I’ve been arguing with myself again. I really liked that phony chocolate study because it so effectively demonstrated a couple of problems I tell my students about, so it’s a spectacular way to illustrate p-hacking and the unreliability of peer review. But as I was thinking about it, and how to present it to a class, it started to sink in that it also raises brand new problems that make it very difficult to use as an example. And then I started reading some other articles that emphasize the ethical concerns in this study.
Here’s Rachel Ehrenberg:
Bohannon and his colleagues decided to create a wrong to prove that wrongs exist. They lied to the public to make their point. Granted, it’s unlikely that anyone will be harmed by eating more dark chocolate. But not only does the caper do a disservice to people who are desperate for meaningful information about health and nutrition, it also undermines all of science and all of journalism. There’s real wrongdoing in both science and journalism (most infamously, see Stephen Glass, Jayson Blair, Janet Cooke, Jonah Lehrer, Brian Williams). But intentionally creating wrong to make a point is both bizarre and potentially very damaging.
“Our key resource as journalists is credibility,” Edmonds told me. “And a deceptive ploy like this could damage that.”
“Don’t lie” is a rather fundamental principle in science.
Another ethical principle is to do no harm, as Chris Lee of Ars Technica points out.
The end of the experiment is that millions of people all over the world were told that chocolate will help them lose weight. The consequence is that all those people who search (in vain) for fad diets—often to help them with their self-image—have been given yet another false data point and another failure to reflect upon.
In terms of ethical analysis, this is an experiment that did not tell us anything that wasn’t known already. On that score alone, the experiment fails to pass muster. Then there are the downsides. The reputation of science journals and science communicators just got a slight additional tarnish. Worse yet, there are people out there who have been taken in by the false reporting, and many of them will never know that the story was false from the beginning.
So if I were to present Bohannon’s fake chocolate study to a class, I’d have to say something like, “Here’s a really vivid example of using misleading statistics and getting the work published and well known. Don’t you ever, ever, ever do this.” That’s a problem.
But like I said, I argued with myself. It’s such a strong example by design — can I salvage it for instructional purposes?
I don’t think I can. What settled it for me is the personal perspective. I have a lab! I work with undergraduates, and I encourage them to try out new ideas. That means I have lots and lots of what I would charitably call ‘preliminary data’ — less charitably, you might call it half-assed data — that needs a lot more work to be useful. Bigger sample sizes, correcting for known confounding variables, just general replication…getting a student to focus for a year on something in that mess and make it a bit more robust is necessary before I could say it was really publishable.
But hey, what if I were to fish around in the data and look for something potentially cool and just fiddle with the stats and publish it in some low quality journal? Especially…what if I were to tell myself I was just doing it to highlight a problem in in science reporting? Wouldn’t that be a service?
And I find myself actively repulsed by the very idea. I am far too conscious of the shortcomings of that data — it’s tantalizing enough that it would be productive to put a student to work on the details, but it could be a dead end, or it could be accounted for by a trivial variable, or it could just be completely wrong. I’m a little surprised at how queasy even the thought of trying something like that made me feel, no matter what the excuse. I could not have assisted in carrying out Bohannon’s study, for instance — I would have begged off on personal ethical concerns.
When I put it in those terms, whether this was a study I could have participated in in good conscience, it suddenly makes it extremely problematic to present it to students, no matter how illuminating the results are. Science is all about the process, not the answer, and corrupting the process further, even to demonstrate the existence of extant corruption, is a violation. I tell students over and over that getting a pre-determined answer in lab isn’t the point — it’s how you get that answer that matters most. It’s a tough lesson to teach. I’m always getting distressed students who think they’re going to fail a course because their experiment didn’t work, and I have to tell them I don’t care that it didn’t work — do you understand why it failed? Do you have ideas for how to correct the problems? Did you learn something? Are you willing to accurately report what went wrong? That’s what I care about.
So how can I tell them that in this case, the end justifies the means?
And once a fake study like this is out there, it’s out there forever. People don’t get their information from medical journal, but from news articles that write about them. Even today, I just searched for “dark chocolate weight loss” and an article touting this study was the third hit. No mention of the retraction.
When discussing safety in training sessions, the things people remember the most are tales of people who did it wrong and got hurt as a result.
When discussing ethics, I think it makes sense to do the same? Pond and Fleishman with cold fusion? Whatshishame with the vaccines? But those people were hurt professionally, which makes them a cautionary tale.
I don’t know – I see the dilemna.
I agree, it’s quite a conundrum: how to teach from this deliberately bad example (errr ‘good example of bad’).
My approach (FWIW) is to not ever say “This was the wrong way to do this, don’t ever do any of yours this way.” but to approach the same message obliquely. As in, “The authors of this paper claim this was a deliberately fraudulent paper; so tell me what mistakes they performed on their fraud stat analysis.” (and if/when they sputter, start showing each error and its implications…)
My learning always seemed to come best from the teachers who asked, rather than told. Sure, they would also demonstrate examples of solving related problems, but then follow with, “how do you do _this_ problem??” Of course the asking I got from those best teachers was not just a request to solve a homework problem, but asking for explanations of how one tried to solve the problem.
…
ugghh, not tryin to tell PZ how to do his very job that he’s been doin very well for very long. just sharin my approach to the OP’s “issue”. FWIW, FWIW, YMMV, etc, disclaimers, etc, etc.
Speaking from the viewpoint of a non-scientist, the phony chocolate study does serve as a useful reminder that scientists are human beings, so may very well be unscrupulous, and may also make mistakes. This means that a study be be deliberately faked and/or simply wrong, but might not be noticed as such. Having said that, one thing that did stand out for me is that, even though these folks were using their scientific knowledge to try to expertly craft a fake study and get it published, they knew they couldn’t risk peer review, as that would likely expose it, so, again, speaking as a non-scientist, this tells me that peer review is actually very good at weeding out faked and/or sloppy science (assuming, of course, it’s not the kind of ‘peer review’ you seem to get at many creationist ‘science journals’, where they simply rubber-stamp anything that seems to support creationism), and that the best people to police scientists’ work for accidental or deliberate wrongdoing are other scientists and the science community.
However, I do agree that there is something of a big problem with the fact that some people might still believe the original study, but not be aware of the ‘it’s fake’ revelation. The only thing that can be done about that is for the original authors to try everything they can to make sure the announcement of fakery gets spread as widely as the original study.
I have to agree that allowing the work to be published, knowing that it is both wrong and at least somewhat plausible, is unethical. (I doubt many skeptics have large problems with Alan Sokal’s famous hoax in Social Text.)
I faintly recall other reports of fake studies used to test peer review in epidemiology, but these were always withdrawn before publication, the results of submission to many journals collected, and then the results of that published with (IIRC) the targeted journals remaining anonymous.
A paper in the BMJ addressed the problems of chance findings and inappropriate subgroup analyses with explicitly simulated trials using different coloured dice as the ‘treatments’. I recall that well enough to find a reference http://www.bmj.com/content/309/6970/1677. (Declaration of interest – one of the authors is a colleague of mine, so I saw this material in a poster on the wall.)
The first objection,
carries little weight. The fake study showed that the popular press has little credibility when it comes to science reporting. I think pointing out that the king has no clothes is a public good, even if if the king doesn’t like being revealed as a poseur.
On the other points: sure, tricking people in to publishing misleading data in and of itself harms the public’s perception of science (those damn scientists keep changing their minds!), but if the effect is to make journalists more circumspect about what they publish and how they present it, the net effect would be to bolster the reputation of science since there would be less hype: “Darwin’s missing link found” (again, and again, and again).
The REAL downside of publishing this fraud (IMO), even including the retraction, and the explanation that is was deliberately done to show how easy it is to make mistakes or “massage” one’s data for fraudulent results. yada yada yada, the effect I’m leading into is that it fuels the “layman’s” mistrust of science results in general. I.E. yet another tinder to add to the mantra: “Yesterday they said eggs are bad, today they say they’re ‘perfect food’. Can’ trust those sciencey know-it-alls…” In short, while trying to make science studies better, it could be, metaphorically, shooting itself in the foot.
That’s one more aspect of this paper PZ needs to include in his teaching of the ethics of this as a bad paper. unintended consequences et al.
oooooo, jaybee@6 beat me@7 to it. shame on me
This is a surprisingly dogmatic approach to morality. And I think it also exhibits confusion about what the authors of the study were trying to demonstrate. If they were really interested in studying the effects of chocolate, then yes, their methods were corrupt, and a violation of scientific process. But this was not their goal, if I’m understanding things correctly. Their goal was to demonstrate the hypothesis that it is easy to have a piece of junk research, with multiple serious methodological flaws, accepted into the scientific literature.
I believe their methods were quite adequate to achieve this goal. And I believe this is a valuable contribution to the advancement of science. By making this information available, we can more readily do what is needed to repair things.
To claim that you can’t behave that way, because it violates the rules is to forget what the rules are for, and to treat the rules as the most important thing. In reality, improving our mutual situation is the most important thing. Rules just serve as handy mnemonics for remembering how to do that.
@6 Jaybee — yeah, I think this should get taught in a journalism class as much as in a science ethics class. The editorial and fact-checking functions have largely disappeared in our mainstream media, even in The Grey Lady, since she was sold to the world’s biggest arms dealer and now relentlessly cheerleads for war. All four major networks are of course pathetic. And in the blog-o-sphere, there’s no editing or fact-checking whatsoever. I continually cringe, even among my very favorite bloggers, at what passes for English and for verifiable truths. “Because, reasons”, has become the go-to catchphrase for every blogger or journalist too lazy to actually examine what the person they’re trashing actually said, and to explain their objections in clear prose. This sort of sloppy writing wouldn’t have passed muster in a 10th grade composition class, but is now ubiquitous. “Wevs” is another, shorthand for “whatever”, which is the most useless bit of brain-farting of the past generation, saying nothing, and wasting those four letters of idiocy. Whenever I see “wevs” or “because, reasons”, I know that an editor should have gob-smacked the writer and told them to re-write the paragraph as a grown-up. This chocolate canard is no different. Blogs and newspapers too cheap and lazy to provide an editorial and fact-checking function simply didn’t. The fault, to me, lies in the journalism, not in the kids having a lark.
I almost wonder if this isn’t instructional so much for students as it is for working scientists – and, more importantly, the mechanisms of science journal control.
I think it is, really, a wake-up call to ensure that the governing bodies tighten up their standards.
Thank you for this. I was feeling pretty pissed off about that study, as what they did was to use the stature of scientists to basically sandbag the journalists just to then make fun of them. It is good to see that many came out feeling the same way.
To expand on this a little, maybe this would be a good discussion in an upper level course in scientific ethics. Combine it with Henrietta Lacks and look at the way scientists either don’t consider or don’t care about the people or the consequences of experiments. Explain how and why what happened with Henrietta Lacks was wrong, even though she personally was not harmed. Explain how the chocolate scientists were trying to point out bad science and bad science reporting, but have in fact caused harm in various ways. You still get to use a good story, but for a different purpose. And it can also reinforce some of the problems with statistics and bad science journals.
FWIW, many decades ago when I was taking a German folklore class, a junior level class, we spent a day discussing the controversy surrounding the use of folk tales, and the ethnographic information of the teller, gathered by Nazis. Even though they were gathered in the same way as any other folklorist would have done it, the fact that they were gathered by Nazis for potential use as propaganda in the War and to further their belief in Teutonic superiority made them ethically tainted. If I remember correctly, people did use the information in their articles and research, but were up front and clear about the source of the material. The fact that the research was done in the same manner as other folklorists would have done helped.
This is a most telling admission in my view. It is essentially a complaint that people shouldn’t try to catch science journalism being stupid because it is already known to be stupid so hush up and don’t point it out.
Deservedly so. The only potentially unfair thing here is that those who didn’t get caught out, i.e. those journals that refused the submission or were simply uninvolved might be tarred by the same brush in the eyes of some, but if it leads to increased scepticism amongst science journalists, I remain in favour of it.
It may be bad science, but I don’t think it is bad journalism. I like to think of it as an undercover operation. A journalist has to lie about who he is and what he does to infiltrate in a suspect organization. In the end the world will be wiser.
Take this in consideration: when you point out an everyday flaw in science journalism, some will say “duh” and some will not care, but few will really take it as an example. But when someone puts up a prank like Bohannon did, some – if not most – will say “gee, see how silly those science journalists are, I’m happy I’m not like them”. The impact, and with it the educational value, may be much higher.
So, with some precaution, in this case I think the ends did justify the means.
Disclaimer: I’m not a scientist nor a journalist. I am a bit annoyed about lazy journalism in general.
I still think there’s some serious power in actually showing the dysfunction of science journals and possible scientific writing by example, versus just telling people “Hey there’s a problem here”. It’s much harder for the folks with egg on their face to rationalize that criticism away or ignore it.
So I’ll have to side with Jaybee on this one:
Sorry, to add-
That Chris Lee editorial was a good exercise in blame-shifting, by the way. Just look at this:
It was just so tantalizing! We couldn’t help but report on this bad bit of science from a journal that we knew doesn’t do peer review!
In response to a post above, I did have problems with the Sokal hoax. I think I’m coming around on it, but it still unethical to submit an article that you know to be completely false.
While there are a number of ethical problems with the “diet study”, this comes back to the point that students need to be educated about these kinds of studies. Not just the “don’t do this in your lab” lesson, which needs to be put out there, but that you should not blindly trust what you see in journal articles. This was obviously not peer reviewed, but even if it were, the peer review process can miss things when done with diligence – and ofttimes peer review is not as rigorous as it should be.
anbheal @ # 30: … our mainstream media, even in The Grey Lady, since she was sold to the world’s biggest arms dealer …
Say what? “Grey Lady” usually refers to The New York Times, which continues to operate as its own corporation.
“World’s biggest arms dealer” – most often signifies the United States collectively. On a corporate level, Lockheed-Martin claims that title, but doesn’t even list anything to do with media per se on the “What We Do” section of its corporate web site (link omitted to evade comment moderation limbo).
So – whatchoo talkin’ ’bout?
This isn’t about pointing out how easily junk science can be published. Johannes Bohannon had successfully demonstrated that last year, for several open access journals, and this was a factor that made him and his collaborators think to take it to the next level in the chain of information. Once something erroneous is published in a disreputable journal, do journalists differentiate between it and real research? Do they even read it in any critical way, and ask the authors about details they don’t understand? The finding was that, in a lot of cases, they don’t. Science journalists were the target, not science itself, and if they were caught with their pants down they deserve to be exposed for it, to hopefully shake them out of complacency.
Oh, and just because nobody else has said it yet: it’s all about ethics in sciencing journalism.
Yes, there are important ethical issues here. But (a) that isn’t much different from what Sokal did, and he is still acclaimed as some sort of hero and (b) unfortunately, I don’t think journalism, especially scientific journalism, has much credibility left…
@#23, kencamargo:
No, you’re wrong there. The whole point of having peer-reviewed journals in the sciences is that they are supposed to maintain a high degree of skepticism about articles and presumably achieve a greater level of accuracy. You might as well not bother publishing at all if this responsibility is abrogated.
What Sokol did differs from this because of both the subject matter and the nature of the journal to which he submitted. Social Text is a journal for highly technical postmodern culture studies. The audience is composed almost entirely of academics working in literature, and it is understood that literary criticism is not absolute — what may seem true and obvious today may seem ridiculously antiquated in ten years’ time. There might be a profound long-term negative effect on society in the long term if Social Text publishes an article which is admittedly nonsense, but the nature of the material is such that any effects would essentially have to be long-term — it’s implausible that anyone will die from postmodern cultural studies.
This, on the other hand, is a much more serious ethical problem — there are, right now, even after the retraction, people who are eating (or buying) chocolate as a result of this study, who will make their health worse in an immediate sense, possibly in ways which are concrete and directly measurable. (It is not at all implausible that there’s a diabetic or ten out there saying “well, the sugar in this is bad for me, but the science says that chocolate is good for me… better keep eating!”)
Note also the difference in representation: if Sokol had not announced his hoax, practically nobody would ever even have heard of his article or the conclusions presented therein. This fake article, on the other hand, received media attention immediately, without any mention of the hoax. Science news carries more weight, and therefore the ethical problems should be taken more seriously.
(Which is as much as to say: nobody cares if you cheat at a game of solitaire, but if you’re playing in the poker world championship you’d better be playing straight.)
So what is an appropriate way to test and improve the methods of science reporting?
@#25, DanDare:
Well, let’s see:
1. You’re saying “test”. If you let false results go all the way to press, you aren’t “testing for bad publishing practices”, you’re “joining in bad publishing practices”. As others have noted above, printing a retraction is nowhere near as effective as not publishing false results in the first place. So, at a very minimum, this sort of plan should include a mechanism for halting publication before, well, actual publication, or else it’s in the “more harm than good” area.
2. For that matter, this ethical aspect of this whole mess could have been avoided by choosing a different “treatment” to fake into legitimacy. If they had chosen a substance which is not available to the public, or which doesn’t even exist (but sounds plausible), they could have released a study which would have been harmless. If you’re already publishing plausible lies, does it matter whether the lies are plausible all the way down?
3. At this point, there isn’t any reason to test. We already know that science journalism is bad. We know that popular press accounts are wildly inaccurate and often flat-out wrong, that non-peer-reviewed journals are effectively rubbish, and that peer-reviewed journals are not as rigorous as, ideally, one would desire them to be. And obviously, the bad publications are happening because there are incentives to publish them. Probably the best bet would be to establish disincentives — say, for example, somehow refusing grant money to people who publish results in non-peer-reviewed journals, or something like that. You would expect this to already be happening — but it may not be. (You would expect that hospitals would communicate with state medical licensing bodies when they discontinue a surgeon’s ability to operate because of malpractice suits, but they don’t always do that, in order to prevent surgeons with a track record of malpractice from jumping from one hospital to another; at least into the 1990s, my state had no requirement for that and may still not have one — I don’t know. My point is that things which seem like common-sense regulation for the prevention of ethical violations are sometimes not actually in place.)
In addition to the value of teaching this in a science ethics course, as mentioned above, I just thought of another useful application. Do any courses teach about the peer review process? I feel that it might be a productive exercise to have students take on the role of peer reviewers and give them a handful of studies to them for review. This study did not have to actually be published for that, but it might give the lesson a bit of extra oomph if the students are later informed that the study made it through.
Regarding the ethics of going as far as they did, I think it was valuable to do it once. Yes, it is already known that bad studies get published, but I doubt the authors of those studies would generally admit that their studies are bad. E.g. Wakefield’s Vaccine / Autism study was even worse in many ways (smaller sample size, fudged numbers, personal reports used when medical records were available, etc.) but he still maintains that he was right.
As multiple people have already noted, the credibility of science journalists is often unwarranted. It’s just unfortunate that good journalists get tarred with the same brush. This is compounded by the unfortunate practice of having headlines written by someone with even less understanding than the journalist. Maybe science journalists need some sort of professional association to police their work.
Regarding the harm of having people running out to buy chocolate because of this study, it’s a drop in the bucket compared to the garbage that people ‘learn’ from sources like Dr. Oz. I just looked and he has frigging juice cleanses on his site, which are likely much worse than eating some extra dark chocolate. Given the money to be make by getting someone like Oz to promote something, I have a feeling fake studies like this are rather plentiful, but created for more nefarious reasons.
I do still think that the people involved in this ruse need to be diligent in chasing down any article they can find online and hounding the authors/publishers until the page is updated or removed.
In summary, the ethics of publishing and promoting this study are quite grey, but I think it was worth doing. Science journalism needs to improve and pay-to-publish journals need to be publicly discredited (or pressured into beefing up their procedures).
Somewhat related question, how many people here have heard of Quirks & Quarks?
It’s a weekly Canadian science radio show (available via podcast) where they interview scientists from around the world about their research. I know it’s broadcast some places in the States. It’s still not perfect (occasional overselling of findings and scientists who aren’t clear about which conclusions are guesses/assumptions), but it removes a step in the telephone game.
I like to mix it in with my religion podcasts because I find it generally less depressing. Even when the subject matter is horrifying (climate change, extinctions, etc.), at least you’re hearing from the people who are doing something about it. There’s also a lot of uplifting content.
Golgafrinchan Captain
I am Canadian.
I ♥ Quirks and Quarks for a looooong time.
I see the ethical concerns, but I think that, at least in this case, the benefits far outweight the harm.
Also, while the authors have definitely a responsibility, the big burden is not on them: it’s on the media who published the study without checking and failed to won the mistake.
How many fashion magazines are going to write “sorry, we incorrectly wrote X but turns out that the study was a fake and we didn’t check. We’ll be more diligent next time”?
Bohannon slapped in our faces the horribly flawed relation that our society has with reality.
We desperately need more of this.
We get this idea that the world we live in is made to be easy to understand for us and none ever ever ever lies to us.
IT IS WRONG.
The faster we go past this attitude, the better humanity will do.
How many of those that believe the study to be true will forget about this study and happily jump on the next fad anyway?
But also, how many will learn how easy is to be deceived and become wiser of it?
I certainly benefitted from the study. My partner too, and I’m willing to think that many, many others did as well.
I completely disagree with Ehrenberg: this was a great service to people who are desperate for meaningful information.
Also, PZ, I don’t understand your personal problem.
Yeah, you could go around fishing data and publishing shit papers.
But are you going to tell the world afterwards?
Are you going to tarnish your own credentials as a scientist?
Bohannon can do it, because he’s a *journalist*.
But you are not a journalist. You wouldn’t get a free pass on this.
@ chigau
I listened to it occasionally when I was a kid and then forgot about it for a long time. Now, I haven’t missed an episode since I got an iPod in 2008’ish. I also used to watch Bob McDonald on Wonderstruck.
Oh yeah, link: cbc.ca/radio/quirks.
How is this different than any other “Devil’s Advocate” situation?
I think that’s one of the hardest, but most important, things to “get” about science; the difference between testing a hypothesis, and attempting to prove a hypothesis. A good experiment does the former, not the latter.
How does everyone feel about the Sokol hoax? One could argue that no one is suddenly going to embrace the excesses of post-modernism based on that one paper, but conceptually is similar in that it was intended to point out misuse of science and problems with peer review.
Not that last thing – Bohannon et al. deliberately avoided peer review by sending their manuscript to a scam journal that doesn’t do any; see comments 4 and 19.
It would be quite redundant so shortly after Bohannon et al. did it.
. , quotation marks not intended.
It’s a category mistake to confuse a conservative writing style with accurate content, or lack of one with lack of the other.
It’s in fact a pretty dangerous category mistake. Take care that you don’t fall for well-written falsehoods.
Conversely, many actual science journals aren’t copyedited anymore. Even the most prestigious one, Nature, has been merrily publishing typos and other spelling mistakes for many years (I distinctly remember a tuberocity in a very important paper from 2000).
Deservedly so:
Lots and lots of your colleagues did report on this bad bit of science without so much as taking one look at the journal.
Most science journalists – worldwide; Americans, don’t make this into “woe is unto our country” – have no idea what they’re doing. In short, comment 21 (and 22).
The real blame, however, falls on their employers. Not only do they assign the task of reporting on science to people who can’t recognize science when it stares them in the face, but they assign the task of reporting on all of science to a single person. Nobody can understand all of science well enough to competently report on all of it.
You can put 6 links into a single comment without triggering moderation. :-)
You know that, and I know that, but the vast majority of people out there don’t; they’ve mostly never heard of peer review, and most of those who have have no idea what it is; they trust in science journalism because they don’t know any better.
Oh yes, this, so much.
Good idea!
Absolutely. As of a few days ago they still hadn’t updated the website of their fake institute.
Golgafrinchan Captain @28: I stopped listening soon after Jay Ingram left. McDonald’s a bit too gosh-golly for me.