Join 3,520 readers in helping fund MetaFilter (Hide)


Social Scientist Diederik Stapel Dismissed for Academic Misconduct
April 28, 2013 12:53 PM   Subscribe

“It was a quest for aesthetics, for beauty — instead of the truth.” The recent dismissal of Dutch social scientist Diederik Stapel for academic fraud [SLNYT] leads the article's author to wonder whether "the scientific misconduct that has come to light in recent years suggests at the very least that the number of bad actors in science isn’t as insignificant as many would like to believe."
posted by Rykey (75 comments total) 16 users marked this as a favorite

 
“It was a quest for aesthetics, for beauty — instead of the truth,” he said. He described his behavior as an addiction that drove him to carry out acts of increasingly daring fraud, like a junkie seeking a bigger and better high.

I'm really looking forward to the comments here, because I think scientific methodoligy is something Metafilter does very well. But that there is the hallmark of an asshole.
posted by nevercalm at 1:07 PM on April 28, 2013 [1 favorite]


Ehhh nobody should really believe social science anyway.
posted by zscore at 1:14 PM on April 28, 2013 [3 favorites]


I certainly don't
posted by thelonius at 1:19 PM on April 28, 2013


I hypothesize that a majority is going to say that all data is faked, just in subtler forms; and someone is going to say that academics commit fraud because the stakes are so low.
posted by tel3path at 1:21 PM on April 28, 2013 [4 favorites]


But he was caught, so the system works, right? Isn't that what they say about fraud and abuse outside of academia?
posted by ceribus peribus at 1:28 PM on April 28, 2013 [1 favorite]


The disproportionate dependence on metrics and credentials rather than quality and value of insights and creativity has led to this shameful series of frauds and fakes and plagiarists galore.
posted by infini at 1:30 PM on April 28, 2013 [9 favorites]


Science has always been this way. The only thing unique about social science at the moment is that they are getting more press because people can actually read and understand some of it. During the same time frame there have been some massive frauds in medicine and genetics as well as the usual daily dribble of smaller frauds in field less comprehensible to the lay people.

This is science working. It isn't perfect and doesn't pretend to be (except maybe in intro courses and grant applications) but it is eventually self-correcting.

So don't trust new social science. Don't particularly trust any new single study science in any field (Nutrition I am looking at you!). Let it age a bit first. See if it holds up to more scrutiny.
posted by srboisvert at 1:31 PM on April 28, 2013 [25 favorites]


And what do you think tel3path?
posted by zscore at 1:34 PM on April 28, 2013


Let's head off the notion that no one in the "hard" sciences has ever cooked the figures or just outright made shit up for a paper right away. And, WRT whether or not you choose to "believe" social science (I think it's fairly ridiculous to ascribe the credibility of any branch of science to your personal beliefs), Stapel was falsifying objective data: the observation of definable behavior. tl;dr--RTFA.
posted by Halloween Jack at 1:34 PM on April 28, 2013 [15 favorites]


If you don't "trust" science, then I guess you trust, what, God? Talk radio? "Common sense"? Of course it's a canard to say "trust" because science is not based on trust anymore than Wikipedia is a trust-based website. You can trust it to some degree for some purposes but you should also verify and keep an open mind that things might not be fully accurate.
posted by stbalbach at 1:36 PM on April 28, 2013


This story would have been a lot better if the writer hadn't bought Stapel's "everyone does it to some degree" line. Outright fraud is a much different problem from unconscious bias and from tactics that are iffy, but accepted. Those should become unacceptable, but this will not solve the problem of people like Stapel. Conflating the two merely gives him cover and undermines science.

The whole point of science is to reduce biases in data— if it's not doing that, this is a problem. But that's about as different from Stapel as tax fraud is from using legal tax shelters. One is a problem of regulation, the other is a crime.
posted by Maias at 1:44 PM on April 28, 2013 [8 favorites]


Let's head off the notion that no one in the "hard" sciences has ever cooked the figures or just outright made shit up for a paper right away.

Has anyone ever claimed that?
posted by normy at 1:45 PM on April 28, 2013 [1 favorite]


Has anyone ever claimed that?

Probably implied by zscore.
posted by Pyrogenesis at 1:47 PM on April 28, 2013


I've been involved in academic psychological statistics for three years, and I'm pretty skeptical of the field right now ;-)
posted by zscore at 1:51 PM on April 28, 2013 [1 favorite]


This is science working. It isn't perfect and doesn't pretend to be (except maybe in intro courses and grant applications) but it is eventually self-correcting.

Just going to repeat this. Every time there's some issue of "oh noes something science said turned out to be false", step back, don't worry, and remember that's how science works. Some things we think we know now are probably false, maybe some fraudulent, we just won't find out which things until we do some more science. This is the normal process.
posted by kiltedtaco at 1:59 PM on April 28, 2013


As fuzzy as postmodernism can sometimes be, can we agree that it's not the same as lying? Bad postmodernism isn't even falsifiable.
posted by justsomebodythatyouusedtoknow at 2:03 PM on April 28, 2013 [2 favorites]


I've been involved in academic psychological statistics for three years, and I'm pretty skeptical of the field right now ;-)

Is that even a thing people can be 'involved in'? Research psychologists don't tend to have statistical consultants and outside of neuroscience I am pretty sure they almost all do their own stats. You're pretty much either a student or a psychologist if you are doing psych research.
posted by srboisvert at 2:04 PM on April 28, 2013


Retraction watch is great, thanks for bringing it to my attention. I see that the honest mistake stories outnumber the fraud ones.

Like some dude who realized he had used palladium instead of rhodium. Or the team that realized too late that they had ordered the wrong mice. Or the team that re-ran all their expirements until they found out that a filter was leaching something into the sample. All these people retracted their own papers as soon as they found out, apologized, and took the trouble of explaining exactly how the mistake was made so that others would know how to avoid it.
posted by Doroteo Arango II at 2:06 PM on April 28, 2013 [3 favorites]


Earlier Mefi post about Stapel. (Note: This current FPP is not a double; I'm just including this link for background, especially for the link to the PDF of the Dutch committees' reports on Stapel's misconduct.)
posted by dhens at 2:09 PM on April 28, 2013


Bad postmodernism isn't even falsifiable.

I have come to think that "science must be falsifiable" is the "correlation is not causation" of scientific methodology.
posted by Pyrogenesis at 2:15 PM on April 28, 2013 [5 favorites]


Research psychologists don't tend to have statistical consultants

If you are using statistics in your work, you either are competent or you have a consultant. Many psych programs have a number cruncher on staff, others collaborate with someone in the math department.
posted by psycho-alchemy at 2:15 PM on April 28, 2013


I'd consider this "science working" over the medium term, and science works great over the long term, but short term Stapel's appointment, grants, etc. all wasted resources.

Worse, these faked results deprive legitimate research and avoid detection by seeming decisive enough to close up avenues of research. So nobody gets grants for subsequent work that might debunk them. So clever but pedantic people don't get grants or promotions.

Even worse, Reinhart and Roghoff faked data helped justify the austerity measures currently impoverishing millions.

Julian Assange has advocated for scientific journalism where primary sources should be reported along with journalistic stories. Can we now have scientific science where datasets and tools are published along with journal articles?
posted by jeffburdges at 2:15 PM on April 28, 2013 [2 favorites]


I don't think it's so much a quest for aesthetics, as it is a quest for grants and other funding.
posted by Chocolate Pickle at 2:18 PM on April 28, 2013


STEM focus, the grant writing race, Science-By-Press-Release, etc. all contribute to a culture that MUST PUBLISH SOMETHING OH GOD, MAYBE IF I JUST CROP THIS GEL.
posted by Slackermagee at 2:20 PM on April 28, 2013 [4 favorites]


Certainly no field is perfect (I work in a "hard" field, and while I think outright fraud is rare, the file-drawer effect and other intellectual misdemeanors have no doubt set us back.) But priming research is really something else. Kahneman, hardly an unsympathetic judge, famously predicted " a train-wreck looming" for the field last year. Said train has probably not yet slid to a halt.

Obviously social science is more difficult in the sense that you're working with systems which are orders of magnitude more complex than those studied in physics, chemistry and biology; an effect size considered utterly unremarkable in those fields is often a cause for celebration in social psych.

The other distinction is that the hard sciences are relatively immune from politicization in the sense that no group senses a political stake in the answers to the questions of whether the Schrodinger equation can model the hydrogen atom or how transcription factors bind DNA. (Human genetics was a partial exception in an earlier era of paucity of data, but this is changing quite rapidly.) People care deeply, on the other hand, about whether stereotype threat is explanatory. The fact that priming research is highly congenial to the politics and general worldview of the people who tend to study it has made it all the more difficult to do good work in that field.

Can we now have scientific science where datasets and tools are published along with journal articles?

We're already on it. We are not nearly there yet, but things are a damn sight better than they used to be.

STEM focus

Slackermagee, I'm not sure I get it? Surely if I'm contemplating cropping my gel, I've already made a decision about my area of focus, no? Would more courses in Brit Lit have steered me away from perfidy?
posted by lambdaphage at 2:23 PM on April 28, 2013 [2 favorites]


This is science working.

A bit too pollyana. It took years/decades to get to the point where somebody got suspicious enough to do some independent research, more to actually get Stapel dismissed.
posted by MartinWisse at 2:29 PM on April 28, 2013


Is it worth trying to differentiate among different scientific fields by asking "What depends on the validity of results?"? If you're a biomedical researcher who fabricates a result, sick people will receive a treatment and not get any better. I'm trying to think what could depend on Stapel's research and I'm drawing a blank.
posted by benito.strauss at 2:35 PM on April 28, 2013


When Zeelenberg challenged him with specifics — to explain why certain facts and figures he reported in different studies appeared to be identical — Stapel promised to be more careful in the future.

Okay, here's what I don't get: if you're an otherwise brilliant person who's decided to take the spectacularly dangerous path of making up data, why in the heck would make what appears to be a rookie mistake like that? Hubris?
posted by Mooski at 2:36 PM on April 28, 2013


>Can we now have scientific science where datasets and tools are published along with journal articles?

We're already on it. We are not nearly there yet, but things are a damn sight better than they used to be.


There's a side issue, which I certainly don't claim to know enough about, but I believe my university would claim they own our data (if we, er, had any) and that we can't just stick it on our websites, as we might be wont to do.

Or so I gathered from some ethics training. I don't think the person doing the training was meant to be telling us we could write grants to remove the university's intellectual property claims...
posted by hoyland at 2:39 PM on April 28, 2013


Wait until people really dig into how money distorts science. It's gotten to the point where a huge percentage of research is some parts of nutrition, medicine, genetics and agriculture is directly funded by commercial interests. And surprise, surprise, the vast majority come to conclusions that benefit their funders. So what that they state their conflicts of interest? It's still a huge waste of time at the very least (the moment I see "funded by the milk council", "almond growers association" or any million of other interests, I immediately stop reading the paper). What's worse, is the fact that this stuff is unethical - how can you intentionally design a study in such a way that it leads to a given conclusion, which you then exaggerate, and not call it anything other than FRAUD? You intend to mislead. The fact that you are doing so on behalf of commercial interests rather than your own immediate ones, does not make it any less of a fraud. There are scientists who supplement their research with this kind of funding, and then there are those whose entire careers are sustained by them.

So when we speak of how much fraud there is in scientific research, we'd be well advised to define our terms, because the answer may vary wildly depending on what we regard as fraud.
posted by VikingSword at 2:42 PM on April 28, 2013 [6 favorites]


Jesus, that M&M experiment is a total WTF! And would have been even if it were done honestly.
posted by 2N2222 at 2:43 PM on April 28, 2013


One strange aspect of this is that Stapel completely and self-consciously sacrificed any chance that he would ever make a real scientific breakthrough in order to pursue this fraudulent program:
The key to why Stapel got away with his fabrications for so long lies in his keen understanding of the sociology of his field. “I didn’t do strange stuff, I never said let’s do an experiment to show that the earth is flat,” he said. “I always checked — this may be by a cunning manipulative mind — that the experiment was reasonable, that it followed from the research that had come before, that it was just this extra step that everybody was waiting for.” He always read the research literature extensively to generate his hypotheses. “So that it was believable and could be argued that this was the only logical thing you would find,” he said. “Everybody wants you to be novel and creative, but you also need to be truthful and likely. You need to be able to say that this is completely new and exciting, but it’s very likely given what we know so far.”
Because everything he claimed had to be something other people already expected to be true.
posted by jamjam at 2:56 PM on April 28, 2013 [2 favorites]


Peter Watts: Because As We All Know, The Green Party Runs the World.
Science doesn’t work despite scientists being asses. Science works, to at least some extent, because scientists are asses. Bickering and backstabbing are essential elements of the process. Haven’t any of these guys ever heard of “peer review”?
posted by the man of twists and turns at 3:40 PM on April 28, 2013 [2 favorites]


I've always had a morbid curiosity, when I was still in academia, about what happens to researchers who are dismissed for fraud. It felt kind of taboo to even discuss fraud and the consequences with virtually every PI I've come into contact with.

I recall that the S.Korean stem cell guy who coerced his students to "donate" their own ova in addition to fraudulent use/manipulation of images ended up starting a company to clone pets, but what happens to other researchers who've been dismissed?

It's not like anyone would ever hire them again for an academic job, and probably the same for industrial research. Do they go sell used cars or work the cash register at the grocery store? Would even a community college hire them on as a non-research lecturer?
posted by porpoise at 3:51 PM on April 28, 2013


If you don't "trust" science, then I guess you trust, what, God? Talk radio? "Common sense"?
people on forums who use memes i understand
posted by This, of course, alludes to you at 4:11 PM on April 28, 2013 [2 favorites]


I am not surprised when I hear that someone has falsified data. I have worked with people whose data I did not trust. I have known labs where data was published that was later found to be incorrect. Sometimes it was retracted and sometimes it was not. I have known labs where an individual would sabotage other's work. I worked in a lab where someone gave someone else deliberately a plasmid that expressed a protein other than the one s/he said it did. The same sized protein, so it took months of work to discover the error.

Like any workplace, there are dishonest people. Any particular field that declares that it does not have the same problems as the rest of the world is being dishonest in that declaration.
posted by sciencegeek at 4:15 PM on April 28, 2013 [2 favorites]


His grad students snitched on him is the part I like.

Next thing we'll be reading the okcupid "research" that people in Seattle don't take showers or baths very often is bogus.
posted by bukvich at 4:46 PM on April 28, 2013 [1 favorite]


There are some fields which have tons and tons of bad science; usually they are fields with political interest and thus funding beyond their capacity. This type of social science is an obvious contender, but there are plenty offenders in the "wet" sciences as well, and there is some forgery within humanities, though the competition is a lot tougher, and more controlling. It's depressing because of the waste of resources. Maybe we could have cured cancer, or we could even have become more knowledgable about the conjugation of verbs in medieval Slavic languages. Instead, we've posted money into fraud.
All politicians should know: obviously, it is good to fund research which might lead to innovation or medical break-throughs. You can help create new jobs, or save lives. Who wouldn't want that? BUT: it is possible to overfund an area within any given line of research. Overfunding is when you provide more money than the available talent can use. (Available includes import from other states/countries). The problem is not only that you pay for more than you get, which is bad in itself, it is far worse. When you have an overfunded field of research, vultures will arrive, and take over the whole field.
Innocent policy-makers might think that the core research will continue, and then obviously a few million are wasted on the vultures. That is ok, one can't make an omelette without breaking eggs. But that isn't how it works out, because the vultures have a legitimacy problem. Research vultures are people who invade a field with too many resources. Generally, they know very little about the field beyond the very basic textbooks. So when they meet the actual scholars/scientists, they have to make the case that their "research" is different because it is more innovative/cross-boundaries/politically correct/industry-directed, whatever. Obviously, the researchers who know the field will immediately call BS. But research vultures are herd animals. They are excellent at creating networks, journals, conferences etc, where they utilize the instruments of peer review against it's purpose. And eventually, the actual scholars will find themselves ostracized from their own field of research. This is necessary from the vultures' point of view, because otherwise, there is always the real risk of the field as a whole losing funds.
As in the FPP, graduate students are the most able when it comes to exposing this. They have less to lose, and more to gain. Even the vultures have to select excellent graduate students, because studying is hard, and more corruptible students fail too much.

What has always puzzled me, observing this from a relatively safe spot, is: the vultures and fakers are almost always highly intelligent and they spend tons of energy on creating their fake science. Why don't they just do right? All of that intelligence and time and power could serve humanity just fine, if it were put into real research.
Also, with a few exceptions, they are always caught. If they put their brilliant minds into actual research, they would have long careers, rather than a short(ish) sniff of glory and then an endless trip downhill.

Also, in this case as in many, it seems safe to say that even if the university management haven't known the problem in advance, they do now, and they need to protect their other employees and students by sending very clear messages out about the situation.
posted by mumimor at 5:03 PM on April 28, 2013 [4 favorites]


Is that even a thing people can be 'involved in'? Research psychologists don't tend to have statistical consultants and outside of neuroscience I am pretty sure they almost all do their own stats. You're pretty much either a student or a psychologist if you are doing psych research.

Um. Plenty of research scientists don't collect their own data. They can be funny that way.

There are droves of research staff who are neither psychologists, nor students, but who may recruit research subjects, obtain informed consent, administer measures, conduct research interviews, prepare research case reports, elicit life narratives from participants, review medical charts, etc etc etc etc.

You could also be negating the work of the staff of institutional review boards.

Also, not sure why you think students who are involved in research wouldn't form opinions based on that involvement.

Also - it's very strange that you think the primary work of research has to do with the statistical analyses. There are so many fudge-able moments before math happens.
posted by vitabellosi at 5:35 PM on April 28, 2013


His grad students snitched on him is the part I like.

Yeah, that's how Marc Hauser was exposed as well. To paraphrase Fight Club: "The graduate students you are training are the people you depend on. We run your subjects, we do your statistics, we make your figures, we write your papers while you sleep... Do not fuck with us."
posted by logicpunk at 5:58 PM on April 28, 2013 [3 favorites]


I have come to think that "science must be falsifiable" is the "correlation is not causation" of scientific methodology.

Then you don't understand scientific methodology, because untestable science isn't science, it's Cargo Cult Science.
posted by Blazecock Pileon at 6:04 PM on April 28, 2013 [4 favorites]


the vultures and fakers are almost always highly intelligent and they spend tons of energy on creating their fake science. Why don't they just do right? All of that intelligence and time and power could serve humanity just fine, if it were put into real research.

I think that if you dug deep enough into their motives, you would find that they had an idea of what should happen, and then when reality turned out to not agree with them, they decided to say screw reality, I'm going to show that I'm really right. You see this a lot lately in political science (theoretical and applied), but it seems to appear in all fields. This is why peer review is important.
posted by mephron at 6:06 PM on April 28, 2013 [1 favorite]


The really bizarre thing about this case is that he didn't even bother to conduct the studies at all.

Convincing high-school kids to eat M&Ms for a research study isn't a particularly expensive or time-consuming task, and having a corroborated venue and pile of signed participant forms would have make him far harder to catch. Visiting the actual train station where he claimed to be carrying out an observational study seems like a pretty obvious first-step to fraud. Besides, wouldn't it be more fun to see if the experiment actually generated the outcome he was hoping for before inventing fake data?

It's tempting to conclude from stories of fraudsters who've been caught that only the mind-bogglingly stupid and lazy commit fraud. (That's a dangerous temptation.)
posted by eotvos at 6:11 PM on April 28, 2013


It's a really interesting article, but then there's this:
The experiment — and others like it — didn’t give Stapel the desired results, he said. He had the choice of abandoning the work or redoing the experiment. But he had already spent a lot of time on the research and was convinced his hypothesis was valid. “I said — you know what, I am going to create the data set,” he told me.
This appears to be the decision the whole rest of his career hinges on -- and yet there's only the barest of motives sketched out. If he was so good at generating plausible-sounding hypotheses based on current research, why such an attachment to this particular experiment? I mean, people throw out weeks and months and years of data all of the time. Again, we're forced to speculate -- career panic? laziness? lust for fame? arrogantly placing his intuition above data? coin flip? Sort of frustrating.
posted by en forme de poire at 6:19 PM on April 28, 2013 [1 favorite]


This type of science is building theories about large and complex systems from tiny amounts of data. The only way I can see to to do this with any kind of integrity is to either dump the experimentation or have some of the people doing it be absolutely focused on data crunching and math.
posted by rdr at 6:49 PM on April 28, 2013


Has anyone done any research into the falsifying of scientific studies? (I'm happy to do this research if hasn't already been done - I promise interesting results).
posted by el io at 7:08 PM on April 28, 2013


Science doesn’t work despite scientists being asses. Science works, to at least some extent, because scientists are asses.

Those who liked this will also like

Hull, D. L. 1988. Science as a process: an evolutionary account of the social and conceptual development of science. University of Chicago Press, Chicago, IL. xiii+586 p. (Amazon warning)

In it Hull concludes that science advances through survival of the fittest. To paraphrase: contending ideas are advanced vigourously, and often in a biased and even vicious manner. Ultimately the truth emerges from the messy battlefield.

Scientists are people. All we can ask of them is that they follow the rules of evidence, deal with all of it, and that they don't make it up. When they don't do that, they lose, sooner or later. Ultimately, the only useful test of a claim in science is the answer to this question: Is it true? We can always test for that.
posted by dmayhood at 7:16 PM on April 28, 2013 [2 favorites]


Has anyone done any research into the falsifying of scientific studies?

I don't know that anyone has gone around trying to falsify other people's studies since it's easy not to get the same results they did no matter how honest and meticulous they were. Just screw something up.

By the time you've checked every possible failure mode to make sure the discrepancy isn't no your end, people working in that field will have already tried to run with it and figured out that they can't get very far without running into a dead end. If no one can replicate your data, it ceases to be relevant pretty quickly, whether you are being deliberately deceptive or just happen to be wrong for some reason.
posted by Kid Charlemagne at 8:03 PM on April 28, 2013


The really bizarre thing about this case is that he didn't even bother to conduct the studies at all.

Convincing high-school kids to eat M&Ms for a research study isn't a particularly expensive or time-consuming task


I'm sure some of his motivation was laziness. (After all, as you say, why wouldn't he at least have checked out the train station to make sure it was as he described?)

Although I am not a psychologist, for the first time at the moment I am collaborating with one and holy hell does it take a long time to collect experimental data. It's been 8 months now and we are just about up to the 50 participants we needed to run. It's only a 30 minute simple task on a computer (clicking buttons in response to certain things, a little bit of a questionnaire, etc). But despite advertising widely, we only get a couple of participants a week. And our research assistant needs about 15 minutes travel time/set up time before and after each participant, and then it takes us another half hour or so to extract the data we need from the recordings/questionnaires, and another hour on top of that to analyse it so that even have numbers we can put into our statistical analysis. If we were to falsify datasets, that would save us around a year of time passing, or about 150 consecutive hours of work.

Even if you are doing something like approaching people at train stations, you probably can't do it all in one afternoon.

And if the normal outcomes are that 50% of the studies do not yield publishable results, then you are having to do twice as much work for 1 published real study than for 1 published faked one.

So yeah, I think laziness was a bigger driver than he admitted to.
posted by lollusc at 8:21 PM on April 28, 2013 [2 favorites]


Science used to be done by geeks. Then it became more respectable than labor, and it's been a race to the bottom ever since.
posted by Twang at 8:22 PM on April 28, 2013 [2 favorites]


If no one can replicate your data, it ceases to be relevant pretty quickly, whether you are being deliberately deceptive or just happen to be wrong for some reason.

I've known one PI whose research ethics I was not so sure about. Watching the way he handled data coming through his lab -- especially results from inexperienced researchers who didn't know enough to advocate for more measured interpretations of their experiments -- made me a little skeptical of his published results.

He's had some successes. But he's also seen quite a bit of controversy, and doesn't have much respect in his field -- an area of study that he pioneered -- mainly because several other labs haven't been able to reproduce some of his key experiments.

I've always found this experience kind of encouraging. The system working as it should, etc. Hopefully it's more the norm than the crazy situation profiled in this article. (I am fascinated by the bit about the graduate students never running their own experiments. Ten years' worth of graduate students! I wonder how many of them ever had suspicions.)
posted by gerstle at 8:36 PM on April 28, 2013


I think there ought to be a course in faking convincing data, or perhaps a program to generate it. You'd enter data about the number of the participants, ages, which gender(s) they were, and tell it the mean and variance of the data you want in each group. Then voila, you'd have a table that would be as statistically significant as you choose. That way referees would look for actual proof that the experiments took place, not just whether the numbers look right.
posted by Joe in Australia at 8:49 PM on April 28, 2013 [1 favorite]


Has anyone done any research into the falsifying of scientific studies?

Its a good thing that the FDA learned about potentially fraudulent work done on behalf of pharmaceutical companies by a contract research firm in Texas and didn't let it go to market.

Because that would be an effective Government and good science.
posted by rough ashlar at 9:03 PM on April 28, 2013 [2 favorites]


Twang - Science used to be done by geeks.

We sacrifice experimental rodents rather more humanely these days.

Rapidly saturating CO2 to render them painlessly unconscious prior to a C4 cervical dislocation.
posted by porpoise at 9:32 PM on April 28, 2013 [1 favorite]


wow, good use of the archaic meaning of "geek" there.
posted by en forme de poire at 9:52 PM on April 28, 2013 [2 favorites]


the vultures and fakers are almost always highly intelligent and they spend tons of energy on creating their fake science. Why don't they just do right? All of that intelligence and time and power could serve humanity just fine, if it were put into real research.

Faking data to get acclaim is a continual backrub from your peers and funders, while actual experimentation is more like a continual face punch by reality.
posted by benzenedream at 10:22 PM on April 28, 2013 [5 favorites]


For years (even before the advent of the aforementioned Retraction Watch), I have struggled to understand my own morbid fascination with NIH's reports of Scientific Misconduct. When I read these brief reports I am left wondering who these people are, and what drove them to jeopardize their careers just for a publication. I wish I could understand whatever it is that brings them to this because I truly believe that these people are not somehow just inherently evil fraudsters. I feel like if we could fix something about the culture that leads to these behaviors, we could somehow protect them from themselves.

This is why I thought the NYT piece was really enlightening--because even though it didn't really offer an explanation, it also didn't portray Stapel as just doing it to get ahead, get grants, or because he is arrogant. I got the impression that he is just extremely/pathologically misguided, feared failing, and fell down a rabbit hole where he essentially became addicted to his fraudulent behavior.
posted by gubenuj at 10:29 PM on April 28, 2013 [1 favorite]


mumimor: "Also, with a few exceptions, they are always caught. If they put their brilliant minds into actual research, they would have long careers, rather than a short(ish) sniff of glory and then an endless trip downhill. "

Hrm maybe. When a single scientist just blatantly makes shit up in an otherwise well-regulated field, they are soon caught. But there are also many fields of inquiry where the entire body of work (+/- some small kernel) consists of a merry but unfundamented game of intellectual follow-the-leader. Upthread, lambdaphage pointed to such a criticism of social psychology; it's also true of (parts of) linguistics, political science, and other disciplines where the theoretical component outstrips the data collection possibilities.

In environments like those, it's possible for a dedicated group of people to have long, successful careers (unto several successive generations of graduate students, even) investigating useless things. Of course, these people are often earnest believers in the tenets of their discipline, and it's not at all clear that these flights of fancy aren't necessary detours along the not-always-linear path of science. So, from the standpoint of internal motivations Stapel is a nasty atypicality, but from the point of view of "wasting a lot of time and money for eventually not a lot of results" it's hard to differentiate him from many other scientists. (This is one of the reasons why it is so hard to get public support for science, and Stapel is totally an asshole for making that harder still.) However, the real scientists vs. frauds, vultures and networkers dichotomy is false -- there is a continuum, and we're all on it somewhere.

Tightening the purse strings to choke off the freeriders isn't necessarily the way forward; we need to encourage values of reproducibility and sharing of primary data, which in so many fields is lacking. A tough funding climate tends to discourage people to be accepting of criticism of their work. When it's no longer a matter of putting food (or health insurance) on the table for science workers, it's going to be easier to have a marketplace of ideas and not of egos.
posted by dendrochronologizer at 10:32 PM on April 28, 2013 [1 favorite]


Then you don't understand scientific methodology, because untestable science isn't science, it's Cargo Cult Science.
also, correlation is not causation
posted by This, of course, alludes to you at 11:04 PM on April 28, 2013


Slackermagee, I'm not sure I get it? Surely if I'm contemplating cropping my gel, I've already made a decision about my area of focus, no? Would more courses in Brit Lit have steered me away from perfidy?

I don't think I meant STEM, sorry, I think it was that focus on commercialization of research. If it isn't going to market, it isn't going to get a grant.
posted by Slackermagee at 11:51 PM on April 28, 2013


Then you don't understand scientific methodology, because untestable science isn't science, it's Cargo Cult Science.

"This isn't falsifiable" is like "correlation is not causation" in that it is parroted by people on forums as if it were some sort of cornerstone of science; it usually exhausts the person's knowledge of the topic; it is usually nothing but an euphemism for "I don't like this"; and actual research scientists don't care a jot about it, because actual science in the making does not depend on knowing a couple of catchphrases.

So yes, also correlation is not causation.
posted by Pyrogenesis at 1:01 AM on April 29, 2013 [3 favorites]


On the data ownership question raised above - in addition to university policy, that's going to be an issue of the individual funding sources. Certain funding sources have certain requirements on who owns that data and how it can be used. And then beyond their general policies, to an extent, some of that is negotiated contract by contract. (I don't work directly with that process - my sense is that data ownership is an area where there's significantly less wiggle room than some other parts of the contract negotiations, but I may be wrong.)

As big funding sources make more of a commitment to requiring open access to data and analysis, universities will be forced to agree to more open terms if they want a piece of the funding pie.

Of course, many researchers have no idea what the data ownership and access terms of their particular grants are, and will share (or withhold) data as they please regardless of the legal requirements/responsibilities. But pressure can be put on by funding sources, and better education efforts made by universities, to encourage open access to data.
posted by Stacey at 7:59 AM on April 29, 2013 [2 favorites]


"This isn't falsifiable" is like "correlation is not causation" in that it is parroted by people on forums as if it were some sort of cornerstone of science

I don't know about catchphrases, but if it isn't falsifiable, it isn't science. So you heard it here from someone who does it for a living: whoever says that catchphrase is 100% correct about science and how it works.
posted by Blazecock Pileon at 8:51 AM on April 29, 2013 [2 favorites]


Tightening the purse strings to choke off the freeriders isn't necessarily the way forward; we need to encourage values of reproducibility and sharing of primary data, which in so many fields is lacking. A tough funding climate tends to discourage people to be accepting of criticism of their work. When it's no longer a matter of putting food (or health insurance) on the table for science workers, it's going to be easier to have a marketplace of ideas and not of egos.

dendrochronologizer, I agree with this, and most of what else you are saying. My experience is that most of the "vultures" have turned up in the over-funded fields because their own field was starved. And in a sense, I am where I am now because there was no room for me in my original field of research (though the basic theories and methodologies are identical in my case). The difference is, in my view, that some people will compensate for their lack of knowledge and experience in a given field by creating a parallel universe, where black is white and rats are cats. This can be done in overfunded fields because there are enough under-qualified people to staff journals and organize conferences independently of the core centers of research, and there are enough politicians with an agenda to fund them if their results match ideology.
posted by mumimor at 1:52 PM on April 29, 2013 [1 favorite]


I don't know about catchphrases, but if it isn't falsifiable, it isn't science. So you heard it here from someone who does it for a living: whoever says that catchphrase is 100% correct about science and how it works.

Much like with the correlation-causation thing, I'm in complete agreement. It is entirely, completely, 100% correct.

Being capable of declaring that just like everyone else who has declared that same thing does not make one wise in the ways of science, is all I'm saying.
posted by Pyrogenesis at 3:09 PM on April 29, 2013


SOPA creator’s latest bill proposes stripping peer-review from science funding
posted by rough ashlar at 5:46 PM on April 29, 2013 [1 favorite]


That link to the SOPA story has a lovely piece of stock photography with the scientist touching her face with her gloved hands. And we won't mention all the pretty colored liquids in various pieces of glassware including a graduated cylinder filled above its maximum measured volume.
posted by sciencegeek at 7:05 PM on April 29, 2013


Disputed Results a Fresh Blow for Social Psychology. Has a bit about Stapel, but the article is about a different social psychologist.
posted by logicpunk at 2:46 PM on April 30, 2013


Re: Stapel and addiction. If data faking is an addiction, the word has no meaning, as I explain in an article I wrote here.

Addiction cannot be an all purpose excuse for anything you do wrong to get what you want: if that's what it is, it's certainly not a disease or disorder and certainly not something we should be wasting money treating. Of course, that's not actually what it is.
posted by Maias at 3:19 PM on April 30, 2013


Ugh. Correlation is not causation, but correlation does not exclude causation. Causal events are typically correlated.

Also, a lot of bioinformaticians may have gripes about the "not falsifiable, not science" assertions. Hypothesis generation through data mining is still definitely a part of science.

Or even less weak - space exploration through looking at the visible universe at different ranges of spectra of electromagnetic radiation or sending probes out and collect and analyze samples. No hypothesis needed. Nothing falsifiable. Yet, that's still "Science!" Right?

I was chatting with a friend's highschool kid while she was doing her science 10 homework. One of the worksheet questions was about "what is the scientific method." There is no single "scientific method." There are "methods OF science." But I highly doubt that her teacher would understand, much less care. Her dad overheard our conversation and gave a small cow - I'm still not sure whether he agreed with me and wanted his girl to "give the right answer" or if he actually bought into the simplification that everyone's taught.

Still, there is definitely Bad Science and Fraudulent Science as well as awful (mis)interpretation of science.

That link to the SOPA story has a lovely piece of stock photography with the scientist touching her face with her gloved hands.

I'm surprised that the photog didn't make her wear a visible vision correction device.

Also, wearing PPE (personal protective equipment) is a bit of a misnomer in some fields. As a molecular biologist, when I had full control of who/what is in contact with my "stuff" (tools, reagents and their containers, equipment) I generally only wore gloves/masks to protect my samples from me (and the bacteria, nucleases, and other immunological biologicals that I shed as a human) rather than the other way around.

In my current lab, I'm almost tempted to strip, bag my clothes, wash down with disposable wipes, and change into clean plastic-bagged clothes after I leave the building and before I get into my car. I'm absolutely appalled with the general practice at my new place of work.

It isn't even a rich lab/poor lab thing (though it's a factor), it's a culture thing. I made the undergrads and my coworkers respect the space and equipment that I used if they wanted to share it. Mostly because if they used the stuff that I maintained and used, their stuff worked. Based on stuff working, they adopted my practices. Life and science is good. New lab; less-motivated undergrads, techs who don't care, management who wants to do it on the cheap, and, well, culture.

I can still get my stuff "to work" but I'm latexed up 99% of the time and keep my forearms/elbows off of any surface.
posted by porpoise at 9:32 PM on April 30, 2013


US Scientists Significantly More Likely to Publish Fake Research
posted by jeffburdges at 8:12 AM on May 1, 2013 [1 favorite]


Also, a lot of bioinformaticians may have gripes about the "not falsifiable, not science" assertions. Hypothesis generation through data mining is still definitely a part of science.

I appreciate the shout-out but I think this is still in the realm of "falsifiable." If you're calling a statement a hypothesis, that generally means it can at least theoretically be tested, and thus potentially falsified, even if the current tech isn't quite up to the measurements you would need (e.g. making claims about 1000s of genes/proteins). Even in the cases where we can't do a direct, definitive test (e.g. when making a claim about history) you can still look for support/contradiction of that hypothesis by looking for other things that it would have necessarily entailed, and testing whether those are true. In that case the enterprise is more like model building, but it's still "falsifiable," I think - you could always find new evidence later that proves your original model was inaccurate, but that doesn't mean it has to be airtight today as long as it makes useful and accurate predictions.
posted by en forme de poire at 1:01 PM on May 1, 2013


Falsifiability is somewhat like Russel's set of all sets which are not members of themselves, which as you may recall, resulted in a logical contradiction because if it was a member of itself then it could not be a member of itself, yet if it was not a member of itself, then it must be a member of itself.

If falsifiability is falsifiable, then there must exist a scientific theory-- that is, a theory which deserves to be called scientific, not necessarily a true theory (otherwise you'd end up asserting that Newtion's gravitation is not a scientific theory, for example)-- which cannot be tested by any evidence now or in the future.

But the very existence of such a theory contradicts falsfiability, since fasifiability asserts that such a theory cannot exist.

Therefore the original premise, that falsifiability is falsifiable, leads inevitably to the conclusion that falsifiability is wrong.

But in that case, falsifiability must not be falsifiable, and we have then exhibited a theory which is not falsifiable, namely falsifiability itself.

Therefore falsifiability is wrong.

That means that assuming falsifiability is falsifiable or not falsifiable each lead to the conclusion that falsifiability is wrong.

The only way I see to save it is to assert that falsifiability is not a theory at all, can never be investigated by science, and has either the status of a priori truth, which is highly dubious, or is a mere rule of thumb like Occam's razor.
posted by jamjam at 3:04 PM on May 1, 2013


Whether you're wearing the gloves to protect you or your samples, touching your face with gloves on isn't going to protect your proteins from keratin contamination or your face from ethidium bromide.

I bet she's wearing contacts.
posted by sciencegeek at 7:15 PM on May 1, 2013


The Republican War on Social Science: They’re winning it.
posted by homunculus at 7:26 PM on May 1, 2013


« Older CRAPCHA...  |  Very obscure and super rare se... Newer »


This thread has been archived and is closed to new comments