The Naked Scientists

The Naked Scientists Forum

Author Topic: How common or rare is fraud in science?  (Read 1728 times)

Offline cheryl j

  • Neilep Level Member
  • ******
  • Posts: 1460
  • Thanked: 1 times
    • View Profile
How common or rare is fraud in science?
« on: 08/12/2014 02:20:55 »
Someone made the comment that you can't trust science because scientists are all bought off by big corporations, big pharm, the government, etc. and that I was obviously naive.

How common is fraud, as in publishing fake data, how is it most often detected, and what are the typical consequences for a researcher accused or proven to have done something like that?


 

Offline Arrual

  • Jr. Member
  • **
  • Posts: 24
    • View Profile
Re: How common or rare is fraud in science?
« Reply #1 on: 08/12/2014 17:20:54 »
My guess would be that its pretty dang common. Only because i see it happen all the time. I have even taken false information from people (because it normally sounds cool) and tried to make it my own. I unfortunatly dont know the consequenses.
 

Offline alancalverd

  • Global Moderator
  • Neilep Level Member
  • *****
  • Posts: 4698
  • Thanked: 153 times
  • life is too short to drink instant coffee
    • View Profile
Re: How common or rare is fraud in science?
« Reply #2 on: 08/12/2014 18:42:51 »
Most scientific publications are at best euphemisms for what actually happened in the lab. But as most scientific publications are only read by the author's doting mother, they are of little consequence.

The ones that matter are generally plausible and in principle reproducible. However governments do like policy-based evidencemaking, so anything that has potentially political consequences is almost certainly distorted and edited to the point of scientific invalidity. There is a telling footnote in the first report of the Intergovernmental Panel on Climate Change to the effect that "water is obviously the principal determinant of climate but as we cannot measure it or model its behaviour, we will ignore it" - and entire careers, massive conferences and huge tax increases  have derived from that statement. 

There's not a lot of evidence of actual lying in drug trial reports, but plenty of convenient halftruths make it into the press, aided and abetted by peer review and meta-analysis. 

History shows us that consensus is usually wrong, especially about big things like the flat earth, the geocentric universe, the fact that there is no conceivable military use for the airplane (US Academy of Sciences, 1908) and "about five computers will suffice for Britain's needs" (Ministry of Supply, 1950).

Off the subject a bit, my favourite example of journalistic makebelieve was reading two completely different accounts (in the Guardian and the Times) of the opening ceremony of the London Sumo championship. Both claimed to be eyewitness reports. But one was written by a woman: women are not permitted to attend the ceremony, so she must have copied her facts from the Sumo Association press release, which suggests that hers was the accurate report even though it was a lie, whilst the man who may indeed have been present at the ceremony may also have been inebriated or unable to read his shorthand notes, so his honest report was factually incorrect.

Beware of "adjusted" data. If the author hasn't also published the raw data set, he's probably hiding something.
 

Offline cheryl j

  • Neilep Level Member
  • ******
  • Posts: 1460
  • Thanked: 1 times
    • View Profile
Re: How common or rare is fraud in science?
« Reply #3 on: 10/12/2014 00:58:58 »
Okay, so if a researcher publishes a study, and then the five studies indicate he was off the rails, does it affect his career in anyway? At the very least, do other scientists snicker when he walks in the room?
 

Offline alancalverd

  • Global Moderator
  • Neilep Level Member
  • *****
  • Posts: 4698
  • Thanked: 153 times
  • life is too short to drink instant coffee
    • View Profile
Re: How common or rare is fraud in science?
« Reply #4 on: 10/12/2014 07:47:05 »
This is apparently the norm for psychology, where any experiment proves whatever hypothesis you first thought of. And it's no big deal in physics, as long as you were honest about your methods and results in the first paper: people may point out sources of error that you hadn't considered, or a weakness in your deductions, but peer review does give you some level of protection.

I think the wheels fall off your career if you make an outrageous claim that can't be substantiated - "cold fusion" was a fine example that got up people's noses because the first publication was a press release, not a peer-reviewed paper or fully-described patent, and nobody was able to reproduce it. "Polywater" was an earlier red herring that caught the imagination of the public but the originators (I think they were Russian) published "properly" and with sufficient technical detail that they were praised for spotting and reporting an unusual and reproducible phenomenon which they simply misinterpreted.

Outright lying can lead to temporary greatness if it is politically convenient (Lysenko, AGW...) or damnation if it isn't important (various nonsensical results in genetics). 
« Last Edit: 10/12/2014 07:50:34 by alancalverd »
 

Offline cheryl j

  • Neilep Level Member
  • ******
  • Posts: 1460
  • Thanked: 1 times
    • View Profile
Re: How common or rare is fraud in science?
« Reply #5 on: 11/12/2014 00:57:09 »
Are the conflicts of interests statements you sometimes see on studies mandatory, and are they generally forthright?
 

Offline alancalverd

  • Global Moderator
  • Neilep Level Member
  • *****
  • Posts: 4698
  • Thanked: 153 times
  • life is too short to drink instant coffee
    • View Profile
Re: How common or rare is fraud in science?
« Reply #6 on: 11/12/2014 08:47:53 »
It is reasonable to assume that any clinical trial of a named device or substance is sponsored by the manufacturer at some level, particularly if the published result shows superiority. But that's hardly surprising: if your prototype car wasn't faster/safer/prettier than your competitors', would you publicise the fact?

The only definitive statement is one of "no conflict" - or it used to be. Once upon a time in fairyland, "sponsored by the FDA" meant "independent" but now the FDA is funded by the industry it just means "all profits go to US shareholders". Thank you, Republicans. And the EU is just as protectionist, only slightly less transparent. 
 

The Naked Scientists Forum

Re: How common or rare is fraud in science?
« Reply #6 on: 11/12/2014 08:47:53 »

 

SMF 2.0.10 | SMF © 2015, Simple Machines
SMFAds for Free Forums