Women in the R00 phase don't apply for R01s as frequently as men

Sally Rockey:

A specific issue that recently has recently created interesting conversations in the blogosphere is whether female K99/R00 awardees were less likely to receive a subsequent R01 award compared to male K99/R00 awardees. We at NIH have also found this particular outcome among K99/R00 PIs and have noted that those differences again stem from differential rates of application. Of the 2007 cohort of K99 PIs, 86 percent of the men had applied for R01s by 2013, but only 69 percent of the women had applied.

She's referring here to a post over at DataHound ("K99-R00 Evaluation: A Striking Gender Disparity") which observed:

Of the 201 men with R00 awards, 114 (57%) have gone on to receive at least 1 R01 award to date. In contrast, of the 127 women with R00 awards, only 53 (42%) have received an R01 award. This difference is jarring and is statistically significant (P value=0.009).
...
To investigate this further, I looked at the two cohorts separately. For the FY2007 cohort, 70 of the 108 men (65%) with R00 awards have received R01 grants whereas only 31 of the 62 women (50%) have (P value = 0.07). For the FY2008 cohort, 44 of the 93 men (47%) with R00 awards have received R01s whereas only 22 of the 65 women (34%) have (P value = 0.10). The lack of statistical significance is due to the smaller sample sizes for the cohorts separately rather than any difference in the trends for the separate cohorts, which are quite similar.

And Rockey isn't even giving us the data on the vigor with which a R00 holder is seeking R01 funding. That may or may not make the explanation even stronger.

Seems to me that any mid or senior level investigators who have new R00-holding female assistant professors in their department might want to make a special effort to encourage them to submit R01 apps early and often.

13 responses so far

  • eeke says:

    One thing that Datahound distinguished that Sally Rockey does not is the fraction of K99 awardees who activate the R00. She says: "Of the 2007 cohort of K99 PIs, 86 percent of the men had applied for R01s by 2013, but only 69 percent of the women had applied." Of course it's lower, because the number of women (as Datahound pointed out) who (presumably) go on to faculty jobs is lower. Rockey omits this key point. I am not convinced that it is a matter of "vigor". Even though Rockey acknowledges that there is bias, I don't like how she seems to brush this off as though the women aren't trying hard enough. And you too, DM.

  • drugmonkey says:

    Data hound reported a non-significant diff of a few percentage pt diff in R00 transition. So....? What am I missing?

  • anonymous postdoc says:

    This post, and the Ginthner report, both point to reduced number of submissions, and resubmissions, in two less-privileged populations of NIH investigators. The thing is, these populations of investigators are hardly monolithic. I wonder if number of grants submitted may be a symptom of a different problem.

    Thinking of my own experience with labs I know which submit a higher number of proposals, they tend to have high numbers of personnel. At worst, the proposals are all ghost-authored by trainees. At best, there are a large number of trainees working on various threads of research to support a number of different proposal types and topics. (See also: the very next post on this blog, which discusses whether it is worth the resources to put a given trainee on a small replication project.)

    A trainee would have to be crazy, desperate, or checked out from reality to join a lab that is not headed by a person with the highest success trajectory she can find, implicit racism or sexism be damned. Where the trainees go, the preliminary data follows proportionately.

    I have been thinking of this in particular after this post, as I have a friend who is a female R00 holder who is delaying submitting R01s because of a perceived lack of preliminary data. She is having a hard time attracting trainees and is making do with the hands she can get. This is hardly a referendum on her research, which is excellent and innovative (recall the R00). A different friend, a male R00 holder, is starting his laboratory with two postdocs.

    I don't think it is enough to simply encourage people to submit more grants, without thinking about why this might not already be happening.

  • DrugMonkey says:

    Agree. Local mentors would, presumably, do more than just say "submit more grants". They should be looking to help with the hows and the whats.

  • PhysicsLurker says:

    I've been following these discussions for a while now, and this recent data point seems quite troubling. If you account for the percent of K99ers who go onto the R00 phase (as eeke suggested), and using datahound's numbers, it doesn't change the picture that much. In that case, 93% of men that went onto the R00 phase submitted an R01 by 2013, whereas 78% of women had.

    I find the 2007 K99ers to be a really interesting group, because this is a cohort that won a fairly competitive award early on which should imply: (1) That they have at least an above-average grant writing ability (2) That they are familiar with the NIH system (3) That they have a good grad-student/post-doc background (4) That they are working on a topic that is of interest to the NIH, and (5) that they've already had a very positive experience with grant writing. Taken all of these together, I really don't understand why anyone in this group would need a mentor to push them to submit an R01. Sure mentors can help with this, but shouldn't just about everyone in this group realize the importance of R01s, having already been through the ringer? Maybe I think too highly of the K99ers because its such a competitive grant, but this group is already at the top of the game coming out of the postdoc phase.

    If we assume that most of the 2007 K99ers started faculty careers in 2009, then this would imply that in the first 4 years of their independent career 7% of these men and 22% of these women haven't submitted an R01. My question is, are these people still in academia? Of course fields differ, but I find it hard to see someone being on a successful tenure pathway who hasn't even submitted an R01 in four years. While having an R01 isn't a requirement for tenure these days, not applying for grants is poorly looked upon. As drugmonkey pointed out, you have to buy the lottery ticket to have a chance to win.

    Assuming the same percentage of male and female 2007 K99ers are in academia in 2013, I can only come up with two other explanations. The first is the confidence gap, and that the men in this group are more likely to put their work out there to get shot down by the review panels. The second is that the women have taken time more time off for family, and hence have less preliminary data for that first R01. When coupled with a confidence gap that would make for a bad combination. While I can't comment on the second part of this, it would seem to me that getting a K99 should help with the confidence gap enough to encourage women to at least submit an R01.

    The other question I have is whether the people that aren't submitting R01s are applying for R21s or other funding mechanisms. That would of course require more analysis.

  • datahound says:

    From my analysis, 92% of men and 89% of women with K99 awards transitions to R00 awards. Sally Rockey indicated that 86% of men from the K99 cohort had submitted an R01 application. As noted by PhysicsLurker, this means that 86%/.92 = 93% of men with an R00 had submitted an R01 application. In contrast, 69% /.89 = 78% of women with an R00 had submitted. From my analysis, 57% of men with an R00 had achieved R01 funding. This indicates a funding rate of 57%/93% = 0.61. 42% of women with an R00 had achieved R01 funding and 42%/78% = 0.54. Thus, it appears that the funding rate for women is lower that it is for men even correcting for the rate of submitting at least one R01 application. Note that this does not correspond to the per application success rate since no data have been presented about the number of applications submitted per submitting applicant.

  • PhysicsLurker says:

    Datahound, is the difference of .61 for the men and .54 for the women statistically significant given the sample sizes? Any idea on the p-value?

  • anonymous postdoc says:

    Regarding PhysicsLurker's points, the reason for my post above is that both the "confidence gap" and the "time off for family" explanations assume equal resources on the part of male and female R00-ers. That assumption is not necessarily valid.

    "Time off for family" does not seem to be a reasonable explanation for the lack of submissions in minority applicants as demonstrated in the Ginther report, and it is not parsimonious to assume multiple causes for failure in different groups when one will do.

    Turning to the "confidence gap", it is usually phrased in such a way as to imply that if only we could get [disadvantaged investigator] to believe in themselves, they could perform just as well as white males. I agree with PhysicsLurker that this cohort of R00-winners are at the top of their game, and so it doesn't make much sense that they are self-sabotaging because suddenly they don't feel good enough.

    Given Datahound's demonstration that these women are still less successful than men in obtaining grants, it would suggest instead that they either laboring under unequal standards of review, unequal distribution of resources, or both. There is some data to support both in the case of women. These mechanisms would be sufficient to explain lack of success in minority applicants as well.

    The problem with this explanation is that the causes of these issues are systemic and not within the control of the applicant. However, I look askance at explanations that would suggest that these applicants are mostly just getting in their own way, as it makes it easy to assume that just getting newer, better, more confident applicants would fix everything.

  • drugmonkey says:

    Thus, it appears that the funding rate for women is lower that it is for men even correcting for the rate of submitting at least one R01 application.

    Multiple additive factors or, as I prefer, "the death of a thousand cuts". Confidence gap, family timelines, preferred lab size, local resource disparity.... it all goes into the mix.

    This is why it is hard to fix an observed population disparity if one insists on finding "The Problem", as Rockey and co seem intent on when it comes to the disparity identified in Ginther.

  • datahound says:

    Physics Lurker: From the sample size, I would estimate a p value of 0.2. Thus, it is not statistically significant. Nonetheless, I would anticipate that analysis of more cohorts (and thus a larger sample size) would allow this to rise to significance (if NIH could provide the data).

    drugmonkey: I agree completely. It is very unlikely that there is "a problem", but rather the death by a thousand cuts. I am working on other analyses that might help define more precisely some of these factors.

  • girlparts says:

    There is a very hard-to-quantify sense I have that, while young women are easily seen as "promising" , and may be well regarded as trainees, that there is a very big difference in how readily they are seen as having established a program and a lab. This came out in the previous drugmonkey post about female PIs being frequently mistaken for trainees. Personally, I submitted plenty of R01 applications during my R00, but I *felt*, accurately or not, a higher bar for how many independent pubs I might need to have before doing so. I only jumped up to an acceptable investigator score after a year with 5 pubs. And one reviewer still questioned my independence.

    Another factor I would propose: are there data on how many of those R01s are sole-authored vs. multi-PI? Partnering with a more established investigator to get that first R01 seems like a good strategy these days. I was discouraged from doing so by other women, because I would be perceived as not being independent from a senior male PI. I'm not sure that male investigators feel this problem quite so acutely.

  • drugmonkey says:

    It is always good advice to pursue an unassailably independent core to your research peogram. Top priority.

  • qaz says:

    Very interesting anecdata - on the study section I'm on, the last five cases where someone specifically complained/brought-up-for-discussion the question of independence, the candidate was always female (and the empire she was in generally male). Truth be told, all five cases were valid questions - some of them were settled on the she's independent side and some not, but all of them were cases where the woman was in the same university (and usually the same labspace) as a more senior investigator who had been an advisor or postdoctoral mentor.

    The question that I find interesting here is whether the reason none of these cases were men is because men don't stay in the same lab/empire or because men don't get questioned about it. Needless to say, I'm going to be watching for this dichotomy as I review grants in the future....

    I'm certainly not saying this is the only reason. I'm just very curious whether this is one of those thousand cuts.

Leave a Reply