Adam Mastroianni writes:
I [Mastroianni] have talked to a lot of folks about how to fix science. . . .
Strangely often, I’ll be talking to someone about all the problems with academic science and everything will be normal and then all of a sudden they’ll reveal a bunch of tyrannical fantasies. Force reviewers to sign their reviews and hold them accountable if the paper’s results don’t hold up! Centralize all journals onto a single website and make everybody submit their papers there! Mount an academia-wide inquisition to identify the data-fakers, the p-hackers, and the over-claimers, and march them off to Science Jail! . . .
What creeps me out is that people seem to imagine all of these improvements happening at gunpoint. . . . I love people, but a lot of them—maybe most?—yearn for their side to acquire absolute power. They seem to believe in the authoritarian school of social change: the only way to create a better world is for a strong central actor to force it on everybody else. . . .
I agree, and this reminds me of a discussion we had a few years ago in a post entitled, No one knows what it’s like to be the bad man, where we discussed Thomas Basbøll’s wonderful line:
Andrew Gelman is not the plagiarism police because there is no such thing as the plagiarism police.
Indeed, there is no plagiarism police. Nor is there a bad-science police.
Just one thing. Mastroianni points to this post by Chris Said as an example of someone who wants to “Mount an academia-wide inquisition to identify the data-fakers, the p-hackers, and the over-claimers, and march them off to Science Jail!” And, indeed, Said’s post is entitled, “The case for criminalizing scientific misconduct.”
But . . . “criminalizing” is not the same thing as an “inquisition,” let alone “authoritarian,” “autocratic,” “fascism,” “Gestapo,” or any of the other scary nouns and adjectives that Mastroianni is slinging around.
Lots of things are criminalized in our society: burglary, selling heroin, tax evasion, dangerous driving, . . . all sorts of things! If there’s a law, it doesn’t have to be enforced by an inquisition, an autocratic system, etc etc. We have a justice system in this country! The justice system is imperfect, and I agree you can make a strong argument that criminalizing scientific misconduct would do more harm than good, and I guess my take on it would be that it would depend how serious the misconduct is and what the penalties are—often it seems that recourses to the legal system just make things worse, as in the case of that former Harvard business school professor who’s suing three psychology researchers for 25 zillion dollars because they had the nerve to investigate and write about suspicious things in her public work—; my point here is just that you should be able to accept that certain things are against the law, and work to improve the justice system, without equating law enforcement with fascism.
More specifically, here are Said’s suggestions:
“Danish-style independent committees . . . In 2017, Denmark passed a law establishing a non-criminal independent expert committee to investigate scientific misconduct, a stronger alternative to in-house investigations. If the committee determines misconduct occurred, it may take actions like notifying the university, publisher, and funding agency, and sometimes even the police. Due to its independence, expertise, and narrowly defined scope, committee decisions are considered more credible than decisions from universities.”
A federal criminal statute on scientific misconduct . . . The laws available to prosecute scientific misconduct are mail fraud, wire fraud, and false statements to the federal government. Yet, these are barely ever used . . . Susan Kuzma, a lawyer for the Department of Justice, . . . proposes a new criminal statute directed explicitly at scientific misconduct.
These might be good ideas, stepping stones on the road to a brighter future; they might be bad ideas, leading us down the slippery slope. To be honest, I’ve thought a lot about problems with science but not a lot about solutions, so I won’t offer any opinions on Said’s proposals. In any case, I don’t see them as being authoritarian, any more than it’s authoritarian to prosecute other crimes that damage society, such as polluting the water supply.
So, when Mastroianni writes that some people are “salivating over the idea of a Science Gestapo,” I’d like to see at least one example of this. The only example he gives of this, doesn’t seem to fit the bill.
Summary
I don’t think Mastroianni helps his argument when he points to a proposal for criminalizing science fraud and labels it as autocratic and fascistic. I agree with his large point that the social processes of science are complicated and that there is no science police.
P.S. Just coincidentally, yesterday this post by Joe Bak-Coleman came up, in which he expressed frustration with the scholarly and media discourse regarding the science reform. In short, research quality issues get framed as battles between people, and then there’s a tendency for disputes to get framed in terms of good guys and bad guys.
Criminalization does one main thing, it enhances the power and authority of the enforcement group. But hierarchical power and authority are the main problems in my opinion. That power will be abused. This is basically the thesis of anarchism and I’ve come to realize that’s as close to my own views as I can find in the history of political thought. The basic problem is hierarchy and authority, and the violence that is inherent in its perpetuation. The data fakers, cheats, and science abusers all got there by twisting authority into funding them, protecting them for years, and promoting their work. Universities largely protect their cheaters and bad science promoters. For every one who gets revealed and has some bad consequences there are how many who carry out entire careers applying institutional pressure to extract good wages and wield power over those who would oppose them and suppress the truth?
The problem is institutions and their hierarchical authoritarian nature. The solution can not come from further institutional power.
The more anarchic solution is for people to do their jobs: hiring committees, peer reviewers, personnel committees, granting agencies. If they actually evaluated the quality of work, then many poor practices would be recognized as such. And any perverse behavior on their part would be something they should be held accountable for. So, if Ivy league institution X is willing to keep Y employed despite shady practices, then that institution would have a black mark on their reputation.
The problem is that information is highly imperfect and asymmetric. Technology has now enabled misinformation and plain old “messaging” that renders these feedback mechanisms meaningless. Hence the appeal of centralized or authoritarian systems as “solutions.” I’m inclined to think of the DEI efforts as an example: reputational effects didn’t seem to work, so we got authoritarian DEI policies which we’ve now seen the backlash for.
I tend to agree with you that authoritarian systems do not offer a solution (though I’m not sure I’d go so far as to say that they got us here to begin with). Neither do grass root efforts appear to work. I wish I could offer something that works, but I don’t see it.
Dale:
I guess that the easy way out is just about always to do nothing. Consider the empty reply we received from the University of California’s time-serving academic bureaucrats when we informed them of academic misconduct performed by their famed sleep researcher. Very frustrating, but that’s how it usually goes.
I couldn’t even get the American Political Science Association or the American Statistical Association to rescind the awards they’d given to plagiarists, and you’d think that would be a slam dunk.
There are some cases where scholarly organizations kick out wrongdoers for academic misconduct, for example Harvard got rid of its disgraced primatologist a few years ago, but usually it seems they’d prefer doing nothing rather than to confront the problem.
Dale, this blog represents part of the Anarchic solution. We discuss together each bringing our own knowledge, and if we’re doing it well, we acknowledge the intelligence and creative input of the others who participate, and in the end we get to some kind of distribution of thoughts about the science, and the process is in the public sphere where everyone who is in good faith is welcome to participate and read about the discussions. People can participate here just by being sufficiently sincere, regardless of if they’re a 15 year old high school student in Bangladesh or a professor of statistics in Australia or a climate scientist in Brazil or whatever.
Compare to the institutional view, even if people within Universities “did their job” by evaluating the quality of the work etc, inevitably then their output goes up the hierarchy and someone higher up makes the decision. Fire this person, retract this funding, publicize this error… Or hire this person, give this funding, squash discussion of this error, or whatever. Much of the reason people don’t “do their job” is that in the end it has very little to do with the final decision. Why evaluate subtle variations in applicants history of publication quality if in the end some Dean just overrides you anyway? And also spending all that time doing a good job just harms your ability to game the system into funding your own research.
That hierarchy exists to maximize income and power for the institution, it will almost always reject the democratic consensus of the peers and such if doing so enhances the organization’s power and money. Hence, Columbia still games the US News report, Stanford still supports the guys who said COVID would maybe kill 1000 people or whatever, the Nudgelords still get a seat at all the powerful tables, TED still has the power-pose talk up on YouTube. The sleep researcher who left out data off his graph still has a well respected trade paperback being sold in bookstores or whatever.
Meanwhile, actual progress in science is being made by lots of people who struggle to get funding for their quality work because the funding all comes from hierarchical decision making with a number of bad incentives and systemic problems (such as who gets chosen for the grant evaluation committees and whether grant reviewers are blind to the names and institutional reputations of the submitters and etc etc).
+1
Only thing I’d add is that the bureaucracy seems to become more powerful as the research becomes more divorced from market value.
Corporations expect their research groups to make them money over the long haul, and if they don’t, they get fired. And while it’s hard to measure the value of basic research, companies can still get a sense as to whether its >$0. As a result, things like AI research, imperfect as it is, has broken out of a lot of the institutional pains you describe.
Meanwhile, social sciences, a lot of which doesn’t have quantifiable value, still needs a way to decide who receives the government funding. How do you decide? I feel like however you decide, you’ll end up rewarding some combination of the best researchers who can overcome the system and the politically savvy who can navigate the system. It appears that with the current system, we are also getting hucksters and honest researchers who are readily fooled by randomness.
Regarding the claim that “Criminalization does one main thing, it enhances the power and authority of the enforcement group”… this is probably true, depending on how you define “main thing” and how you define “power”, but it is probably false if you define those terms another way.
Murder, robbery, rape, theft, fraud, are all criminal acts. I would argue that the main thing criminalization has achieved is to make those acts less common than they would be if they were not criminalized.
There isn’t a great dataset but some studies in the 1990’s suggested that 40% of police officers admitted in anonymous surveys to domestic violence against their partners. Other versions suggested it was closer to 30%. We should probably believe it is way higher than in the general population.
Murder clearance rates have been trending down for decades. Before the 90’s the data isn’t particularly reliable, but since the 90’s clearance rates have dropped every decade. It’s down to like 50% these days. Mostly it’s just easy to clear a case where you find one person standing over the body of the other person and they admit to the crime. If that isn’t the case the probability that the police solve the crime is pretty low.
Here’s reported forcible rape rates in the US: https://www.statista.com/statistics/191226/reported-forcible-rape-rate-in-the-us-since-1990/ it seems likely that the massive uptick was the “Me too” movement making people more willing to report? Does law enforcement make rape less likely, or just make reporting it less likely because the victims are so poorly treated? Back in the 2010’s a boy was caught in the middle of the act of raping a woman on Stanford’s campus and held by a number of other students until he was arrested. The judge gave him a slap on the wrist saying something about how he didn’t want to ruin this promising young mans future career or some shit.
In the 80’s we ramped up criminalization of drug possession, sale, etc. We incarcerated millions of young black kids, we had massive gang wars, the DEA got attack helicopters and essentially light infantry and the Coast Guard got funds to do drug interdiction, and soforth and so-on the budgets went way up…. and drug crime increased and drug availability increased, and 25% of black men wound up in prison, and the more effort we put into the “enforcement” the more money there was to be made in the illicit drug market and the bigger the effort there was.
In 2009-ish we criminalized the OTC sale of nasal decongestant. Briefly the price of methamphetamines went up, finally making it worthwhile for the cartels to get into the meth market at industrial scale. Now meth is everywhere, extremely cheap, there are vastly more people harmed by it, and cartels have sewn up that market very well. Plus, chronic sinusitis is way up by comparison. It still requires a full Rx to get sudafed in Oregon.
I would like for law enforcement to reduce crimes like murder, assault, rape, robbery, but I’m not at all convinced that it does. But there seems to be even less evidence that criminalizing things less inter-personal like copyright infringement, or data fraud in science would have much positive effect at all. Particularly if there weren’t substantial funding to enforce the data fraud brigade… and do we really want a division of the FBI digging through everyone’s lab books doing random audits and such? I’m not sure guys with guns in my wife’s bio lab trying to find out if someone really did that surgery on Dec 18th or not is likely to improve anything much. And I’m pretty sure the perpetrators would use their positions of power to get back at those who accused them wielding the law enforcement like a weapon.
A frequent claim is that we need institutions because people are otherwise likely to behave badly. Kropotkin in his essay “Are We Good Enough” turned the tables… we need to eliminate institutions precisely because they are the source of the power that lets people behave badly. I tend to agree with him.
Are you really claiming that robbery, murder, rape, etc. would be occur at the same rate ( or lower) if they were legal?
Phil, police rarely if ever stop crimes, they mostly mop up afterwards. In almost every case where a person prevents a crime it’s through self defense. I’ve read many accounts of people calling the police and getting response times measured in hours to days.
Would more people try crimes if there weren’t a police force “deterring” them? I’m not sure. How much deterrent effect do the police have? Things I’ve read are pretty mixed about the issue. It’s far far from clear that police reduce interpersonal violence. Causality is hard to establish there. It’s particularly hard when you realize that another function that police have is to impose rules that help oligarchy and cause poverty and income inequality which is known to result in increased violence.
Decriminalizing gangs in Ecuador resulted in dramatically lower violence. So it seems to be the case that “would fewer people commit violence if gangs were legal” was answered yes there. I know police and the legal system have a lot of propaganda a “obviously they reduce crime” sort of “common sense” behind them, but that common sense is not supported by very clear data. Decriminalizing drugs in Portugal resulted in a lot of reduced violence and soforth. Similarly in CA in Mendocino county after the decriminalization of cannabis.
What I’m trying to say is, it’s far from obvious to me.
Daniel, whether the police stop crimes is beside the point. The fact that if you commit a crime you may be punished is the point.
If you genuinely think murder, rape, and robbery should be legal then…well, I was going to say “you’re an idiot” but you are surely no idiot. So I’ll just say I strongly disagree.
“Criminalization does one main thing, it enhances the power and authority of the enforcement group…. The basic problem is hierarchy and authority, and the violence that is inherent in its perpetuation.”
Agreed. The violence used by the IRS to fund academia, and it’s subsequent corruption, should be ended by the government not funding academia.
I’m curious though. You seem to strongly favor the left as a practical matter. The same left that routinely “enhances the power and authority of the enforcement group” to enforce societal outcomes they – and you – seem to like quite a bit. Are you willing to go all anarchist with regard to those policies as well?
Anon:
Wait a minute. “The violence used by the IRS”??? Ummmm, I don’t know about that. A quick google led me to this page, which says:
This doesn’t sound so violent to me. Maybe you’re thinking of the FBI?
Andrew, then go ahead and refuse to pay the IRS. Eventually men with guns will show up at your door. I didn’t mean to imply that the armed wind of the IRS would be the ones doing it. The paying of taxes is not voluntary, and when push comes to shove, it’s the threat of violence that backstops that power.
Anon:
I’ve never refused to pay my taxes—that wouldn’t be cool!—but one year I made a mistake, and I got a letter from the IRS with a bill for the missing amount plus a fine. I paid it. I guess if I hadn’t paid it, eventually the cops would’ve come and arrested me? Or maybe they would’ve just put a lien on my salary? I don’t know, but even if I got arrested for it, I wouldn’t call that “violence” on the part of the cops.
None of these definitions apply to the IRS sending me a letter telling me I had to pay my back taxes plus a fine, nor would they apply to someone coming to my apartment to arrest me. Nobody’s trying to injure, abuse, damage, or destroy anyone here; they’re just saying you have to pay your taxes.
And what that all has to do with corruption in academia is . . . zero. You might as well say that the gas line going into my building involves violence because if I stop paying my bill, they’ll cut off the gas. Or you could say that the local state park is violent because if I try to enter without paying, they won’t let me park in their lot. American Airlines is violent because they won’t let me on their plane if I don’t buy a ticket. The local 7-11 is violent because they won’t give me free Big Gulps—and, even if they were to do that, they wouldn’t let me set up a lemonade stand inside their store. Etc etc etc. If you use the term “violence” in this way, it has no meaning.
“I don’t know, but even if I got arrested for it, I wouldn’t call that ‘violence’ on the part of the cops.”
I would say it is pretty standard to think of this as violence. At least since Weber, we describe the state as having a “monopoly on violence” and this of course includes police power.
And the OED’s first definition is “the deliberate exercise of physical force against a person, property, etc.; physically violent behaviour or treatment”.
Dean:
Even if I ignore the letter with the bill telling me to pay my back taxes, and even if the government doesn’t just take it out of my paycheck, and even if the cops come to my door (I think there are still a couple more steps that I’m missing here), it’s still not “deliberate exercise of physical force,” because when those cops come and knock, I’ll open the door and walk out willingly and let them take me to jail. If I resist arrest, that’s another story.
But, again, at that point, if you’re saying that the IRS is “using violence” because of the extreme hypothetical that I underpay my taxes, then I don’t pay when the bill comes, then I avoid various other opportunities to pay, then I resist arrest . . . in that case, just about everyone is “using violence,” including, as I said above, the gas company, the state park system, American Airlines, 7-11, etc. My neighbor is using violence because if I try to build a treehouse in his backyard without his permission, he might well kick my ass. This has nothing to do with the IRS.
I am against authoritarian leftists. I’m not a fan of the US tax code one bit. Yes in general I’d prefer a more anarchic approach to everything. I am happy to admit I don’t know what that would look like in its entirety. In some ways that’s because it would look very different from place to place most likely.
Examples of the kind of anarchism I think has been enormously successful is the development of huge quantities of free software through voluntary cooperation of people working on Linux kernel, R language, Stan, Julia, Python, etc etc. all that put together is probably responsible for say 20% of the global GDP. Id like to see that sort of thing in education and science yeah. I don’t know how to get there. Empowering presidents of USC or Columbia or Harvard isn’t it though.
Daniel:
I can’t speak to Linux, R, or Python, but I will tell you that both Stan and Julia were created by way of a big boost from the taxpayer, in the form of a Department of Energy grant from 2010 or so. And also an indirect benefit from the taxpayer, given that Alan Edelman (one of the Julia founders) and I both benefit from a lot of taxpayer-funded education and research support over the years.
How it was funded vs how it was developed are in some sense two different questions. Julia and R for example aren’t just the language but the whole ecosystem, packages developed by large numbers of people coordinating without direction from a “boss” who determines who does what. Someone who wants to make a new package for Julia doesn’t have to “get permission” from the “boss of Julia”. Someone who wants a new feature in photoshop DOES have to get such permission.
Of course, in the present day, with the whole NSF and soforth system in place it’s great that you guys happened to to manage to get funding from the govt and that funding wasn’t used to say develop further power posing The world isn’t anarchist today, but some projects are more anarchist than others. Adobe Photoshop is developed exclusively by people at Adobe directed by managers who tell them what features to add this quarter, determine which bug fixes will get priority, and fire workers who spend their time on something other than what the boss told them to do.
Daniel
I think you are overly promoting open source. Commercial products have their place and are not evil. It could be my economics training, but I actually believe that markets with well-defined property rights are the best mechanisms we have developed for many of society’s decisions – certainly not all, however. Open source is something to be appreciated, even marveled at. But as a model for all technological development I think it fails. AI will prove to be an interesting case. Much AI development can be “open source” but it is likely that it will look more like commercial development. Without the training data, having open source code for AI will be of limited use.
You seem to view the profit motive as something evil – I do not. I want to see commercial software providers compete for producing easy to use software that works well, and I believe they are entitled to make money off that. Where that system goes awry is when the market ceases to be competitive, as it is with many software products. We (the collective we) have done a poor job of dealing with market power in my opinion. The profit motive need not result in unrestrained market power. Open source has produced some wonderful developments and these can compete with commercial development. They can and should coexist. I think it is a mistake to choose one over the other for all technological development. Similarly, I think it is a mistake to want R&D all conducted by government entities or all by commercial entities – these can also coexist. The problems – and there are many I believe – come in the details: the incentives and regulations that these entities operate under.
Dale, I 100% disagree with the mere existence of copyright. Ideas and their expression are obviously a non-rivalrous public good, and copyright is en enclosure of that through force.
Markets are not evil, I love markets. Profits are not evil per se, when they come from providing a quality service in a competitive environment without protectionism. What’s evil is property rights to wide swaths of resources created out of thin air by the threat of guns and imprisonment and the profits that exist entirely because of those state guns creating situations to move wealth into the hands of small numbers of oligarchs. Some example mechanisms are copyright, patent, tax exceptions (carried interest etc), spectrum licenses, and even to some extent real estate title, mineral rights, and similar laws. Similarly, when it comes to running pipelines through indian reservations, or private ownership and control of massive corporations like Amazon utilizing patent rights as a club to prevent competition, banking and monetary supply control laws that can favor the wealthy, and many other systems which mechanically favor the increase of wealth by the few. This is where I think you and I agree that Economics went off the rails and doesn’t have proper answers to these market power questions. we may disagree on the specifics, but neither of us thinks it’s a good thing that 100 year old recordings and movies and books are being lost to history because it’s illegal to distribute them and the copyright owner prefers to keep them out of circulation to prevent old stuff from competing with Taylor Swift or whatever right?
The best mixture in my opinion has to be something like public ownership of all large corporations. A mechanism that might work for example would be to simply tax shares themselves away from founders as corporations become larger and sell the shares into the market to be bought up by the public, and/or granting them equally to citizens (or granting shares in holding companies). And making more corporations into highly regulated public utilities (where they become monopolistic, Google or Apple might be good examples).
But I think we’re far from the original topic and I’m wary of our host’s preference to keep from wandering too far. I’ve made some provocative statements, but anyone who wants to discuss them in depth can do so on mastodon/fediverse @[email protected] rather than going further off topic here.
Daniel
Well this old dog tried to learn something new. I set up an account on Mastodon but I can’t figure out how to send you a message there. I really don’t have much interest in setting up and regularly using yet another social medium service so I don’t have the patience to figure it out. I was able to find your page, but I can’t seem to post on it and when I go to my page I don’t even see a way to send a message. Supposedly you can find me at @[email protected].
Hi Dale. To start a discussion with someone on Mastodon simply mention their username in your message. I followed your account, and sent a message mentioning you. You should see that message if you log in again, and can reply there.
“refuse to pay the IRS. Eventually men with guns will show up at your door.” 😂
About 500-1000 people per year are convicted of not paying their taxes. The majority serve prison time. The individuals who take them into custody are armed. The people holding them in prison are armed.
Gb:
As I wrote above, what that all has to do with corruption in academia is . . . zero. You might as well say that the gas line going into my building involves violence because if I stop paying my bill, they’ll cut off the gas. Or you could say that the local state park is violent because if I try to enter without paying, they won’t let me park in their lot. American Airlines is violent because they won’t let me on their plane if I don’t buy a ticket. The local 7-11 is violent because they won’t give me free Big Gulps—and, even if they were to do that, they wouldn’t let me set up a lemonade stand inside their store. Etc etc etc. If you use the term “violence” in this way, it has no meaning. This has nothing to do with the IRS.
Intuitively, I would reject the use of criminal law to prosecute scientific misconduct. This is because the existing laws are usually sufficient to prosecute the really bad apples (the Paolo Macchiarini type). However, I have read a lot of disturbing things about institutional cover-ups by universities, ranging from silencing victims of sexual abuse to ignoring clearly falsified research data. And that is just the tip of the iceberg, the stuff that is so bad that it cannot be left unaddressed in the long term without seriously undermining the credibility of science and turning certain subfields of science into what can only be described as Akerlof’s lemon market.
I just do not think that criminal law is the answer. Civil law, however, might be applicable if we look at the narrow field of scientific fraud*. Those who fund research should make sure that if the research they have paid for is not carried out, as evidenced by fraudulent research results, they should demand their money back. It would also help if scientific publishers invested more money and effort in cleaning up their journals when cases of rampant fraud are reported to them. However, the financial incentives seem to point in the opposite direction, so there may be room for institutional change driven by the public.
*As opposed to the wider field of scientific misconduct, which is a subset of general misconduct in academia.
Unfortunately, our justice system also seems to be broken. It is no longer clear that violating a law means you will be punished – it increasingly depends on your politics. Is it far-fetched to imagine that civil or criminal prosecution of bad science will depend on whose interests are served by that science? I don’t think we are quite that bad yet, but rapidly heading that direction. So, criminalizing such behavior is not just ill-advised for many of the reasons given, but also because it relies on a system which itself is breaking.
Let me add another limitation of criminal law: It stops at the border. In a case like Macchiarini’s, the problem was not only the active cover-up by scientific institutions, but also the fact that it happened all over the world. This highlights the need not to look for one policy that will cure all science, but for several changes, each of which would hopefully make the situation a little better.
Raphael K, in an off-hand way, coupled Paolo Macchiarini and Akerlof’s lemon market. I know/knew a great deal about the former but never heard of the latter and was entirely unaware of the (very depressing) 1970 (!) paper by George Akerlof.
The Wikipedia entry, https://en.wikipedia.org/wiki/The_Market_for_Lemons, concludes that it therefore a good thing that the health insurance industry knows everything about each individual’s true health because this avoids being swamped by lemons, individuals who are unhealthy.
I think you are misreading the conclusion regarding health insurance. I don’t think that article (more importantly, most economists) claims that it is “good” that health insurers know everything about each individual. What the claim should be is that without any such information, the market might well be a market for lemons – only the sickest would end up being insured and at very high premiums. This might even preclude the market from existing if the willingness and ability to pay for insurance in these high risk individuals is not high enough to afford the premiums.
To the extent that insurers can obtain information about individual risks, the lemons market may be avoided. Premiums would then match the risks of insuring each individual. But the result would be much like the result of market for many things – for example, luxury yachts. If you believe that health insurance should not be a “luxury market” then the economically efficient solution may not be to your liking (it is not to mine). Forcing people into larger risk pools avoids the luxury solution but introduces problems of asymmetric information and/or cross subsidization. An example of this tension can be seen in the disparate premiums for employer sponsored plans that require participation of all employees (usually with exceptions for employees with other coverage in place) and the open voluntary market for purchasing insurance. Voluntary insurance raises issues of asymmetric information and results in higher premiums than markets where participation is required.
I’ll admit that many economists blur the lines between what is “good” and what is “efficient.” But asymmetric information poses real problems and the good may not be consistent with what is efficient. I believe health insurance in one such area.
Sorry to be sounding like a broken record, but….
The US is the only developed country where this is an issue. In every other developed country (to a very good first approximation*) no one is uninsured. And, of course, there’s no such things as medical bankruptcy or surprise charges. (Out of network docs/hospitals is a unique American insanity.)
And to the best I can tell, the idea that medical care is better in the US is nuts; the only medical care that’s “better” in the US is transplant surgeries, but that’s only because the insane carnage on our roads provides a better supply of spare parts than is available anywhere else. (In Japan, if you are a passenget in a car with a drunk driver, you will lose your license.)
*: One can always figure out a way to mess up (the Joan Rivers problem** could happen anywhere), but unexpected medical costs are rarer by multiple orders of magnitude. Everywhere else.
**: Dying after an off-insurance procedure by a fly-by-night doc. Most places don’t cover cosmetic procedures, of course.
David
Is your comment meant as a response to mine? I don’t see any relevance at all. There is nothing in what I’ve said that indicates support for the US health care system. My own views are quite the opposite. But it remains true that economic efficiency and “good” are two concepts – that hardly overlap at all when it comes to health care. I can’t figure out what you are reacting to.
Next time you have a health issue, even dentistry, go get a second opinion where they have no record of you and tell them you are paying in cash.
Most likely you get offered 50- 99% off and the next available slot. Because this “insurance” is a huge headache for everyone involved.
Try it.
Anoneuoid
This comment was not up to your usual standard – I have no idea what it relates to. The health care insurance market in the US is one of the most dysfunctional markets I can think of. I could write pages about how poorly it operates. I also know that nobody pays list price for anything – and, getting real data on prices paid is nearly impossible, so studying these markets is also nearly impossible.
A better example of how badly it works (than your example of paying cash) is the part D Medicare prescription drug plans. It turns out that 70% of people don’t change their plan each year. My wife and I change our plan every year – you have to because the coverage it provides for various drugs changes drastically every year. Medicare even provides good tools for researching the comparative coverage amounts for every drug you need. But 70% of people don’t bother using those tools and just keep the plan they started with. My plan this year just informed me they were changing their price – increasing my monthly premium by 800%. So, I’m changing to a plan with a $0 premium (how can they charge 0 and make money – it is because Medicare pays them a subsidy). I could go on and on, but I really don’t understand what your point was.
My point is to try it, and avoid this:
https://en.wikipedia.org/wiki/Chargemaster
Just get a second opinion (don’t need to actually use that provider) and see for yourself. I would love to hear back about the experience too.
Sorry to not make sense.
I was trying to say that the definition of “insurance” that allows the insurance providers to play the game of trying to not actually pay out claims (either before the fact by not insuring or charging more, or after the fact by denying claims) so they can make more money is a uniquely American thing, and that thus the ideas of good and efficient here depend on the perversity of that definition.
That is “good’ and “efficient” don’t occur in other countries because the object of the game is to provide everyone with the medical care they need.
sexual abuse =/= scientific misconduct. what does sexual abuse have to do with the credibility of science?
good luck with the institutional change driven by the public.
rampant fraud is very rare.
I am sure research fund granting agencies would gladly spend resources trying to demand their money back in civil court.
Broken record: Independent replication
All that’s needed is for this to be funded at scale, possibly with its own journal.
Then people with standards can start using that, and everyone else can keep working with the wild speculations.
That will still exist, I’m sure, since it appears to fill some kind of societal niche (eg, prestigious middle-class jobs program). But it would play a role more like the modern Catholic Church does, with far less influence.
The question is, what is your goal? There is so much fraudulent science out there that it would be a huge waste of money to refute it all with independent replications. It is cheap to fake research, and it is also cheap to ask your partner in fraud to ‘independently’ confirm your results. I am sure that mandatory independent replications would not stop fraudsters from publishing. I like replications, there should be more replications, but it is not the policy to end all research misconduct. Independent replications would probably help with the many misguided researchers who publish statistical noise while firmly believing it to be a signal.
As an aside, I also read on this blog the recommendation to do a non-independent replication, i.e. to do your experiment with a small sample, improve the protocol and then do it again with a larger sample size.
The rate of fraud is negligible and will not stand up to replication. The goal is for people to publish replicable *methods*. If others cannot read your methods and get the same results, then no one should trust the results.
The real problem is the second group: “the many misguided researchers who publish statistical noise while firmly believing it to be a signal.” Also the related “soft-fraud” which is more like weaselly behaviors (leaving out inconvenient results, etc).
We need to know “if I do x, y, z; then I will observe A, B, C”. Once there are phenomena to “hang our hats on”, we can move forward on the theoretical front. Ie, the theories are then constrained to fit those observations.
Are you sure that the level of fraud is negligible? I think it is very difficult to estimate. The number of fraudsters may be small, but there are people who produce a hundred papers a year. These people are hired not only by the notorious King Saud University or some local university in the backwaters of a less developed country, but also by prestigious universities in Western countries. Some fraudsters become university presidents, like my former university president, Simone Fulda (University of Kiel, Germany). It may not be the Ivy League, but I assure you it is not a backwater. Millions of dollars (or euros) of research can no longer be taken seriously because of this scandal. So I am not sure that fraud should be neglected in areas where there is a lot of money (e.g. some areas of biology and medicine and physics/engineering).
Anyway, I do not want to argue against replication, except that I do not see it as a panacea for science. I appreciate your point about replicable methods and agree that often, better explanations of the methods would help.
About 10% of the spinal cord reproducibility project replicated. For the cancer reproducibility project it was about 25%.
The NIH budget is about $50 billion/year. So that is about $40 billion per year from that funding agency alone. And we only have rough heuristics to guess which study/experiment would replicate vs not. So it really may as well be the entire thing.
A few outright fraudsters are negligible and will be filtered out anyway.
How much fraudulent science is out there?
You think independent replications would be conducted by collaborators on the research being replicated?
I’d like to see universities develop consortia for investigating research misconduct. At the institution hosting the researcher, the only role would be to facilitate access to materials and people; the investigators would be from other institutions without an incentive to gloss over misconduct.
Failing that, imagine how much fun it would be to have demonstrably robust research misconduct practices included as part of university accreditation criteria.
Republicans in Florida think they know how to deal with ideas they don’t like: https://www.nytimes.com/2024/11/21/us/florida-social-sciences-progressive-ideas.html
I tend to agree with them that post-modernists in some departments are a problem, but this does not seem like the way to deal with it.
The linked article says nothing about post-modernists.
fyi, post-modernists was never a thing in the socials sciences, they infested the humanities.
I don’t pretend to have a solution to this problem, but it will help to have a better developed understanding of the structural forces (and not just individual corruptibility) that underlie it. As it happens, I just published a paper that includes an impressionistic, but I think instantly recognizable, account of clientelism in academia. I posted this excerpt in a Substack that you can read here: https://open.substack.com/pub/peterdorman/p/clientelism-alive-and-well-in-academia?r=b8ew&utm_campaign=post&utm_medium=web
The key point for this discussion is that clientelist academic networks are most likely inevitable and even have positive aspects, but they are problematic when they crowd out more disinterested evaluations of scholarly merit. And that’s what we’re talking about, right? A star professor can enlist grad students and junior acolytes to support their research program, and inevitably some of these programs will encompass shaky methods and conclusions or even outright misconduct. But the crisis occurs when the institutions that are supposed to be above such personalist networks abdicate or are even colonized by them. At this point it becomes an empirical question: what motives disincline Stanford, Harvard, the various journals, etc. from exercising independent judgment of star academics and their networks? How important are economic considerations? Status enhancement through association with influential stars? Direct capture?
If you think of it in terms of clientelism, aspects of Daniel Lakeland’s anarchism suggest themselves: measures to reduce the concentration of valuable resources—publication prospects, jobs & tenure, promotion, prestige—in those occupying a privileged status. A certain amount of clientelism is unavoidable, but it needs to be pushed back into the interstices of institutional academia and not allowed to overcolonize it.
Thanks Peter I will take a look at your post in a bit.
I like these pie in the sky ideas! Yes, let’s reduce “the concentration of valuable resources—publication prospects, jobs & tenure, promotion, prestige—in those occupying a privileged status.” How would you do that? How should valuable resources be allocated? A lot of it is already allocated by non-academic considerations, for example, demographic merits. Do you think that should be even further developed?
What would “exercising independent judgment of star academics and their networks by Stanford, Harvard, and the various journals” look like? What is “independent judgment “? Independent of what? How would it be practically exercised? How would you define these “star academics”? How would you define their “networks”?
This post has reminded me of some of Andrew’s comments in:
https://statmodeling.stat.columbia.edu/2017/10/18/beyond-power-pose-using-replication-failures-better-understanding-data-collection-analysis-better-science/.
Incidentally, I received a few weeks ago Harvard Law School Executive Education’s Quarterly Newsletter, advertising “Meet Amy Cuddy: Amy Cuddy, a social psychologist who focuses on presence, non-verbal behavior, performance under stress, and more, will be speaking in the 2025 Women’s Leadership Initiative. Watch Amy’s TED Talk that has been viewed over 65 million times, Your Body Language May Shape Who You Are→”
I’m just a hoi polloi commenter, so haven’t really thought out the full explication, but I suspect two major issues, influencing some of the problems area:
1. The massive size of Big Science and Big Academia. Just look at number of researchers. It’s not like just Moissan and Michelson off doing their 19th century breakthroughs, while wearing dapper suits and smoking cigars. You have a massive enterprise of people slugging away like slaves working on the Pyramids. ;)
2. The tournament system for grants and professorships. I have personally seen hype science in effect. And making the downsides low and the upsides high leads to more cheating. The behavioral literature is rife with findings to this effect.
2.5. Joking (OK, half joking)…the whole thing makes me think of the Voltaire comment on the British, “In this country, it is thought to shoot an admiral from time to time pour encouragez les autres.” And I never saw anyone shot in the USN. But I did see how they handled collisions or groundings. “A grounding at sea will ruin a captain’s whole…day.” And it doesn’t matter if the skipper is asleep when it happens. The Navy doesn’t like unlucky skippers. :)
Anon:
Regarding the tournament system (point 2 in your comment): yeah, I’ve often thought this is something that’s very funny about academia (and not just science).
On one hand, academia and science are fundamentally cooperative efforts, both locally (researchers collaborate with each other) and globally (in that research results build upon what was done in the past). And, most of the time, doing science feels cooperative, both with our colleagues and with researchers of the past and future. When I’m writing a paper, I’m working with coauthors and also feel myself in conversation with the authors of the cited papers (even if I’ve never met these people) and with future readers of my work.
On the other hand, so much of academia is structured as tournaments: getting into a top university, getting to work in a good research group, getting an academic job, getting papers published, getting grants, getting awards, etc. Some sort of competition seems necessary, but some of these tournaments just seem unnecessary. For example, various orgs within the university set up “seed grant” programs creating mini-tournaments to get small amounts of funding, and this requires a process of mini-grant reviews.
Unrelated, but also an issue, is the seemingly inexorable growth of paperwork and legalism within the academic and corporate worlds.
One of the often unweighed risks is that criminalization would represent novel levers to suppress scientific inquiry. Scientists routinely engage in research that is unwelcome by governments and industry, how certain can we be that scientists that produce this valuable research won’t wind up dragged through a criminal trial and punished for purely political reasons? It’s easy to forget this when fraud surrounds hypotheses that truly don’t matter, but harassment and suppression of science in other domains is much more salient and a very real threat.
Perhaps less extremely, it’s unclear what the philosophy of punishment is for criminalizing misconduct. Is it to reform? Most would agree that conduct rising to criminal levels under any reasonable definition warrants excommunication from the academy. Is it a deterrent? What is the marginal gain in deterrence of imposing fines or whatnot beyond risk of losing a very secure job and career? More generally, do we have evidence for deterrent effects of criminalization that would apply here?
As with criminal justice, it may be worthwhile to think beyond punitive approaches and more deeply consider the conditions that give rise to bad behavior in science. Where does it crop up and where doesn’t it? These are scientific questions that we can certainly ask but may not if we’re lost rhetoric, terminology and debates that miss nuance of how science unfolds across disciplines.
As we all know, the research that scientists routinely engage is not only, or even primarily, unwelcomed by governments and industry by also by various groups inside and outside academia. These groups have no qualms about dragging people through criminal trials and punishing them for purely political reasons.
Consider the vitriol spewed at imagined academic misconduct here. It shows that the philosophy of punishment for criminalizing academic misconduct would be that warm feeling of revenge.
Is there evidence for deterrent effects of criminalization anywhere?
For some time to come the inconsistencies, mistakes, and backroom manoeuvring during Covid will dominate the public image of “The Science” and it seems likely that committee hearings will expose more of all of these. This history suggests that official powers to sanction bad science will be wielded by entrenched authorities against their critics, to the detriment of scientific progress and the public good.
There is an open call for papers on this and related topics in the journal Theory and Society currently: https://link.springer.com/collections/ajjejhjfge
Olip:
Interesting. I guess I should write something for this issue, as indeed it seems right up my alley. I’m concerned about a couple of things in that call for papers, though:
1. They write, “And how should we think about those who rigorously enforce data-sharing norms and scrutinize research practices (sometimes called ‘data police’)?” For reasons discussed in the above post, I don’t think the term “police” is at all appropriate. “Vigilantes” would be more accurate.
2. The focus of their sample topics is almost entirely coming from within the academy (for example, “Is it legitimate to suppress scientific findings due to concerns about misuse? Is it acceptable for academic journals, professional societies, or universities to censor, censure, or formally punish scientists for their morally undesirable empirical conclusions?”), but many of the problems they’re discussing come from outside. Here I’m thinking of the news media (NPR, Gladwell, etc.) that regularly promote bad science, non-academic organizations such as Ted and the now-defunct Edge foundation that have promulgated the science-as-hero narrative, ideologically-based outside funding of research, and politicians or political operatives who try to suppress dissenting views of American history—this last item is most notorious at the secondary school level regarding the textbook approval process, but it happens at the university level too.
This is not to deny all the problems coming from within academia—for example, the administrators of the University of Nevada have no one to blame but themselves from hiring that “torment executioner”—; I just think it’s a mistake for them to entirely focus on the motives and actions of people on the inside, without considering the outside pressures and incentives affecting them.
Hmmm, they want 1000 to 6000 words . . . Maybe this comment can be the core of my article for them!
Quickest solution I can come up with is to claw-back all grants, stipends, research funding, etc. given to people or research groups who commit fraud. Ideally with penalties tacked on. If the professor or lab can’t pay, the university is liable. Or split the liability between the professors and the university.
I don’t know what solutions universities will come up with, but under that regime I’m positive they’ll find solutions. I suspect there are already general anti-fraud stipulations written into a lot of funding agreements; they’re just not enforced and treated as meaningless boilerplate.
While I am very sympathetic to the idea of material consequences for acts of scientific misconduct and fraud, I am not so sure about criminalization.
Indeed, this is a rare occasion where I’ve heard discussion of “criminalization” rather than “decriminalization” — though of course, the tightening of regulations and banning certain actions are how one speaks of the former.
As others have mentioned, criminalizing acts puts legal power in the hands of governments to carry out actions against those who perform those acts. I am very much not a libertarian/anarchist, but even I can see that the outcomes will depend strongly on who is in power and how politicized enforcement agencies are.
John G Williams above mentions Florida as an example of how the state government has aggressively used its power to intervene in all levels of education. One of the figures at the heart of that process is Chris Rufo, who was also the person who kept spoon-feeding the mainstream media details about the alleged plagiarism committed university presidents. He has been very open about how he intends to push his idea of education into the rest of the US, and there should be no doubt that he and others like him would use any new legal powers against misconduct to hound researchers at all levels, however spurious the charges may be.
We should not just keep in mind Andrew’s favorite examples of the impossible shredder, bottomless bowl of soup, or the more recent appearance of “torment executioners” in academic papers. Consider things of the scale of the UEA email “climategate” scandal: it isn’t hard to see how a motivated authority could go after the (perhaps too withholding) researchers for misconduct. Or even the Flint water crisis: the local authorities denied that the water was contaminated. This is before we start on topics like gender and sexuality.
There is already a good example at the federal level of how the US government uses its legal powers to control the direction of scientific research: under the influence of the NRA, Congress passed the Dickey amendment to stop the CDC from collecting statistics on gun violence. It’s not so hard to imagine the tobacco lobby could push for research showing evidence of the harms of smoking/vaping be prosecuted as misconduct.