From LLNL scientists
https://www.science.org/content/article/nsf-grant-decisions-reflect-systemic-racism-study-argues
The analysis supports earlier studies finding similar racial disparities in the funding of scientists by other federal agencies, notably the National Institutes of Health (NIH). And its authors—a team led by geochemist Christine Yifeng Chen, a postdoc at Lawrence Livermore National Laboratory—attribute the gap in NSF funding rates, with white scientists at the top and Asian researchers at the bottom, to “systemic racism.”
The NSF funding disparities “have cascading impacts that perpetuate a cumulative advantage to White [principal investigators] across all of science, technology, engineering, and mathematics,” they write in their study, posted earlier this month on the Center for Open Science preprint site.
6 comments:
I read that a followup analysis showed that the gap can be explained by the fact that older faculty have much greater chance of getting grants than early career. The study is for the last 20 years where the fraction of Asian faculty went way up by a factor of 3, and whites where down by a factor 1.35. This is also also true of blacks and hispanics that that grew by 1.35 and have a much smaller gap compared to whites. All showing is the known gap between younger and older PIs.
Also premise of the study sounds weird, if Asians have such harder time getting grants why are we hiring an increasing fraction of them? It is clear that there is something off about the way the authors presented the data.
Some other factors are that for NSF funding the engineering-computer science sections have a much higher fraction grants from Asians than from other group but for NSF this category has the lowest acceptance of about 18% percent, while some fields like earth science which are mostly white have 36% acceptance rate. The reason for this is that engineers have many other funding sources while for earth science it is rather limited. In any case this also means there is gap that has nothing to do with racial bias.
In fact there could other variables like people born in non-native english speaking countries may have lower odds of grant success, but if this is true of any country not just Asian countries means it is not based on racial bias. I think the vast majority in the growth of Asians fraction are from abroad so this could be a big factor.
NIH has been doing these studies for years now and know about effects of topic
selection and that the older PIs get disproportionates amounts of money. The authors also go on about a 2011 NIH study showing bias in funding rates against blacks but several other followup studies have take into account all
the other factors like age, topics and so on found these gaps go away. Some of these NIH studies also considered gaps in funding with Asians but found if you only look at US born Asians the gap goes away.
What I do not get is why the authors do not even bother to propose other possible explanations. They see a gap therefore the gap must be racism. I am not sure what kind of scientists they are but yes you can measure a gap but their study cannot directly measure racism, only infer it is the explanation. I would have thought that a scientists would know that whenever you have to infer on unobserved cause you have to consider what are known as confounding variables. In fact without a direct measure they can never prove racism in this kind of study.
But hey do not let the scientific method get in the way of a good narrative. I hope LLNL can do better than this kind of shoddy work.
From the Science article
"finding similar racial disparities in the funding of scientists by other federal agencies, notably the National Institutes of Health (NIH)."
This is talking about a 2011 paper
" D. K. Ginther, W. T. Schaffer, J. Schnell, B. Masimore, F. Liu, L. L. Haak, R. Kington, Race, Ethnicity, and NIH Research Awards. Science. 333, 1015–1019 (2011)"
However the same group did a followup study
"D. K. Ginther, J. Basner, U. Jensen, J. Schnell, R. Kington, W. T. Schaffer, Publications as predictors of racial and ethnic differences in NIH research awards. PLoS ONE. 13, e0205929 (2018)
Showing that a large fraction of the gap could be explained.
Also there is another study showing additional effects.
"T. A. Hoppe, A. Litovitz, K. A. Willis, R. A. Meseroll, M. J. Perkins, B. I. Hutchins, A. F. Davis, M. S. Lauer, H. A. Valantine, J. M. Anderson, G. M. Santangelo, Topic choice contributes to the lower rate of NIH awards to African-American/black scientists. Sci. Adv.5, eaaw7238 (2019)"
It could be the same effects could be arising in the NSF study and the authors
need to make this clear from the start. I checked the LLNL preprint to see if they cited these two other papers and they did. From their p reprint
"However, several studies over the past decade have shown inequalities in the allocation of research funding, most notably at the NIH: a 2011 study showed that Black PIs were funded at roughly half the rate as White PIs (2). Subsequent analyses revealed additional inequalities across race (3–8),"
These references 3 and 6 of the preprint. The authors are completely misrepresenting these papers which are actually showing the opposite of their claims. It could be they did not read the references which is just really bad science, or they did this deliberately which is a form of scientific misconduct.
The more you look at this paper the more and more odd things. I think something is odd about how they are plotting things as well, maybe someone can figure it out.
This thing is a repeat of the 2011 NIH debacle.
The above cites
"D. K. Ginther, W. T. Schaffer, J. Schnell, B. Masimore, F. Liu, L. L. Haak, R. Kington, Race, Ethnicity, and NIH Research Awards. Science. 333, 1015–1019 (2011). "
However this work has now been largely discredited see below.
If you look at the followup paper
D. K. Ginther, J. Basner, U. Jensen, J. Schnell, R. Kington, W. T. Schaffer, Publications as predictors of racial and ethnic differences in NIH research awards. PLoS ONE. 13, e0205929 (2018)
The authors discussed how the initial 2011 suggesting racial bias
results caused a big stir in NIH so the formed a panal to look into it which pointed
out that many other factors could explain the different funding rates.
Below is what they say.
"In response, the NIH Director established a high-level Working Group on Diversity in the Biomedical Research Workforce (WGDBRW), and their report [9] pointed out that potentially important explanatory variables were missing from the previous analysis [2]. The report argued that the ability to
distinguish between the competing explanations of the black/white NIH funding gap—application merit, investigator characteristics, or bias in the peer review process—was insufficiently explained by variables included in the analysis, prompting a the need for a more detailed evaluation."
In light of the previous works on funding gaps and how it turned out you would think the authors would mention this or think maybe they should consider this.
Instead of doing real work they add little jems like below in their supplementary section. How on earth is this in a science paper?
"S6.3.1 On the “ascendancy of whiteness” We note that “whiteness,” which includes white identity and the racialization as white or assumed whiteness of marginalized groups or individuals, is a nebulous term that can shift to include or exclude specific groups depending on social context. The condition of being white—of enjoying the privileges of whiteness—is often made to seem neutral, inviting, and inclusive of racial, sexual, and other minoritized identities (5, 22); in other words, whiteness is (explicitly or implicitly, through choice, coercion, or assimilation) often the goal of expanding equality in outcomes (rather than opportunities) (3) to historically excluded groups. The “ascendancy of whiteness,” a phrase derived from Rey Chow’s The Protestant Ethnic and the Spirit of Capitalism (40), is used to describe how white hegemony is consolidated not through the exclusion of the “ethnic,” but rather through liberal multicultural inclusion and its violence on and separation of ethnic communities (41). This “logic of assimilation, the default to whiteness” (3), whether conscious or not, is an inherent part of race and racialization in the United States. While the authors have neither the space nor expertise to fully discuss the history and implications of whiteness and its various manifestations within the scientific disciplines, the data we present demonstrate the privileges, power, and material gain that whiteness confers"
LLNL assuring things designed to wipe everything off the face of the earth still work
and do so in a way that that embraces fairness, equity, and anti-racism.
Evey study of anything that alleges "systemic racism" fails to take into account racial demographics in the population being studied. Black incarceration rates fail to take into account Black criminal offense rates. Here, Black grant award rates fail to take into account Black Ph.D. rates. Choose a preferred conclusion and then twist the data to fit.
8/10/2022 5:27 PM
In studies of these kind if they find some kind of gap and they simply assume bias. You cannot measure bias directly you must infer and inferring a cause that cannot be measured is always problem in the social sciences where there are tons of additional variables. In physics and chemistry labs you can usually tie down most of the variables, in the social science not only cannot not control them but it is very likely there could additional variables you have never identified. It puts the burden on you to show that other variables are not coming in to play.
There is example after example of social science studies showing a gap. A followup study which takes into account more variables sees the effect reduced, another study sees it is reduced more and so on. The problem is if the original work is all hyped out and makes it to the media and it becomes the accepted as truth and all the followup studies are forgotten about. Also they can use twitter asymmetry. They tweet out the great new paper, shortly afterwords if flaws are found, you can bet your bottom dollar they will never tweet out that they found out there are issues, their effect reduced, and the data did not really support their claims.
I noticed in the NSF study the authors never mention other possible effects.
I have doubts about just how honest this stuff really is, the authors have a lot of data and put a lot of work into yet they misrepresent other studies. They also claim that if other factors are relevant they do not have the data to consider this, yet some very simple estimates can easily be done to show that other variables besides bias must contribute to the gaps they see. It makes me wonder if they actually did some of these because they are so easy to do but found results going against their narrative so they just left it out. I guess if you get lots of tweets saying "this work is so amazing", "thank you so much for this needed study", " this just scientifically confirms what we already knew in your hearts", "heart emoji" it is worth it.
Also why do people so addicted to complement tweets?
Post a Comment