I know complaining about Republican chutzpah is like complaining about the sun being bright, but still, this is flabbergasting. Particularly nauseating is the claim that investigating an intimidatory fishing expedition by the majority party in government is somehow preventing people from holding government to account.
Does this mean Scott LeM can FOIA La A/l/th0use's email?
You mean Kitty LeMew? God, that made me laugh.
Me neither. Does this have to do with an amorous cartoon skunk?
Particularly nauseating is the claim that investigating an intimidatory fishing expedition by the majority party in government is somehow preventing people from holding government to account.
"I have never seen such a concerted effort to intimidate someone from lawfully seeking information about their government." That's right, GOP spokesman! Writing a blog post is the most amazing violation of civilized norms in your lifetime! The president of the American Historical Association is the second coming of Richard Speck!
Every day it becomes a little bit clearer that the entire Republican Party, from the top to the bottom, is nothing but gaping assholes.
That's a little bit harsh towards gaping assholes. What did they do to deserve the comparison?
It's interesting, this does seem like a law whose balance should be shifted more in the direction of academic freedom. FERPA probably would take precedence where students' names and personal information are involved, and the helpful state statute database links to this court decision saying emails aren't automatically subject to request just because they're stored on government computers, they need to be about government business - so Cronon can probably safely withhold a lot of his emails from them.
But this doesn't prevent anyone from digging through emails to see if, for example, a given professor has been improperly using government systems to engage in political activities, because that is theoretically an issue of public concern, and that kind of request could still have a chilling effect.
9: The whole trickle-down thing.
Not that it matters to the ethics of it, but the latest TPM story says that the emails were on Cronon's personal account. They explicitly don't source or evidence that, though presumably it came from Cronon. So legally speaking they may not even be subject to the Open Records Request, or indeed to the university's policies. But then again that TPM report contradicts reporting by others that also seems to be sourced from Cronon.
emails aren't automatically subject to request just because they're stored on government computers, they need to be about government business - so Cronon can probably safely withhold a lot of his emails from them.
There may also be privileges that apply. I work with state agencies, not universities, and so the concepts don't carry over perfectly, but there's a privilege in NYS, similar to a federal privilege, for intra-agency communications. The agency doesn't have to spill internal deliberative documents that don't make up part of a final action that will affect the public. There's almost certainly an analogous privilege in Wisconsin, but I'm having trouble wrapping my head around applying it to a professor's emails -- none of what he does is what I'd think of as a final administrative action.
12: The request email is quoted in the 2nd OP link: "Copies of all emails into and out of Prof. William Cronon's state email account..."
13: Unless it was dealing with university administrative/executive decisions, I can't imagine he would be at all protected by that.
Well, there are no "the emails", right? The request isn't for particular emails answering a given description, it's all emails answering a given description. So while there might be some really juicy emails that the GOP hopes to get its hands on (something I rather doubt), Cronon can't just say "the email you want is elsewhere". If he's exchanged email with a student named Alec, or mentioned AFSCME in professional correspondence, using his University email address (and such correspondence would be carried out with that address), that falls under the request's remit.
12: If they're seeking emails from his personal account, while I don't know Wisconsin law so there's a bare possibility there's some quirk there, they can almost certainly go pound sand. Your entire life isn't subject to FOIL because you have a government job.
ActBlue has a way for out of state people to give to the petition drives that are prerequisite to recalling the Wisconsin 8. I gave.
Seeing the R efforts blow up in WI seems like the best way to set limits for US house Rs.
FERPA probably would take precedence where students' names and personal information are involved,
Someone on Crooked Timber pointed out that even if FERPA-protected emails aren't produced to the GOP, someone still needs to look at them to determine that they are protected by FERPA, so there's still some loss of confidentiality.
Of course, in all likelihood, the person who does that will take two seconds to glance at the email, say, "oh, to/from a student", flag it as such, and move on.
Sort of related: I got an email a week ago, from the president of my university, "reminding" us all that while we have every right to communicate with government officials and members of the U.S. Congress, we are not allowed to do so using university letterhead or university email addresses.
This wasn't really in response to any event that I know about, and generally seems totally weird. Also we are in DC so none of us even have any representatives anyway.
Seeing the R efforts blow up in WI seems like the best way to set limits for US house Rs.
Blow up? They passed their bills. Moreover, while everyone was gaping at Wisconsin, they've been up to no good in other states as well.
Unless it was dealing with university administrative/executive decisions, I can't imagine he would be at all protected by that.
If it worked at all, it would probably only protect emails within his university or with other state workers. But while I'm not sure it would (on two levels: I don't know WI law, and I don't know how this would work out even in NY), it might. The public policy behind the privilege is literally to avoid this kind of chilling effect -- the public has a right to documentation of everything state actors actually do that affects the public, but not to every thought or communication that doesn't culminate in action, because if state employees had to be on stage all the time, they'd be too cautious to get anything done.
It doesn't apply perfectly, but it's close enough that if I were his lawyer, I'd be trying to figure out if the issue were worth briefing.
Well, I forgot about the restraining order in WI.
Legislative success or failure matters only if it produces votes in the next election. Passing bills that fund your opponents and not you is a failure.
Actually, it looks like I completely misread the TPM story. They're talking about Cronon's correspondence with TPM. D'oh.
23: Counting chickens.
21: I don't know if there's such an exemption in practice, but I don't think there's one written, and the Wisconsin law is a lot more sweeping than the FOIA, at least going by the preamble:
Declaration of policy. In recognition of the fact that a representative government is dependent upon an informed electorate, it is declared to be the public policy of this state that all persons are entitled to the greatest possible information regarding the affairs of government and the official acts of those officers and employees who represent them. Further, providing persons with such information is declared to be an essential function of a representative government and an integral part of the routine duties of officers and employees whose responsibility it is to provide such information. To that end, ss. 19.32 to 19.37 shall be construed in every instance with a presumption of complete public access, consistent with the conduct of governmental business. The denial of public access generally is contrary to the public interest, and only in an exceptional case may access be denied.
Sometimes I wish we weren't ruled by thwarted, crabbed little insecure spirits like this, but that's a little like wishing to be ruled by Son of Univac instead.
I don't know how ND and WI laws compare, but as a ND state employee I am pretty sure any E-mail that I send using the government supplied system that isn't covered by things like FERPA or HIPPA can be requested by pretty much anyone in the state. There are very very few exclusions.
Counting Chickens
Maybe. But along with the headlines, which themselves have an effect, this is unusual for WI:
https://secure.actblue.com/page/wiscrecall
There's another actblue page representing 30% more. Gingrich's 1994 shutdown backfired as well.
I don't know how ND and WI laws compare, but as a ND state employee I am pretty sure any E-mail that I send using the government supplied system that isn't covered by things like FERPA or HIPPA can be requested by pretty much anyone in the state. There are very very few exclusions.
Same here. I don't use my work email for anything but work. I don't even access my personal email account on work computers.
I have a feeling we're going to see a lot more of this (i.e., of right-wing harassment of academics who either work at state universities or have federally-funded grants). All the publicity that FOIA requests got in relation to Climategate, and Cuccinelli's attempts to subpoena records of research done at the University of Virginia, probably opened the floodgates on this one.
29: But at some level it does not matter--this potentially ushers in a whole "crowdsourced" right-wing industry following the climategate and O'Keefe edited video model. Even for folks who have been discreet, are they prepared to defend every line of every e-mail that might be taken out of context by some right-winger on a fishing expedition? Chilling effect on discussion? This is the icy breath of tyranny. Bring your own PC and use private e-mail accounts for everything possible.
Or back to the phone. Which I'd hate, personally, but that's the way it goes.
31: they're welcome to the e-mail chains documenting the latest bike parts I've bought. Tires this week!
The thing that's really weird about Climategate and other right-wing climate-related activity is that a lot of people seem to think that they are entitled not only to the final, published results of publicly-funded research, but to every intermediate stage of the work product. There are people demanding to see the data, the code that was used to process the data, even to have the code explained to them. You know what? When I write a paper, the code I use to produce the plots is a fucking mess. I can barely explain it to my collaborators, much less any random person who decided to subject me to an inquisition. It isn't pretty, it isn't documented, and often I do a lot of fiddling with things by hand in such a way that the files I have on my computer at the end of a project don't fully reflect all the steps that were involved in producing the published result. Maybe I'm just an atypical slob, but if people really think the norm should be having every step of a process fully documented for public consumption, that's asking a huge time commitment from people. I realize there are some scientists who think this is how things should be, but I think it's nuts. The point of replicable results is that other experts should be able to understand what you did well enough to do the same thing independently, not that any random person can find a detailed explanation of every single step and repeat it identically.
There are people demanding to see the data, the code that was used to process the data, even to have the code explained to them.
The last thing is ludicrous, and the middle thing is problematic, clearly, but more public data has always struck me as a positive, especially if it's data that's expensive to collect.
the code I use to produce the plots is a fucking mess
That isn't even the main problem. These dumbasses generally couldn't (or disingenuously refuse to) understand the science even if your code was immaculate. Look at the giant sweaty fapsession they had when they saw one researcher use the word "trick" in reference to a data normalization method, when even a nitwit like me understood perfectly well what he meant by it.
36: And most of them "understood" it too, they just chose to misrepresent it.
35: How public data can be is really dependent on where it comes from. Data you can put in a simple table is one thing, but this is almost never what you have. For almost any reasonably big experiment, you can't ever really release data without also releasing some set of software tools and calibration information to go with it, because the data doesn't mean anything on its own. In some cases this is a totally reasonable thing to do (astrophysics is one field where all or nearly all of the data seems to be public). In particle physics, the argument we get from experimentalists is that their data is huge and complex, so even serving it up to the outside world in a reasonable way would be a big investment. Plus, the code is huge and hard to understand and use if you don't learn it as part of the collaboration with a social network you can go to for frequent help. Also, the process of understanding the data is complicated enough that any outsider who tries to do it is almost certainly going to find things that aren't really there and waste everyone's time. Then there's the argument that processing the data will be an ongoing effort for years, and the people who built the machines deserve the first crack at figuring out. If they did somehow release it publicly in a useful form, they could have their Nobel Prizes stolen by some less careful person who manages to find the interesting thing faster. There are still arguments over whether the data should be released toward or at the end of the lifetime of a huge experimental collaboration, but I've heard from people who have tried to dig into such data years after the experiment ended that even compiling the code on modern machines can be a huge challenge at that point.
Plus, the code is huge and hard to understand and use if you don't learn it as part of the collaboration with a social network you can go to for frequent help.
I don't think the people who want, or pretend to want, access to the data have a conception of science or scientific research that can acknowledge the importance of being part of particular social networks for understanding.
As for 35's "the last thing is ludicrous," I think the problem is that if you've gotten to the point of releasing the data and the code, there will be people who think they've found a crucial error (or who might, in fact, have found a bug, probably one that doesn't affect the results in a major way if at all). This is how the right-wing misinformation machine seems to work on the climate front, at least, with all the continuing attacks on the "hockey stick". Researchers are free to ignore it all, of course, but it's not clear that ignoring it is a viable thing to do once reporters start taking these people seriously.
So I tend to fall back on thinking that even releasing code is a bad idea, since competent people who want to check what you've done really should be writing their own code anyway, and incompetent people have a higher barrier to making a nuisance of themselves if they can't get their hands on such things.
In particle physics, the argument we get from experimentalists is that their data is huge and complex, so even serving it up to the outside world in a reasonable way would be a big investment.
That seems reasonable, and this obviously seems like the kind of thing that people would be more comfortable with if it was a funded effort. On the other hand, this:
the process of understanding the data is complicated enough that any outsider who tries to do it is almost certainly going to find things that aren't really there and waste everyone's time.
doesn't strike me as particularly sympathetic, especially on projects that are so huge that replication would be basically impossible. Okay, other people might come up with fallacious interpretations of the data, but they might also come up with better or novel interpretations; that's the point.
I've heard from people who have tried to dig into such data years after the experiment ended that even compiling the code on modern machines can be a huge challenge at that point
That doesn't seem like a reason not to release the data.
40.1: well, okay, so the problem is right-wing hacks politicizing science and shitty science reporting. I understand the need to balance that against the potential benefit somebody else might get from the data.
Is there a tradition of informal within-field data sharing? That'd probably get a good chunk of the benefit. Still, data mining on large, hetergeneous pools of experimental data is a powerful technique, and it seems like a shame to handicap it in the name of James O'Keefe-types.
So I tend to fall back on thinking that even releasing code is a bad idea, since competent people who want to check what you've done really should be writing their own code anyway, and incompetent people have a higher barrier to making a nuisance of themselves if they can't get their hands on such things.
That seems basically reasonable, although a large, funded effort to keep old data accessible by developing open source tools for working with it also seems like it would be good (I realize it would never happen).
I have mixed feelings on the merits of (eventually) releasing code from particle physics experiments. Being a theorist, and not really wanting to join an experimental collaboration, I often find myself wishing I could dig into their data for myself. And I don't find all of these arguments fully convincing. I'm just saying that it's a big, complicated, messy issue, and I'm afraid that it would be a disaster if all these right-wing fishing expeditions turned into Congresspeople deciding it would be a good idea to start passing legislation demanding more openness in scientific research.
Having worked in a political office, I have been on the receiving end of an open records fishing expedition -- there's even some court filing on the internet somewhere with a lame attempt to twist something I wrote in an email into evidence of nefarious coordination with homelessness activists.
The fortunate thing is that it was not nearly the least temperate thing I had expressed in my email about certain of the people opposed to the permanent supportive housing facility. But "Nazi harpy", entertaining as it was, never made it into the record.
From what I understand of certain province's FOI/privacy laws, the most commonly applied exceptions have to do with personal information. But it depends on how personal information is defined and how you determine what constitutes personal information. However, the exceptions apply at the line-by-line level, not at the document level. So you'd get a lot of documents with lots of personal information redacted, even documents where only a few words are left visible, and every redaction would have an annotation naming the exception being applied (22c or something like that) rather than a reply that simply says "sorry, no docs."
I have no idea how Wisconsin's law is applied, but I wouldn't be surprised to see a release of heavily redacted e-mails and a fake scandal about the redactions.
Having caught up with the thread finally, there's actually some fairly detailed discussions in the archives world about scientific data archives and preservation and access, but I haven't followed any of it. I guess there's some split between "save all data + results" and focusing on saving the results and only some key intermediate stuff. You might think it's an easy answer, but I guess some of the datasets are still quite large relative to storage costs and it's not realistic to save all of it forever. As this comment becomes increasingly speculative, I will guess that the example I hears was regularly updating meteorology stuff, where the data is pretty much constantly accumulating and will continue to do so indefinitely.
34
... Maybe I'm just an atypical slob, but if people really think the norm should be having every step of a process fully documented for public consumption, that's asking a huge time commitment from people. I realize there are some scientists who think this is how things should be, but I think it's nuts. The point of replicable results is that other experts should be able to understand what you did well enough to do the same thing independently,...
Not every step, but if the results in a paper depend on a computer program then the version of the program which produced the results should be available in some form. How are other experts supposed to understand what you actually did without the code? Reproducing what you did by duplicating your programming errors is likely to be rather difficult without your code. Maybe you think you don't have any programming errors and maybe you are right but an awful lot of published work is wrong because it is relying on buggy programs. This is a serious problem and making code more widely available would mitigate it. And it does take more effort to do things right but presumeably that is the object of the exercise.
For an example of the benefits of providing code see here. And before dumping on Levitt (of Freakonomics fame) give him credit for making his code available which many academics are reluctant to do because they are afraid people will find mistakes in it.
Reproducing what you did by duplicating your programming errors is likely to be rather difficult without your code.
But you're not supposed to be duplicating programming errors. If you're actually trying to duplicate the research, you're supposed to be duplicating whatever it is the paper said the code was meant to do. And if you independently write code that does what the original code was meant to do, and you get different results, that's how you find errors in the original paper.
Whatever's wrong with the original code, if it's producing results that correspond to independently written code intended to do the same thing to the same data set, the errors in it (which there always will be) are unimportant.
What kind of an important error could there be that couldn't be found by that method? (I could easily not have thought of something, but nothing springs to mind.)
49: The problem is that any realistic data analysis, there is going to be a data cleaning step. Usually journals don't give authors enough space to describe what they did, so the easiest way to figure it out is look at the code. Also, if you write a program that can't duplicate the results of another program, it's hard to know if the problem is in your code or in their code, unless you can compare them side-by-side.
49
But you're not supposed to be duplicating programming errors. If you're actually trying to duplicate the research, you're supposed to be duplicating whatever it is the paper said the code was meant to do. And if you independently write code that does what the original code was meant to do, and you get different results, that's how you find errors in the original paper.
The idea of a reproducible experiment is that you describe what you did in sufficient detail for someone else to reproduce your results. These results may still be wrong as there may be errors in the design or analysis of the experiment but they should be repeatable.
A paper describing a measurement of say the melting point of lead should not just give a number but describe how it was obtained. A bunch of papers just giving different numbers would not advance the field much.
If you write your own computer code and get different results you don't know (and often have no good way of finding out if the other code and data are unavailable) whether you have a bug, the other researcher has a bug, the raw data is different or the programs are actually doing different things. It is better for the code to be available.
Victoria Stodden is on something of a one-woman crusade/jihad/[whatever the Buddhist analog is] to make releasing code a standard part of reproducible computational science, and has pretty much convinced me. A key point, as said above, is that if I put in the (not-inconsiderable) time to re-implement your computations from scratch and can't, without your code it's really hard for me to know which of us is wrong. Yes, cleaning up code to the point where it can be shared is more work and a pain - but I don't do it, how much can I trust my own results? The issue of motivated mis-understanding is a real one, but I am not sure that not releasing code makes things any better on that front, and I've been convinced the scientific value is real.
I recommend her slides from the AAAS symposium she organized recently; also those of Keith Baggerly.