The Flat Earth Society

Other Discussion Boards => Philosophy, Religion & Society => Topic started by: Tom Bishop on November 02, 2018, 11:08:54 PM

Title: 1 In 4 Statisticians Say They Were Asked To Commit Scientific Fraud
Post by: Tom Bishop on November 02, 2018, 11:08:54 PM
https://www.acsh.org/news/2018/10/30/1-4-statisticians-say-they-were-asked-commit-scientific-fraud-13554

From a report:

Quote
A stunning report published in the Annals of Internal Medicine concludes that researchers often ask statisticians to make "inappropriate requests." And by "inappropriate," the authors aren't referring to accidental requests for incorrect statistical analyses; instead, they're referring to requests for unscrupulous data manipulation or even fraud. The authors surveyed 522 consulting biostatisticians and received sufficient responses from 390. Then, they constructed a table that ranks requests by level of inappropriateness. For instance, at the very top is "falsify the statistical significance to support a desired result," which is outright fraud. At the bottom is "do not show plot because it did not show as strong an effect as you had hoped," which is only slightly naughty.

I cannot say that I am surprised in the least. Further, these requests for fraud aren't coming from some businessman. According to the article it is scientists who are making these requests.

Edit: Please move this to the science board.
Title: Re: 1 In 4 Statisticians Say They Were Asked To Commit Scientific Fraud
Post by: Rushy on November 02, 2018, 11:29:28 PM
The unfortunate thing about statistics is that you can build any number of false claims on top of perfectly valid data as long as you present your data in a specific order. I'm surprised the rate isn't far higher. I myself have been asked before to present certain data sets in a way which makes them appear more positive than they really are. I've also been asked to emphasize correlations of questionable quality merely to show the hypothesis as more solid than it otherwise would be. No one has ever asked me to outright falsify data, though.

The problem with these inquiries is that normally they can't be denied on a factual basis. Many times, you're not actually committing fraud, that is, you're not saying something that's patently untrue. Presenting the truth in a specific order isn't illegal, and if it was, then all media organizations would go under in a matter of hours. In the end, statistics, particularly statistical models, are dubious science at best. It's never a good idea to lay too much faith on statistical models, especially ones presented by people with a direct interest in saying a specific idea. The underlying facts are often correct while the conclusions are nothing but sometimes educated and usually heavily biased opinion.

Title: Re: 1 In 4 Statisticians Say They Were Asked To Commit Scientific Fraud
Post by: AATW on November 07, 2018, 11:40:23 AM
So...who compiled those statistics then and how do we know that number is accurate?
Title: Re: 1 In 4 Statisticians Say They Were Asked To Commit Scientific Fraud
Post by: disputeone on November 07, 2018, 11:52:02 AM
This is a very interesting thread, I should have paid it more attention. Tom do you have any information on the people who are pushing for the fraud? I'll do some digging myself. You can usually follow the money in cases like this.

Rushy you make a very good point on how statistics can be misrepresented.

The unfortunate thing about statistics is that you can build any number of false claims on top of perfectly valid data as long as you present your data in a specific order. I'm surprised the rate isn't far higher. I myself have been asked before to present certain data sets in a way which makes them appear more positive than they really are. I've also been asked to emphasize correlations of questionable quality merely to show the hypothesis as more solid than it otherwise would be. No one has ever asked me to outright falsify data, though.

The problem with these inquiries is that normally they can't be denied on a factual basis. Many times, you're not actually committing fraud, that is, you're not saying something that's patently untrue. Presenting the truth in a specific order isn't illegal, and if it was, then all media organizations would go under in a matter of hours. In the end, statistics, particularly statistical models, are dubious science at best. It's never a good idea to lay too much faith on statistical models, especially ones presented by people with a direct interest in saying a specific idea. The underlying facts are often correct while the conclusions are nothing but sometimes educated and usually heavily biased opinion.

This is what I think really goes on, it's much easier and more effective to just present the parts of the truth that support your argument than to lie. I think more people should read about propaganda and how it effects group psychology. It can arguably be used for good or evil but it's power and effect are undeniable.
Title: Re: 1 In 4 Statisticians Say They Were Asked To Commit Scientific Fraud
Post by: Rama Set on November 07, 2018, 12:54:47 PM
So...who compiled those statistics then and how do we know that number is accurate?

The creditations for the paper are available if you follow the link to the paper. It seems legit. What’s important to note is that the paper says nothing about what the statisticians did with the requests. A big missing part of the argument.