant review into context, does it not?
Slide 28: Median number of years to complete a biomedical doctorate was 6 in 1980, increased to 7 by 1995 and now sits north of 6.5 as of 2005.
Slide 29: Documents the long sustained slide from 1970 to 2006 in the percentage of medical school faculty which are "new". The dramatic drop (~6.5% to ~3% new) from 2004 to 2006 makes you really wonder how this trend has played out in the past three years.
Slide 31: The average age of first Assistant Professor appointment for PhD's has gone from 32 in 1970 to 38.5 in 2006. Compare with Slide 28 for age of completion of doctorates. A four year differential in 1980 versus an eight year differential in 2005.
*An initial version of this post failed to make this distinction accurately