Bishop Hill Who would be in Professor Hardaker's shoes?
Dec 31, 2009 Climate: CRU Climate: other FOI Journals As the Climategate analysis starts to flow from Steve McIntyre's keyboard, it's interesting to note the theme of "climategatekeeping" emerging from the first few posts. It seems clear that there have been multiple instances of attempts to suppress or delay sceptic papers and just as many examples of warmist papers being rushed through to print on the nod. This angle to the climategate affair has been given added impetus in recent days by the extraordinary revelations of Spenser and Christy in their American Thinker article, showing how the journal editor at the International Journal of Climatology (IJoC) conspired with Hockey Team members to delay the appearance in print of a sceptic paper (Douglass et al).
IJoC, which is a journal of the Royal Meteorological Society of the UK, appears more than once in the gatekeeping stories. In 2004, McKitrick and Michaels submitted a paper to the journal looking at the connection between temperature and economic growth. This would have been threatening to the warmist cause because it suggested that some of the observed temperature increases might not be due to carbon dioxide. With depressing predictability, one of the reviewers was Phil Jones and as sure as night follows day, the paper was rejected.
But even this is not the end of it. In 2008, McIntyre became involved in the ongoing debate between Douglass and Santer. McIntyre's interest was piqued by Santer et al 2008, the latest attempt by the Hockey Team to refute Douglass's finding that the instrumental records and the climate models didn't match. Having found some interesting issues in the text of the paper, McIntyre requested Santer's model data in order to help him test the findings more fully. His request received a startlingly rude response:
I gather that your intent is to “audit” the findings of our recently-published paper in the International Journal of Climatology (IJoC). You are of course free to do so. I note that both the gridded model and observational datasets used in our IJoC paper are freely available to researchers. You should have no problem in accessing exactly the same model and observational datasets that we employed. You will need to do a little work in order to calculate synthetic Microwave Sounding Unit (MSU) temperatures from climate model atmospheric temperature information. This should not pose any difficulties for you. Algorithms for calculating synthetic MSU temperatures have been published by ourselves and others in the peer-reviewed literature. You will also need to calculate spatially-averaged temperature changes from the gridded model and observational data. Again, that should not be too taxing.
In summary, you have access to all the raw information that you require in order to determine whether the conclusions reached in our IJoC paper are sound or unsound. I see no reason why I should do your work for you, and provide you with derived quantities (zonal means, synthetic MSU temperatures, etc.) which you can easily compute yourself.
I am copying this email to all co-authors of the 2008 Santer et al. IJoC paper, as well as to Professor Glenn McGregor at IJoC.
I gather that you have appointed yourself as an independent arbiter of the appropriate use of statistical tools in climate research. Rather that “auditing” our paper, you should be directing your attention to the 2007 IJoC paper published by David Douglass et al., which contains an egregious statistical error.
Please do not communicate with me in the future.
Ben Santer
What Santer was saying was that the raw model output was freely available and that he thought McIntyre should convert this into the synthetic MSU numbers using the published algorithms. He was saying that the synthetic MSU data was therefore not raw data but an intermediate and that he shouldn't have to release it, a case that was extremely weak, to say the least.
McIntyre followed up with requests to NOAA, who had funded the research, as well as to Glenn McGregor, the editor of IJoC, asking about the journal's policy on data availability. Although McGregor indicated that he would find out from the publisher, it was over 2 months later before McIntyre had a response. It turned out that the journal did not require data to be made available as a condition of publication and to add insult to injury, McGregor now invited McIntyre to ask Santer for the data, in the full knowledge that Santer had already refused to release it.
The freedom of information request to NOAA, however, turned out to be effective, and at the end of January the data was made available, the bureaucrats claiming unconvincingly that it was due to to be published anyway.
At this point yours truly enters the story. At the start of 2009 I decided to take up the issue of a data policy with Professor Paul Hardaker, the Chief Executive of the Royal Meteorological Society. As the publisher of IJoC, Hardaker was the one man in a position to do something about the lack of a data policy. His reply was quick in coming and was very encouraging, indicating that he would take the issue up with the RMS publications committee in May.
Shortly after this, another Climate Audit reader, Geoff Smith appears to have taken the issue up with Phil Jones, a co-author of the Santer paper, encouraging him to adhere to the highest standards of data availability and also telling him about Hardaker's intentions. This seemed to be a worry to Jones, who set out his concerns in an email to Santer:
I'm not on an RMS committee at the moment, but I could try and contact Paul Hardaker if you think it might be useful. Possibly need to explain what is raw and what is intermediate.
Santer agreed that this was a sensible course of action and in March, Jones contacted Hardaker setting out the Hockey Team position on materials availability:
I had been meaning to email you about the RMS and IJC issue of data availability for numbers and data used in papers that appear in RMS journals. This results from the issue that arose with the paper by Ben Santer et al in IJC last year. Ben has made the data available that this complainant wanted. The issue is that this is intermediate data. The raw data that Ben had used to derive the intermediate data was all fully available. If you're going to consider asking authors to make some or all of the data available, then they had done already. The complainant didn't want to have to go to the trouble of doing all the work that Ben had done. I hope this is clear.
Santer seemed pleased with this approach, but held out the possibility of stronger actions:
If the RMS is going to require authors to make ALL data available - raw data PLUS results from all intermediate calculations - I will not submit any further papers to RMS journals.
This seems a remarkable position for a publicly-funded scientist to take but despite this, Jones seemed to agree. In his reply he also suggests that pressure has been put on the RMS over another journal article.
I'm having a dispute with the new editor of Weather [another Royal Met Soc journal]. I've complained about him to the RMS Chief Exec. If I don't get him to back down, I won't be sending any more papers to any RMS journals and I'll be resigning from the RMS.
In June I contacted Professor Hardaker once more to see what progress had been made on formulating a policy and once again received a cordial response. The committee, it seemed, had looked at some example policies Hardaker had shown them and had tasked him with drafting a formal policy for presentation at the next meeting in the autumn.
Professor Hardaker is now in a position that appears "interesting". He can see clear evidence that his editor at IJoC has been tipping the scales in favour of the Hockey Team - delaying sceptic papers and sending them to hostile reviewers, accelerating critiques of them. Even if the threats of Santer and Jones to refuse to publish in IJoC were never relayed to him, he will also now know that this is what the two Hockey Team members are thinking. He can have little doubt that if Jones and Santer boycott RMS journals then the rest of the cabal will follow too. Could this have commercial implications for the journals? Who knows? But meanwhile Hardaker can also be in little doubt that the scientific method requires that the RMS insists on openness in data and in code. Santer's argument that intermediates can be withheld seems to me to be without scientific merit.
Hardaker's position must be profoundly uncomfortable. The "editor problem" aside, there is no doubt what the correct decision regarding the materials policy is in scientific terms. This one acceptable course of action must also be absolutely clear to Hardaker, who is on the board of Sense About Science, a body which promotes "good science and evidence in public debates". It is hard to see how he could credibly permit the withholding of any data or code while holding this position.
I wonder what decisions were made at that meeting in the autumn? I'll write and see.
Reader Comments (35)
Excellent post. I await the outcome with great interest.
Bishop,
Sending papers to "hostile" reviewers doesn't in itself seem a bad idea. I doubt that I would have learned what I have from reading at the warmist sites. The warmists might consider this site hostile.
But the result should have been some well-focussed comments, possibly objections, not the prevention of publication.
Likely there are many more papers submitted for publication than get to press and someone has to decide which papers are worthy. If the only reviewers are hostile, then the editor is in someones pocket.
Until retired, I was a college/university professor for 37 years, and both published in refereed journals and acted as a referee. Santer, Jones et al clearly have subverted the refereeing process at a few journals, and have achieved at least partial control over the scientific debate on AGW. This is really much more important than the quality of their science. This subversion goes to the very heart of science and is profoundly corrupting of the scientific enterprise, even more so than outright fraud.
One result is that none of their own papers is in fact refereed. Also, the journals they control, like IJOC, cannot claim to be part of the refereed literature.
Moreover, they have contaminated the climatology literature and the work of honest scientists who innocently depended on it, They now must go back and try to figure to what extent their own work is in error. Whole careers might have been wasted, and climatology has been channeled into a dead end.
The scandal has already diminished the respect the broader public has for all science, and it will discourage talented graduate students from pursuing careers in climatology, juist when we most need them.
Very informative post. The American Thinker article was enlightening. The hole keeps getting deeper. When did Santer become King? ..."Please do not communicate with me in the future."...away with you!
It's the Mother of Science Battles: Open science vs Clandestine science.
The Team would be wise to read history. Even the Berlin Wall could not hold back the truth.
Less scientists than hoodlums, some people.
Santer, Jones et al clearly have subverted the refereeing process at a few journals, and have achieved at least partial control over the scientific debate on AGW. This is really much more important than the quality of their science.
Bob I strongly disagree. No one should rely on what any Journal publishes if the "materials and methods" which underlie the conclusions are not archived so as to be easily available to anyone interested. This is a requirement of the Scienific Method, period. Because if they are archived and available, then they can be reviewed by any number of functionally qualified people to see if they support the results, without anyone getting to decide who is qualified.
If they are archived, the publication's reviewers also will be much more likely to do a reasonable job almost no matter who they are, and the publication will also get a better idea about the capacities of the reviewers it chooses. The fact is that the real peer review starts only after a paper is published, as has just been proven again by Steve McIntyre in the case of Climate Science. I have been mystilfied since day one as to why this requirement to publically archive is not known as a given for doing Science. In the case of Medical Science, no one would put up with "science" the way Climate Science does it, especially the ground level practitioners and their patients.
Another way to bottom line it in the case of "Climate Science" is that if the materials and methods are not publically archived, there isn't any science to review, no results to either promote, defend or criticize. In this respect he ipcc and its elite Climate Science have simply not been doing Science. "Let the Big Lie Die!"
My apologies for speaking so strongly, but this facet of the way Climate Science does "science" has severely bugged me now for 9+ years. It's essentially the first thing I ran into when I decided to look at AGW. thinking I was going to be learning some new science: some version of the TAR conclusions were released, but where was the science which underlied and resulted in the conclusions? It's right there in front of everyone. Why did it seem it was only me who noticed it? Not that I really knew who else did, but I was wondering, "Why aren't people raising holy hell about this?"
It's almost funny to realize that if Climate Science had been done as it should have been done according to the Scientific Method, we wouldn't even be talking about all this mess. As I'm sure you all know, Anthony Watts, enc.., are the only people who have so far even checked the "thermometers", which you would have thought a real Scientist might want to do if s/he was going to study the Earth's temperature. For god's sake!
All submission to publications would have been either rejected or soon withdrawn. But why is this so difficult to see? We trusted Climate Scientists was our first mistake. I got cured of that pretty quick. But alot of people still can't believe it and a lot of others have, you know, other "priorities".
The more we learn about these 'gatekeepers' and the way they denied access to their data, the more i am aghast at why this has been allowed to go on and on and on.
The various editors of the specific journals must have had a vague inkling of the Scientific Method - and they must have been capable to see that this issue was handled differently in other disciplines.
I am even more stunned by the behaviour of Jones, Santer etc, threatening not to give any papers to the RMS to publish - Who did they think they were?
Writers of Hollywood blockbusters?
As zoologist, I find it beyond belief that data are withheld in this fashion - they are not the personal property of the investigators. They are not intellectual property even - not the data.
As for the arrogant, top-down brushing off evidenced in Santer's and Jones' e-mails - what a bunch of totally unprofessional, discourteous curs!
What an arrogant little prick Mann is! And I noticed the Uriah Heap of the AGW crowd, Barrie Harrop, is chiming in as support.
As far as I understand it, the scientific method requires independent verification of results. I can't see how that implies that materials should be archived in a publicly accessible place - however good an idea that may be - nor that it implies that anyone can demand the data and code used - however much we feel that publicly-funded scientists should respond to such requests.
Steve McIntyre - again, as far as I've understood him - doesn't claim to be replicating the science, only checking the statistical methods used by climatologists. No-one has failed to use the scientific method simply because they failed to make their data and code universally available; saying that they did is a mistake.
What many are forgetting is that this is not a scientific issue, it is a political issue. While as scientists we can gnash our teeth in anguish over what the AGW crowd has done, the real issue is why they are going it. Clearly it is a political agenda. At COP15, there were many references to the IMF issuing a Tobin Tax. Other references to a world wide organization to control the carbon emissions. Smacks of a One World Government, does it not?
I think we would be wiser to listen to what Lord Monckton has been saying and take a political stand as well.
I don't think Hardaker should worry too much, the Team's credibility has taken a real knock. There are plenty of other pro-agw researchers still wanting to publish.
Happy new year bishop.
Rich
"As far as I understand it, the scientific method requires independent verification of results. I can't see how that implies that materials should be archived in a publicly accessible place - however good an idea that may be - nor that it implies that anyone can demand the data and code used - however much we feel that publicly-funded scientists should respond to such requests."
Please explain how someone can get temperature data from 1937 independently. The problem here is that climate "science" is observational and almost all of data that is not reproducible in any way. If this were experimental science your comment would be valid, but then again, these dopes would have been dead and gone long ago since their science would have been shown to be wrong. Compare Climategate to "cold fusion" in 1989.
There is a furhter twist to the Michaels McKitrick story. Michaels and Mckitrick published a fuller version of their work :McKitrick, R.R. and P.J. Michaels (2007), Quantifying the influence of anthropogenic surface processes and inhomogeneities on gridded global climate data, J. Geophys. Res., 112. Gavin Schmidt published a critique of this paper in the IJOC. See Schmidt, Gavin (2009) Spurious correlation between recent warming and indices of local economic activity. International Journal of Climatology.
Now comes the interesting part. First and most important the Scmidt paper seems to contain two serious errors. The first is the apparent belief that spatial autocorrelation in the independent variable necessarily implies that the estimated significance of the coefficients of independent variables needs to be adjusted. This is wrong. What matters is whether there is spatial autocorrelation in the RESIDUALSs of the estimated equation (and if there is most economists would look first for misspecification rather than "correcting" for autocorrelation). There is no detectable autocorrelation in the McKitrick Michaels equation. The second error arises when Schmidt inserts GCM model output troposphere data instead of actual satellite data and shows that the socio-economic variables still are significant. He interprets this as evidence that the non-zero coefficients found by M&M must have arisen by chance. What he did not seem to realise was that his estimated coefficients were different in sign from those of M&M. So far from showing that the M&M results arose by chance they in fact strengthen the case that socio-economic variables have an independent positive effect on measured surface temperature even allowing for actual or modelled tropospheric trend data.
So much so technical. But the beauty of this is that, thanks to climategate we have the review that Phil Jones wrote for this paper. It can be found here under review scmidt.doc
http://www.climate-gate.org/cru/documents/
You will see that Phil Jones gave a very light touch review that missed these errors - indeed seemed to endorse the first as a valuable reference for future work.
And the devastating reply by Mckitrick and Nierenberg is with the IJOC at present. See
McKitrick, Ross R. and Nicolas Nierenberg (2009). Correlations between Surface Temperature Trends and Socioeconomic Activity: Toward a Causal Interpretation . Submitted to International Journal of Climatology.
which can be found at
http://sites.google.com/site/rossmckitrick/
It will be very interesting to see what happens. To me it seems that S09 is simply wrong. It's a real test for IJOC. If they don't publish the response it will reflect very badly on them.
Looks like somone's leaning on Richard North......
http://eureferendum.blogspot.com/2009/12/rather-preoccupied.html
Spread the word.
Rich,
As we have seen repeatedly with Mann, Rahmstorf, Steig, Briffa, Jones, et al, the stats and code are often pathetically bad. How is someone supposed to check how bad Mann's numbers are if they can't get access to his numbers? Mann's hockey stick is due to the use of inappropriate data and screwed-up, homemade statistics code that he wrote. Had he simply done his Principal Components Analysis with a regular, commercially prepared stats package, he would have struck out. If no one got to see his code, no one would be able to figure out what caused him to have screwed up so badly.
J Peden,
Thanks for the stuff you have published in the past. My take on the climate science mess is like yours. What kind of scientist fails to check his instrutments? The Watts' discoveries impugn the competence of the whole field. Not only those responsible for the maintenance of the sites, but any scientist who published a study relying on temperature data. It is incredibly irresponsible to publish warnings of impending doom without making at least a cursory check to insure that your data isn't corrupted.
The same thing applies to Mann's hockey stick. What a joke! Can anyone imagine another field of human endeavor that would accept a finding that completely overturned everything previously believed without a question? I think the most gobsmacking aspect of my inquiry into all this came when I read of Mann's initial response to McIntyre's first request for the data he had used. Steve expected some kind of due diligence package. Surely Mann could simply append the file in an e-mail reply. But there was no file! No one had ever asked for the data, including those who supposedly did the peer review.
The world is doomed, but no one ever bothered to check the math. Simply incredible.
"The world is doomed, but no one ever bothered to check the math."
Nicely put, Stan.
The posting refers to "Spenser and Christy in their American Thinker article". It's Douglass and Christy. This error appears a couple of places in the posting.
Bishop,
As a research scientist myself, I'm not sure if requiring code be made available is reasonable, or even prudent. But I don't see any reason why intermediate data should be unavailable, so I agree with you there. "Intermediate data," "final data," "raw data" ... these are all the data from which the results and conclusions are derived. Having this information can only improve reproducibility, and therefore strengthen the authors' claims (or allow these claims to be corrected).
But code? Well, two points: 1) often times I write code that will be used in multiple experiments of various types. I've generally spent a lot of time creating it, and would like to be able to explore outputs for various goals. To give that away and let others discover things I haven't had a chance to discover could undermine my productivity. If the relevant algorithms are described in sufficient detail, then others should be able to reproduce them, and check their data against my intermediate and final results.
2) I wouldn't want other researchers to become too dependent on using my code anyway. It's easy to just push a button and get the same results again and again. True reproducibility means that someone else has done the work to create the same experiment (i.e., their version of the code) and then produced the same results.
Imagine some sort of physical or chemical experiment. Giving away code would be like saying, "Here's my lab, my materials, and my equipment ... use this to get my results." Better, I think, to describe the materials and equipment used, and see if others can produce the same output I did.
Just my thoughts on the matter.
Cheers.
@Ted Carmichael
I strongly disagree. Code should be made available to the referees, and be put into the public domain as soon as the paper is accepted.
This is because code is as important as data in reaching the conclusion, and because the code was paid for by public money and thus belongs to the public.
If your code is so simple it can be easily reproduced, you do not have any competitive advantage by keeping it secret. If your code is complicated, you should be able to write the next paper before the competition fully understands your code. If you cannot take advantage of your head start with the complicated code, then you are not worthy and it is right that the competition writes the next paper.
Indeed. Claerbout's principle:
"Steve McIntyre - again, as far as I've understood him - doesn't claim to be replicating the science, only checking the statistical methods used by climatologists."
Rich, the statistical methods are their "science", and McIntyre has had a very hard time getting the code and exact data used, both as an official ipcc reviewer and as qualified statistician in his own right. Then look at what he found.
No one should have to do a completely seperate study to refute a study that can't stand up by virtue of it's own methods. It's the AGW hypotheses which have to stand up scientifically first. This should be obvious, but it isn't, partly because Climate Science has simply gotten away with it . Science is not Politics. The ipcc and its elite Climate Scientists are not doing Science. It's worse than anyone would want to imagine, but it's true.
Here is the replication policy of Econometrica, one of the premier journals in econometrics and mathematical economics
"Econometrica has the policy that all empirical, experimental and simulation results must be replicable. Therefore, authors of accepted papers must submit data sets, programs, and information on empirical analysis, experiments and simulations that are needed for replication and some limited sensitivity analysis. (Authors of experimental papers can consult the more detailed posted information regarding submission of Experimental papers.)
This material will be made available through the Econometrica supplementary material web-page. Submitting this material indicates that you license users to download, copy, and modify it; when doing so such users must acknowledge all authors as the original creators and Econometrica as the original publishers.
At the same time the editors understand that there may be some practical difficulties, such as in the case of proprietary datasets with limited access as well as public use data sets that require consent forms to be signed before use. In these cases detailed data description and the programs used to generate the estimation data sets must be provided, as well as information of the source of the data so that researchers who do obtain access may be able to replicate the results. This exemption is offered on the understanding that the authors made reasonable effort to obtain permission to make available the final data used in estimation, but were not granted permission. We also understand that in some particularly complicated cases programs may have value in themselves and the authors may not make them public. Similarly, there may be compelling reasons to restrict usage, and if we agree we will post a notice on the web site regarding such restrictions.
Requests for an exemption from providing the materials described here, or for restricting their usage, should be stated clearly when the paper is first submitted for review. It will be at the editors’ discretion whether the paper can then be reviewed. Exceptions will not be considered later in the review and publication process."
Seems pretty reasonable to me and several other journals in economics have simliar policies. This seems like best practice.
I think some are also missing the fact that publications such as "Nature" require material and methods to be archived so as to be reasonably available as a very condition for publication, then don't enforce it. The only reason McIntyre was able to show what Briffa was doing - the exact opposite of what his "statistical methods" should have been doing - was that the Royal Society enforced their own requirements for publication of another paper Briffa wanted published which used the same data he'd used first 10 years earlier.
If you believe in AGW and the concept of Global Mean Atmospheric Temperature, Briffa managed to "prove" that the GW "mean temperature" only strikes one tree on the Yamal Penninsula. It's a Cherry Tree.
"Nature" has just published an editorial condemning the [not happening] ~"harrassment" of the poor McIntyre-victimized elite Climate Scientist Team. McIntyre was similarly [not] harrassing "Nature"[!] because it did not enforce its own conditions for publication of Mann's and Santer's papers either, if I recall correctly - needs checking as to the exact papers in "Nature". McIntyre is talking about it at Climate Audit. I also saw some of it there back when McIntrye was in the process of trying to get "materials and method's.
"It's worse than we/you thought." You wouldn't believe it if you hadn't seen it. "Nature" is ruining itself too.
@Richard Tol
You said: This is because code is as important as data in reaching the conclusion, and because the code was paid for by public money and thus belongs to the public.
Well, actually, no ... that's not how research is done currently. (At least in the US. Not sure about other countries.) Government funding agencies do not get the rights to developments that they fund through research grants. Any derived benefit remains with the researcher, other than what they publish or lease ... it is their work, and their property. (Although usually the university shares in the patents, if developed at a university. Also, most grants require a statement of how the results will be disseminated.)
This is true not only for software, but also new inventions, novel algorithms, drugs, compounds, materials ... you name it.
The funding agencies provide capital that supports this research because it works pretty well as it is. They want to advance the state of the art. And they know that society as a whole benefits tremendously, even if only individuals are incentivized. You can make an argument to change this, but you should understand that you are advocating a fundamental and far-reaching shift in how things are currently done.
Paul from Boston, stan and JPeden,
My point was only the very narrow one that replication as a necessary part of the scientific method does not necessarily imply providing universal access to data, code, methods or anything else. This was the assertion by JPeden that I wished to challenge: Another way to bottom line it in the case of "Climate Science" is that if the materials and methods are not publically archived, there isn't any science to review, no results to either promote, defend or criticize. In this respect [t]he ipcc and its elite Climate Science have simply not been doing Science. "Let the Big Lie Die!"
Public archiving is a good way to facilitate replication but not a necessary condition for it. It is clearly - to me - possible for one research group to replicate, independently, the experiments of another without either releasing anything to third parties. The requirement of reproducibilty will have been met. We can argue that it could have been done better but not that it hasn't been done.
There are sticks a-plenty with which to beat the Team. I fear bad arguments drive out good ones.
@Ted Carmichael
You are talking about government subsidies for R&D, where indeed the subsidised party is paid by public money to acquire private (intellectual) property.
I am talking about government subsidies for fundamental research, that is, the sort of research that is not supposed to lead to marketable products.
@Richard Tol
I'm talking about research grants such as from the National Science Foundation. The focus on basic science doesn't mean that it can't lead to marketable products, only that it is not required to. But the NSF doesn't hold any resulting patents from the research it funds. Nor the NIH or DARPA, or any other government granting agency that I am aware of.
Many private research grants often work the same way, such as with the X-prize.
So some research can be proprietary or such that the "materials and methods" don't have to be public. So what? We are not talking about that kind when it comes to the Climate "Science" the ipcc and its elite Climate Scientists are doing, and that's all there is to it!
But even if it is, then the science is certainly not "settled", and they are admitting it in contradiction to their own [non scientific] "standard". The noble Climate Scientists should have been tryiing to disprove their own hypotheses just like real scientists do, but they weren't . For example, for a mere $10 billion I will assemble a bunch of "scientists" who will "prove" that GW will bring about the closest thing possible to Heaven on Earth, using the Climate Scientists' own way of doing "science". It gets totally rediculous to do science the way the ipcc, enc., does it, as they have proven!
@Ted
The NSF is funded by public money. The grants go to universities, which are either public institutions or private institutions set up for the common good. How can a university professor claim that her non-commercial research is private?
@Richard
Well, because she's the one who developed the whatever-it-is. It is the result of her labor and ingenuity.
It's kind of like being a patron of the arts. You may be supporting, say, public performances and creative acts, but that doesn't mean you own the rights to the art that is created. Those rights remain with the artist. Sometimes the artist puts on a show that, like Shakespeare, is already in the public domain. Sometimes the artist creates a commission for public display (in which case it is owned by the public). But sometimes the artist is supported in their work with no strings attached, and they keep the fruits of their labor. Society still benefits because new art is created, but society doesn't own it.
The university will often share in resulting patents resulting from research at that university. If they're smart about it, it will be a small percentage and they will streamline their 'to-market' practices. But the NSF doesn't require a piece of those patents. Only, perhaps, an acknowledgement that they funded the research behind it.
Please note that I'm not arguing that the system is perfect. There are details that may need adjusting. However, in the broad strokes it not only works, but it's worked for a long, long time.
@Ted
Your are right at a legal level. There is nothing in a NSF that forces one to disclose non-commercial research findings. I think that the NSF should rewrite its rules. Fortunately, journals are taking corrective action (see the Econometrica example above).
"in a NSF" > "in an NSF grant"