Do Republicans have a point in questioning NSF guidelines?

by Massimo Pigliucci

You’ve probably heard of the uproar that has recently been caused by a bill introduced by Texas Representative Lamar Smith, the chair of the US House of Representatives’ Science Committee. Lawrence Krauss has commented on it on behalf of the Center for Inquiry, and so has the Chronicle of Higher Education, the American Physical Society, The Scientist magazine, and Science Insider, among others.

At issue is draft legislation that would require the director of the National Science Foundation to declare that each grant approved for funding is:

i) “... in the interests of the United States to advance the national health, prosperity, or welfare, and to secure the national defense by promoting the progress of science;
ii) ... [of] the finest quality, is groundbreaking, and answers questions or solves problems that are of utmost importance to society at large; and
iii) ... not duplicative of other research projects being funded by the Foundation or other Federal science agencies."

Like almost every Republican move in a while (like, years, if not decades), it sounds superficially reasonable, until one digs just a little deeper to uncover the deeply ideological and highly disturbing motives behind the proposal. That Smith’s intentions are anything but benign was immediately betrayed by his request to Cora Marrett, the acting NSF director, asking for specific information about five particular grants awarded recently by the agency. According to Science Insider, Smith wanted comments of the copies of each reviewer, together with the notes of the NSF officers managing the process of peer review. Why? He said that “I have concerns regarding some grants approved by the Foundation and how closely they adhere to NSF’s ‘intellectual merit’ guideline ... the proposals about which I have requested further information do not seem to meet the high standards of most NSF funded projects.”

Needless to say, this is a highly unusual — perhaps unprecedented — move by a US Congressman, and one wonders exactly why Smith feels better qualified than the NSF officers and scientists who have actually reviewed the proposals in question (Science Insider does not identify the research topics with which the five grants are concerned, but one can make highly educated guesses...)? If you look up his bio, you will find that Smith graduated in 1965 from the the Episcopal School of Texas, in 1969 from Yale University, and in 1975 from Southern Methodist University’s Law School. That is, his technical background is as a lawyer, which makes him just as unqualified to be concerned about NSF’s standards of peer review as it makes me unqualified to represent anyone in a court of law.

It should therefore be clear that this is just another attempted Republican hatchet job to undermine the nation’s standards of reason and evidence in favor of their narrow minded ideological agendas. That said, let us take a closer look at the three points above, and see how unreasonable they are when considered on their own merits, as opposed to as thinly disguised trojan horses.

Quite frankly, all three standards seem pretty reasonable to me. The third one simply asks that NSF certifies that the money isn’t going to duplicate efforts. This is already standard NSF procedure, as Principal Investigators (i.e., the scientists submitting grant proposals) cannot submit the same proposal to different agencies except under very specific circumstances (e.g., if they are young investigators at their first attempt at funding), and even then they will have to withdraw their proposals from all agencies but one if they happen to be multiply funded. (A possibility that, given recent funding rates at federal agencies, especially NSF, has ridiculously low priors.)

The second one should be qualified because it contains the clause “of utmost importance to society at large.” As long as we agree that scientific knowledge is in itself important to society, we are good to go. After all, NSF’s very mission is not directed toward practical applications (for which we have NIH, the USDA, the DOE, etc.), but toward basic research. Still, not all basic research is actually worthwhile, or at the very least not worth the money, so setting standards is reasonable.

The first requirement, “to advance the national health, prosperity, or welfare, and to secure the national defense by promoting the progress of science” also seems to withstand scrutiny as long as one changes the “and” to another “or” (otherwise one could read the sentence as saying that all research has to specifically aid national defense, which would definitely be too restrictive). Again, this assumes the idea that a nation’s health, prosperity etc. also depends on the accumulation of basic scientific knowledge, an assumption without which the very existence of NSF wouldn’t make any sense.

So why shouldn’t NSF support such standards, discounting the specific and pernicious intentions of Representative Smith? Indeed, NSF — to some extent — already does address these issues. The process of peer review that examines each grant proposal is, at its ideal, designed precisely to guarantee that the funded science is worthwhile by the highest intellectual standards of the relevant community (i.e., the community of scientists in a given discipline). And NSF has for years required Principal Investigators to write a “statement of broader impact” in which they explain why and how their proposed research will benefit society at large. What, then, is the big deal?

The thing is: a lot of funded basic scientific research doesn’t really have a social impact, broad or narrow. Which is why most grant proposals’ statements of such impact are boiler plate stuff copied and pasted from one grant to another (trust me, I’ve done it), and usually can be summarized into a few rather unconvincing points: (1) the grant will fund some student (undergrad, grad or both), who will learn “valuable skills” of a sort; (2) well, it is well known that basic research often leads to unexpected and unplanned important applications “down the line”; and (3) the one usually informally given by scientists when pressed to explain how, exactly, spending hundreds of thousands of dollars and years of research time investigating the sexual habits of an obscure species of butterflies in the middle of Panama will help anything at all: “it’s intrinsically interesting, of course.”

Let’s start with (3) first: I’m not sure how to cash in on the concept of intrinsic intellectual interest, though I suspect there may be some more or less reasonable ways of doing so. But I can tell you that most of my colleagues (myself included, of course!) use that nice phrase as a synonym for whatever we happen to find interesting or challenging. There is, of course, nothing wrong in having idiosyncratic, even arbitrary, taste in intellectual matters — just like I cannot really fault you for preferring milk to dark chocolate, however strongly I feel you are actually wrong in your bizarre preference... But when one is asking a federal agency funded by taxpayers to shell out significant amounts of money one needs a bit more than just “I find it interesting” as a justification.

Which brings me to number (1) above: it is surely of some social value to train a few students in the ways of scientific research (though all too often for undergraduates this amounts to little more than pipette washing). But it also likely isn’t worth the hundreds of thousands or even millions of dollars (depending on the field) of your average NSF grant. Indeed, good small liberal arts colleges manage to give their students a nice taste of what it means to do research with comparatively small budgets, for the simple reason that learning how to do science doesn’t require being involved with cutting edge and/or expensive projects.

We are then left with (2): the contention that basic research is necessary because it leads eventually to applications of direct human interest. This is certainly true, as far as it goes, but you will notice that for a crowd that prides itself in abiding by the weight of evidence, most scientists are exceedingly recalcitrant to provide actual quantitative evidence of such connections. Yes, you do get a number of anecdotal stories, usually the same select few, recycled over and over. Or you get the argument that pretty much any complex example of applied science would simply not have been possible without some basic research preceding it.

While both the anecdotes and the generic basic applied argument are reasonable, they are a bit of a red herring, since they do not actually answer the fundamental question. The issue is not whether basic science can lead to human applications, but how often, and whether NSF’s funding strategy is the most efficient way to get us there. [1] As far as I know (and by all means, let’s crowdsource this), neither NSF as a federal agency, nor certainly individual Principal Investigators, have been making that argument in any detailed and verifiable way.

All of which does not, I hasten to say, justify the latest Republican assault on basic science. Nor does it mean that we shouldn’t have a National Science Foundation. But researchers themselves have for years called for a more rational approach to scientific funding and peer reviewing, which certainly has to include transparent ways of allocating money, and certainly more than just generic, boiler plate statements of “social impact.”

This is a complex conversation to have, and it is not one at which only scientists ought to be invited either. NSF, NIH, the USDA, the DOE etc. are all funded at taxpayers’ expense, and they are all under Congressional oversight — as they ought to be. It is therefore up to scientists to make a good case to the public, and yes to policy makers, that funding basic research is not a luxury, but rather a practical and cultural necessity for an open society. So far, they haven’t been doing a particularly good job, which I think only helps demagogues like Lamar Smith. We really ought to be able to do better than this.

———

[1] Interestingly, there is now a whole field of “translational research” that attempts to deal with this sort of question in a systematic way.