If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(NPR)   A new study finds that two thirds of published biomedical papers contained fraud or plagiarism. These results may not be reproducible   (npr.org) divider line 64
    More: Fail, research papers, immunologists, independent media, The Lancet  
•       •       •

1065 clicks; posted to Geek » on 02 Oct 2012 at 11:43 AM (1 year ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



64 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest

Archived thread

First | « | 1 | 2 | » | Last | Show all
 
2012-10-02 09:28:10 AM
Two thirds of retracted papers, not all papers. Specifically there are 2000 of them since 1977, which it says in the article.
 
2012-10-02 09:30:10 AM
Did subby read the article or just the abstract? Two-thirds of retracted papers contain fraud or plagiarism, not two-thirds of all papers. Big difference.
 
2012-10-02 09:30:51 AM
What he said
 
2012-10-02 09:35:01 AM

sweetmelissa31: Two thirds of retracted papers, not all papers. Specifically there are 2000 of them since 1977, which it says in the article.


Well, it does say "These results may not be reproducible".
 
2012-10-02 11:48:15 AM
img-s3-01.mytextgraphics.com
 
2012-10-02 11:51:42 AM
Does this really surprise anyone? When you have a profession in which your livelihood depends on publishing a ton of papers, all with positive findings (since journals rarely publish negative findings), is it really a shocker that people will fudge?

Ain't saying it's right. Just that I understand.
 
2012-10-02 11:57:12 AM
welcome to American Crony Capitalism.


the finest money can buy.
 
2012-10-02 11:57:40 AM
And here I was prepared to shiat on all things biomed.

/chemistry
//subby needs to retract that headline
 
2012-10-02 12:00:11 PM
Not surprised. This is partly the reason why I didn't pursue this field after college.
 
2012-10-02 12:03:14 PM
How many papers refer to works that were later retracted? I think that would make for a more interesting field of study.
 
2012-10-02 12:07:03 PM

ThurmanMerman: Does this really surprise anyone? When you have a profession in which your livelihood depends on publishing a ton of papers, all with positive findings (since journals rarely publish negative findings), is it really a shocker that people will fudge?

Ain't saying it's right. Just that I understand.


Linux_Yes: welcome to American Crony Capitalism.


the finest money can buy.


KHIKYDRTFA?
 
2012-10-02 12:18:51 PM
Citations on everything gets tedious.
 
2012-10-02 12:23:01 PM

hovsm: Citations on everything gets tedious.


You think you're kidding.

/does it anyway
 
2012-10-02 12:23:26 PM
What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.

*Attempts to reproduce results of a paper, to use the nanoparticle/surface for an experiment*
*Procedure fails repeatedly*
*Contact original author*
"Oh yeah, well, you also have to heat it at 60*C when you're stirring the solution."
*NOT IN THE PAPER*

/FFFFFFFFFFFF.
 
2012-10-02 12:28:03 PM

sweetmelissa31: Two thirds of retracted papers, not all papers. Specifically there are 2000 of them since 1977, which it says in the article.


Cue the anti-vaxxers, climate change deniers and creationists - "Science is WRONG!!!!"
 
2012-10-02 12:28:25 PM

blahpers: ThurmanMerman: Does this really surprise anyone? When you have a profession in which your livelihood depends on publishing a ton of papers, all with positive findings (since journals rarely publish negative findings), is it really a shocker that people will fudge?

Ain't saying it's right. Just that I understand.

Linux_Yes: welcome to American Crony Capitalism.


the finest money can buy.

KHIKYDRTFA?




You're assuming the papers were retracted for reasons other than that the authors got caught.
 
2012-10-02 12:28:33 PM

Felgraf: What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.


To be fair, with a lot of nano it can get pretty equipment-dependent, and people don't usually want to put "faff around with your setup until you get dimensions of 5nm" or similar into their paper.
 
2012-10-02 12:30:27 PM

Felgraf: What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.

*Attempts to reproduce results of a paper, to use the nanoparticle/surface for an experiment*
*Procedure fails repeatedly*
*Contact original author*
"Oh yeah, well, you also have to heat it at 60*C when you're stirring the solution."
*NOT IN THE PAPER*

/FFFFFFFFFFFF.


Try this one
 
2012-10-02 12:46:12 PM

Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway


Who said I was kidding.
 
2012-10-02 12:47:02 PM

hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.


Sorry, that was a question.
 
2012-10-02 12:47:20 PM

Jim_Callahan: Felgraf: What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.

To be fair, with a lot of nano it can get pretty equipment-dependent, and people don't usually want to put "faff around with your setup until you get dimensions of 5nm" or similar into their paper.


Okay, this is true. Sometimes the size of the stirbar, or even the stirring speed, is important and people don't realize.

But when the paper leaves out "Oh yeah, you need to be heating this *as* you add the dry powder, or it won't all dissolve and you won't be able to make the membrane you want", that's kinda critical!

/Though I suppose it doesn't help that I come from the physics side, so some of the chemistry sections are basically "Mix these chemicals, THEN MAGIC HAPPENS."
 
2012-10-02 12:51:15 PM

hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.


I was adding a few citations to a research proposal last week. By the time I finished looking them up and putting them into EndNote, the entire morning was gone.

Doing the citations on the thesis is going to be painful.
 
2012-10-02 01:00:53 PM

Felgraf: What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.


This probably happens in all fields of science. I've tried to code simple numerical models from the descriptions given in papers, only to find that they left out a parameter value, or didn't fully specify the numerical scheme necessary to get stable results, etc. Or I work with their data, which it turns out they preprocessed in a vague way that I can't figure out.
 
2012-10-02 01:01:45 PM
Watched a TED talk the other day that also mentioned that selection bias made it so that papers with positive results were accepted and published into journals far more than papers with negative results.

So that a doctor relying on journals would only see the positive papers and not the negative ones.

http://www.ted.com/talks/ben_goldacre_what_doctors_don_t_know_about_ th e_drugs_they_prescribe.html


The bias was also not always intentional, EX, a company does a review on a drug that they consider marketing, they get bad results, they scrap the project (IE, they won't market the drug due to the bad result). But since they're not going to market the drug, they also don't bother to publish their negative result.

Meanwhile, another company tests an almost identical drug, gets a positive result, and publishes the paper.

Multiply this out to several companies on both sides and you end up with 45/50 positive results being published and maybe 5/50 negative results being published, so that from the outside it looks like 45 positive vs 5 negative when it's closer to 50/50.

// numbers pulled out of my ass for demonstration, not a representative sample, results not published :P
 
2012-10-02 01:01:48 PM
There is always at least some unintentional plagiarism in papers whether because of citations or how research statistics are explained.
 
2012-10-02 01:16:23 PM

lordargent: Watched a TED talk the other day that also mentioned that selection bias made it so that papers with positive results were accepted and published into journals far more than papers with negative results.

So that a doctor relying on journals would only see the positive papers and not the negative ones.

http://www.ted.com/talks/ben_goldacre_what_doctors_don_t_know_about_ th e_drugs_they_prescribe.html



Ah, if you haven't read Ben Goldacre's book, you're missing out. It's a good read.

Anyhow, this isn't really new. Not too long ago was a study done by a scientist at Bayer attempting to reproduce a bunch of high profile cancer studies. Turns out only a small handful of them could be reproduced.

These sorts of problems are bad, but they get worse when you add in the insanity that is the funding environment. Bad behavior, or at least bad thinking, becomes reinforced in that sort of environment.
 
2012-10-02 01:19:52 PM

mysticcat: Did subby read the article or just the abstract? Two-thirds of retracted papers contain fraud or plagiarism, not two-thirds of all papers. Big difference.


Maybe subby's doing some weird meta-commentary on the fact that people plagiarize without understanding what they're stealing and end up taking things out of context and misusing them.

But probably not.
 
2012-10-02 01:29:16 PM

mysticcat: Did subby read the article or just the abstract?


Hah, reading is for the weak. We have anti-science dogma to uphold!
 
2012-10-02 01:33:55 PM

hstein3: Anyhow, this isn't really new. Not too long ago was a study done by a scientist at Bayer attempting to reproduce a bunch of high profile cancer studies. Turns out only a small handful of them could be reproduced.


Yeah, although to be fair not being able to reproduce results exactly isn't always due to anything being incorrect or even fudging. This is especially true if you are only relying on say, whether you can get a result that was significant in one statistical analysis to be the same in another. Slight changes in undocumented parameters, undocumented preprocessing of data, etc can all effect downstream statistics quite a lot.

Of course, there is a bias in the whole "only results significant at a 5% cut-off are valid" as well.
 
2012-10-02 01:35:34 PM

jaytkay: sweetmelissa31: Two thirds of retracted papers, not all papers. Specifically there are 2000 of them since 1977, which it says in the article.

Cue the anti-vaxxers, climate change deniers and creationists - "Science is WRONG!!!!"


The scientific system is a good system. When you get scientists into the mix, that's wher the whole thing gets turned upside down.
 
2012-10-02 01:40:07 PM

lordargent: The bias was also not always intentional, EX, a company does a review on a drug that they consider marketing, they get bad results, they scrap the project (IE, they won't market the drug due to the bad result). But since they're not going to market the drug, they also don't bother to publish their negative result.


Interestingly, we *just* published a paper that had negative results (and backing up another group, who ALSO had negative results). I suppose it helps that these were both countering an earlier paper.
 
2012-10-02 02:26:06 PM
And people wonder why I snicker at the phrase "evidence-based medicine."
 
2012-10-02 02:33:33 PM

Felgraf: What I find more irritating is in the nano-filed, people tend to LEAVE CRUCIAL STEPS OUT OF THEIR GODDAMN PROCEDURES.

*Attempts to reproduce results of a paper, to use the nanoparticle/surface for an experiment*
*Procedure fails repeatedly*
*Contact original author*
"Oh yeah, well, you also have to heat it at 60*C when you're stirring the solution."
*NOT IN THE PAPER*

/FFFFFFFFFFFF.


That's how they know you read the paper.
 
2012-10-02 02:35:33 PM

Nurglitch: And people wonder why I snicker at the phrase "evidence-based medicine."


Because 0.007% of published articles were retracted due to misconduct? You are dumb.

(67% of retracted articles were found to be due to misconduct, but only 0.01% of articles are retracted in the first place.)
 
2012-10-02 02:46:37 PM

Ambitwistor: Nurglitch: And people wonder why I snicker at the phrase "evidence-based medicine."

Because 0.007% of published articles were retracted due to misconduct? You are dumb.

(67% of retracted articles were found to be due to misconduct, but only 0.01% of articles are retracted in the first place.)


I was thinking moreso the sheer volume of papers that are published being more material than a single person could digest and use. But hey, I like the way you jump to conclusions. It makes you seem pretty smart. We should hang out some time.
 
2012-10-02 02:53:24 PM

Bondith: hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.

I was adding a few citations to a research proposal last week. By the time I finished looking them up and putting them into EndNote, the entire morning was gone.

Doing the citations on the thesis is going to be painful.


Psst. Look up Mendeley. It was a life saver for me.
 
2012-10-02 02:59:19 PM

Nurglitch: I was thinking moreso the sheer volume of papers that are published being more material than a single person could digest and use.


Oh, you mean an observation irrelevant to both this thread and the utility of the evidence-based medicine literature. Gotcha.
 
2012-10-02 02:59:55 PM

ThurmanMerman: Does this really surprise anyone? When you have a profession in which your livelihood depends on publishing a ton of papers, all with positive findings (since journals rarely publish negative findings), is it really a shocker that people will fudge?

Ain't saying it's right. Just that I understand.


And don't forget, the apple only goes to the guy that finds it first.

/Everyone else chases crumbs
 
2012-10-02 03:10:09 PM
I don't know about biomedical papers; but almost all of the computing papers I've read should have never been written.

If you subtract out the fancy talk, the math, and the references - 99% of the time you're left with absolutely nothing worth talking about.
The people who write the other 1% are truly amazing. I'll never be one of those guys.

But I really, really, really hate that I'm being forced to participate in a giant mental circle jerk where I take something absolutely not worth mentioning and turn it into 12 pages of really smart sounding crap that *might* get published in a journal nobody reads (except other people who got their crap published too).
 
2012-10-02 03:18:30 PM

Nurglitch: Ambitwistor: Nurglitch: And people wonder why I snicker at the phrase "evidence-based medicine."

Because 0.007% of published articles were retracted due to misconduct? You are dumb.

(67% of retracted articles were found to be due to misconduct, but only 0.01% of articles are retracted in the first place.)

I was thinking moreso the sheer volume of papers that are published being more material than a single person could digest and use. But hey, I like the way you jump to conclusions. It makes you seem pretty smart. We should hang out some time.


What does that have to do with evidence-based medicine? The whole point of science in general is that people don't need to be experts in everything. That goes doubly for clinical trials and how we train physicians. Of course, I wish more physicians had more of a background in science beyond an undergraduate degree...

Fark_Guy_Rob: I don't know about biomedical papers; but almost all of the computing papers I've read should have never been written.

If you subtract out the fancy talk, the math, and the references - 99% of the time you're left with absolutely nothing worth talking about.
The people who write the other 1% are truly amazing. I'll never be one of those guys.

But I really, really, really hate that I'm being forced to participate in a giant mental circle jerk where I take something absolutely not worth mentioning and turn it into 12 pages of really smart sounding crap that *might* get published in a journal nobody reads (except other people who got their crap published too).


I'm in bioinformatics, this is what I hate about half of the bioinformatics papers published in journals only other bioinformaticians read. Try and make something actually useful and publish it so other scientists will use your method/program.

stainedglassdoll: Bondith: hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.

I was adding a few citations to a research proposal last week. By the time I finished looking them up and putting them into EndNote, the entire morning was gone.

Doing the citations on the thesis is going to be painful.

Psst. Look up Mendeley. It was a life saver for me.


Yup, thats what I use as well, worked much better for me during my thesis than most of my peers using EndNote.
 
2012-10-02 04:05:49 PM
87.56% of statistics are made up.
 
2012-10-02 04:30:17 PM
I think they started out saying that 2/3rds of all Climate AGW papers were but the director of the department threatened to have them shot while hiking in the woods so they quickly changed to biomedical research.
 
2012-10-02 05:03:34 PM

Bondith: hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.

I was adding a few citations to a research proposal last week. By the time I finished looking them up and putting them into EndNote, the entire morning was gone.

Doing the citations on the thesis is going to be painful.


Yes, yes it is. Endnote makes it a little faster, but you still have to put everything in.

I defend in two months. Moving on to postdoc in 4. Month 3 is "relax", move and try to figure out what I'll be doing there.
 
2012-10-02 05:32:15 PM
Retracted is the operative word. F@&k-tards subby
 
2012-10-02 06:18:49 PM

RogueVortex: Yes, yes it is. Endnote makes it a little faster, but you still have to put everything in.

I defend in two months. Moving on to postdoc in 4. Month 3 is "relax", move and try to figure out what I'll be doing there.


Good luck! I defended back in February, had a great external examiner who was more interested in being collegial and prompting good discussion so it went smoothly and was actually enjoyable. I had unofficially started doing work with my post-doc group last September while I was writing my thesis and started full time the day after I submitted my revisions.

A month break probably would have been nice but I wasn't moving so it wasn't a big deal.
 
2012-10-02 06:21:31 PM

entropic_existence: RogueVortex: Yes, yes it is. Endnote makes it a little faster, but you still have to put everything in.

I defend in two months. Moving on to postdoc in 4. Month 3 is "relax", move and try to figure out what I'll be doing there.

Good luck! I defended back in February, had a great external examiner who was more interested in being collegial and prompting good discussion so it went smoothly and was actually enjoyable. I had unofficially started doing work with my post-doc group last September while I was writing my thesis and started full time the day after I submitted my revisions.

A month break probably would have been nice but I wasn't moving so it wasn't a big deal.


Good luck! I've probably got about two more years, myself.

... The thought of leaving the protective cocoon of academia is honestly a TETCH scary. Even if I'd get paid more.
 
2012-10-02 06:57:51 PM

entropic_existence: stainedglassdoll: Bondith: hovsm: Bondith: hovsm: Citations on everything gets tedious.

You think you're kidding.

/does it anyway

Who said I was kidding.

I was adding a few citations to a research proposal last week. By the time I finished looking them up and putting them into EndNote, the entire morning was gone.

Doing the citations on the thesis is going to be painful.

Psst. Look up Mendeley. It was a life saver for me.

Yup, thats what I use as well, worked much better for me during my thesis than most of my peers using EndNote.



I've been using Endnote for a long time, and I've been a bit too reliant on the online integration with ISI instead of having local copies of my PDFs. I'll give Mendeley a try.

As long as people are talking about this, has anyone had any experience with Zoetero?
 
2012-10-02 06:58:40 PM

Damnhippyfreak: Zoetero


Zotero that is.
 
2012-10-02 06:59:59 PM

Felgraf: entropic_existence: RogueVortex: Yes, yes it is. Endnote makes it a little faster, but you still have to put everything in.

I defend in two months. Moving on to postdoc in 4. Month 3 is "relax", move and try to figure out what I'll be doing there.

Good luck! I defended back in February, had a great external examiner who was more interested in being collegial and prompting good discussion so it went smoothly and was actually enjoyable. I had unofficially started doing work with my post-doc group last September while I was writing my thesis and started full time the day after I submitted my revisions.

A month break probably would have been nice but I wasn't moving so it wasn't a big deal.

Good luck! I've probably got about two more years, myself.

... The thought of leaving the protective cocoon of academia is honestly a TETCH scary. Even if I'd get paid more.


I'm still in Academia, just not a student any more.
 
2012-10-02 07:02:06 PM

Damnhippyfreak: I've been using Endnote for a long time, and I've been a bit too reliant on the online integration with ISI instead of having local copies of my PDFs. I'll give Mendeley a try.

As long as people are talking about this, has anyone had any experience with Zoetero?


Yeah, I think there is an import function but I haven't used it. I have a Dropbox folder that holds all my papers and Mendeley just watches that. Does a good job of parsing PDFs to get all the metadata too, although some have to be corrected. Papers taken off of Pubmed that are only open access because of NIH is irritating as those PDFs have a cover page on them that gets parsed so that the manuscript title is NIH Access.
 
Displayed 50 of 64 comments

First | « | 1 | 2 | » | Last | Show all

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »






Report