1/3rd of Scientists admit to Research Violations

[Durandal]

It sounds like the study included social "scientists" as well as real ones. In the social sciences, it's far easier to bullshit data and because of the huge sources of error, much easier to get away with it during the process of peer review. I'd bet that the percentage of dishonesty goes down when you take sociologists and psychologists out.

As likely as it seems, that is stated nowhere in the article. Wishful thinking and implication aren't the same thing.

They mentioned that some scientists ignored some rules for protecting human subjects. So that implies that some scientists in this study worked with human subjects, which would be medical researchers or social science researchers.

[tharkûn]

I would like to know how many of these researchers who skew study results just happen to work in medical research. No offense to medical researchers out there, but it's not exactly a secret that the drug companies do not pay you to make their products look bad. Hence, for example, it took so long to admit that Tylenol could cause liver damage that it was virtually common knowledge by the time we finally got a study confirming it.

I have no idea about what percentage of those people happen to be in medicine, but our lab has been paid, and paid well to say that one project from pharma was utter crap. There are some scientists and labs known for skewed results, but most people know that and serious pharma avoids them like the plague. These days the costs of the inevitable lawsuits grossly exceed any possible profit you could make by covering up side effects until they can't be ignored.

Even if the drug companies were amoral greedy bastards, the cost/benifit ratio these days isn't forgiving of bad science - eventually it WILL bite you in the ass.

I didn't know that! What pain relievers should I take then?

Whatever one your doctor says is best for you. Virtually all drugs cause liver damage to some degree, notice how just about every drug advertisement has liver problem contradictions. The reason Tylenol is more pronounced than other drugs is that it is so commonly used and abused; to a lesser extent long term alcohol abuse increases improper breakdown of the drug and liver damage.

Tylenol has side effects in the Liver, traditional NSAIDs ulcerate the stomach, COX-IIs cause cardio problems, etc. As long as you are taking more than 2 grams per day, drink regularly/heavily, or have other problems I'd go with Tylenol as being the safest. The majority of diagnosised acetaminophen induced liver damage comes from intentional overdoses, of the remainder you have unintentional overdoses, morons who down half a bottle of Jack Daniels with a gram or two of acetaminophen, and hepatitis patients. Frankly Tylenol's health record would be ludicriously better if it weren't the drug of choice for attempted suicide.

[Joe]

I completely believe these numbers, they're perfectly in line with what I'd expect given the prevalence of fraud in other professions. Though to be fair, I imagine scientists would probably come out favorably in comparison, since 33% is certainly not as big a number as it could be.

This study, as I see it, is not really a unique indictment of scientists, it's an indictment of people in general. The fact is, a substantial chunk of the population is going to be willing to lie and cheat, though in varying degrees, to get ahead, and we shouldn't be surprised that scientists are human too. That doesn't excuse the fraud, but it does put it in perspective.

[DPDarkPrimus]

Two words to smack down the idiots who argue this means science is bullshit:

PEER REVIEW.

Yes, that's right, peer review. If a bunch of other scientists run the same tests that the forgerer supposedly did, and come up with different results, guess what? The fraud is exposed.

[tharkûn]

Yes, that's right, peer review. If a bunch of other scientists run the same tests that the forgerer supposedly did, and come up with different results, guess what? The fraud is exposed.

The problem is a lot of crap is never double checked. In theory all science should be duplicated by another researcher to do exactly what you propose, however no one gives out funding for that except for the most critical studies. If you are studying something which few people care about or something which has gaping holes for error to creep in it is entirely possible for a double checking test to never be run. In some cases you KNOW only the most cusory peer review occurred, for instance I once worked through the geometry of a crystal lattice and found a value that was physically impossible - no one ever did a second X-ray diffraction and none of the peer reviewers worked the geometry.


When it is IMPORTANT the fraud will be exposed by peer review, when it is shlock throughput in some journal whose sole reason for being is to provide a publication niche and obscure reference it most often won't.

[wolveraptor]

Does the article mention how many of these frauds actually get through the peer-review process? How much of this bullshit is stopped in its tracks before it does real harm?

[Il Saggiatore]

It sounds like the study included social "scientists" as well as real ones. In the social sciences, it's far easier to bullshit data and because of the huge sources of error, much easier to get away with it during the process of peer review. I'd bet that the percentage of dishonesty goes down when you take sociologists and psychologists out.

As likely as it seems, that is stated nowhere in the article. Wishful thinking and implication aren't the same thing.

They mentioned researchers having improper relationship with test subjects.

If I had improper relationships with my test "subjects", I would be sent to a psychiatric hospital straight away.

[Darth Wong]

The problem with the public perception of situations like this is that they will take an example of a situation where scientists screwed up and nobody caught it, and they will presume that this is indicative of the general state of all science, so that even the most well-tested theory should be assumed to be unreliable and/or fraudulent. Because hasty generalizations are, after all, how the mind naturally works, unless you've trained and educated yourself otherwise.

[Nova Andromeda]

But a surprising 33 percent confessed to other kinds of misconduct -- such as claiming credit for someone else's work, or changing results because of pressure from a study's sponsor.

...

A significant number --15 percent -- said they had changed the design, methods or results of a study in response to pressure from a financial sponsor.

In addition, 7 percent admitted ignoring "minor" rules for protecting human subjects. And 6 percent said that they failed to report data that contradicted their previous work.

...

Scientists, Martinson said, are "one of the hardest-working groups of people that I know." But he said there may be something about their working environment -- the mountains of rules, the pressure to compete for grants and to produce results -- that leads them to compromise their ethics.

"A lot of other professions engage in a lot of misbehavior -- look around at corporate America," he said. "There's been this kind of idea that scientists ... are super-humans or something, that they're immune from these kinds of pressure. But scientists are human."

...

--Emphasis mine.
-I find this "survey" to be highly biased. One only needs to look at major portion of this "misconduct" which is described as "... changed the design, methods or results of a study in response to pressure from a financial sponsor." So scientists are quilty of misconduct when they change their grant proposals upon review? That is totally absurd.
-In addition, at the very least 7% of scientists deal with human subjects? That is utterly false and it is clear that MD's and probably sociologists are being lumped together with full time scientists in an attempt to muddy the waters.
-The article also says that being a scientist leads one to compromise one's ethics "... that leads them to compromise their ethics." This is truly offensive and doesn't even deserve a rebuttal.

[wolveraptor]

-The article also says that being a scientist leads one to compromise one's ethics "... that leads them to compromise their ethics." This is truly offensive and doesn't even deserve a rebuttal.

Re-read it. It says nothing of the sort. All it says is that conditions of high-pressure and stress lead scientists to become worried when trying to produce the results. This may make them compromise their ethics. Being a scientist itself does not do any of these things.

[Nova Andromeda]

-The article also says that being a scientist leads one to compromise one's ethics "... that leads them to compromise their ethics." This is truly offensive and doesn't even deserve a rebuttal.

Re-read it. It says nothing of the sort. All it says is that conditions of high-pressure and stress lead scientists to become worried when trying to produce the results. This may make them compromise their ethics. Being a scientist itself does not do any of these things.

--You're a fucking idiot and I take personal offense at your apologetics. The message is clear: "Scientists, Martinson said, are 'one of the hardest-working groups of people that I know.' ... something about their working environment ... leads them to compromise their ethics. "
-It doesn't talk about high-pressure work environments in general. It specifically targets the working environment of scientists. Perhaps you would like to explain how one can be a legitimite scientist and be subject to a scientific working environment.