science fraud is common place

Where did you address my point??

"In Betrayers of the Truth, Broad and Wade speculate that for every major of case of fraud that comes to light, perhaps 100 go undetected, and that for each major case, another 1000 lesser violations are committed." (url removed from quote because it won't let me post it)

SO if only 1% of science fraud gets exposed how many can we extrapolate goes undetected....you do the math.

I can't believe that this hasn't yet been commented upon.

"For every major case of fraud that comes to light, perhaps 100 go undetected." Therefore, yes, 1% of science fraud gets exposed. That's a tautology. If only 1 in 100 cases of science fraud are noticed, then 1% are noticed.

So I'll do the math, like you asked. If 1% of science fraud gets exposed, 99% of science fraud goes undetected. Fine. That's a tautology.

Now, what percent of science is science fraud? The category of "science" appears nowhere in this analysis. You can draw absolutely no conclusions about it! All you're saying is how much science fraud is detected - but the percent of science that is fraud can be anywhere from 100% to 0.000000000000(...)000000000000000001%!
 
I can't believe that this hasn't yet been commented upon.
I can.

"For every major case of fraud that comes to light, perhaps 100 go undetected." Therefore, yes, 1% of science fraud gets exposed. That's a tautology. If only 1 in 100 cases of science fraud are noticed, then 1% are noticed.
No, you are making an error in your thoughts.

So I'll do the math, like you asked. If 1% of science fraud gets exposed, 99% of science fraud goes undetected. Fine. That's a tautology.
No, that is not what it means at all.
 
I like how you say "no" and then proceed to completely neglect to tell me where my mistake is.

(Unless you're trying to tell me that I should be doing 1/101 instead of 100, in which case I say, for the love of God, nobody's giving us exact numbers here, and rounding makes everybody's lives easier.)
 
Originally Posted by justintime View Post
Where did you address my point??

"In Betrayers of the Truth, Broad and Wade speculate that for every major of case of fraud that comes to light, perhaps 100 go undetected, and that for each major case, another 1000 lesser violations are committed."

SO if only 1% of science fraud gets exposed how many can we extrapolate goes undetected....you do the math.



I can't believe that this hasn't yet been commented upon.

"For every major case of fraud that comes to light, perhaps 100 go undetected." Therefore, yes, 1% of science fraud gets exposed. That's a tautology. If only 1 in 100 cases of science fraud are noticed, then 1% are noticed.

So I'll do the math, like you asked. If 1% of science fraud gets exposed, 99% of science fraud goes undetected. Fine. That's a tautology.

Now, what percent of science is science fraud? The category of "science" appears nowhere in this analysis. You can draw absolutely no conclusions about it! All you're saying is how much science fraud is detected - but the percent of science that is fraud can be anywhere from 100% to 0.000000000000(...)000000000000000001%!

You must have missed the link to the survey that declared 2% of scientists admit to committing fraud which is higher than the 1% quoted in other studies.The survey also concedes the actual frequencies of misconduct could be higher.
How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0005738

You also missed jt512 done the math results which can be found in his post# 680 where he is rather surprised how easy it was to calculate the total percent of science fraud. I am even more surprised at the muted response from the readers in reaction to the staggering conclusion JT512 arrived at.
Check JT512 post. http://www.internationalskeptics.com/forums/showpost.php?p=9744168&postcount=680

jt512 wrote: OMG! 200% of science is fraudulent!
:jaw-dropp
 
My point is that saying "For every major case of fraud that comes to light, perhaps 100 go undetected" is nothing like saying "for every scientist who admits that they might have done it once, there are a hundred who don't". The number of scientists who admit to committing fraud in anonymous surveys is probably going to be higher - a lot higher - than the number who were caught. Plus, the frequency of fraud among scientists who have admitted to (or been caught committing) fraud is going to be higher -presumably much higher - than the frequency of fraud among scientists who haven't.

Otherwise, yes, you end up with the nonsensical >100% fraud rates.

In essense, you can't conclude anything because you're trying to guess at something for which you have absolutely no data.
 
I can't believe that this hasn't yet been commented upon.

"For every major case of fraud that comes to light, perhaps 100 go undetected." Therefore, yes, 1% of science fraud gets exposed. That's a tautology. If only 1 in 100 cases of science fraud are noticed, then 1% are noticed.

So I'll do the math, like you asked. If 1% of science fraud gets exposed, 99% of science fraud goes undetected. Fine. That's a tautology.

Now, what percent of science is science fraud? The category of "science" appears nowhere in this analysis. You can draw absolutely no conclusions about it! All you're saying is how much science fraud is detected - but the percent of science that is fraud can be anywhere from 100% to 0.000000000000(...)000000000000000001%!

My point is that saying "For every major case of fraud that comes to light, perhaps 100 go undetected" is nothing like saying "for every scientist who admits that they might have done it once, there are a hundred who don't". The number of scientists who admit to committing fraud in anonymous surveys is probably going to be higher - a lot higher - than the number who were caught. Plus, the frequency of fraud among scientists who have admitted to (or been caught committing) fraud is going to be higher -presumably much higher - than the frequency of fraud among scientists who haven't.

Otherwise, yes, you end up with the nonsensical >100% fraud rates.

In essense, you can't conclude anything because you're trying to guess at something for which you have absolutely no data.
You raise a good point, Yerushalmi. For every fraud exposed a 100 goes undetected does not equate to the the number of scientists involved in fraud because a scientists can commit several frauds thereby skewing the results.

What we do know is 2% of scientists admit to fraud. According to JT512 that would translate to. "OMG! 200% of science is fraudulent!". But your point well taken would not mean 200% of scientists are fraudsters. That wouldn't make sense.

What the numbers would reflect is the percent for fraudulent submissions but not the actual number of scientists committing the fraud. But however you look at surveys you have to conclude science fraud is common place when you have 2% of scientists admitting to committing fraud and for every fraud exposed a 100 goes undetected which approximates 200% of science is fraudulent and that is just the tip of the iceberg because the survey also concedes the actual frequencies of misconduct could be higher. Try to shrink that perception to a manageable size so the scientific community is not viewed by the general public as both criminal and dysfunctional as well.
 
when you have 2% of scientists admitting to committing fraud and for every fraud exposed a 100 goes undetected
Yet:
1) no evidence whatsoever is presented for the latter statistic;
2) no evidence whatsoever is presented for the assertion that number of admitted frauds is equal to the number of exposed frauds.

Combined, these two basically discount any ability to draw the conclusion you wish to draw.
 
Otherwise, yes, you end up with the nonsensical >100% fraud rates.

In essense, you can't conclude anything because you're trying to guess at something for which you have absolutely no data.

JiT knows this. From his link:
Aside from such guesses, however, or indications in surveys, exact figures do not exist.
 
Yet:
1) no evidence whatsoever is presented for the latter statistic;
2) no evidence whatsoever is presented for the assertion that number of admitted frauds is equal to the number of exposed frauds.

Combined, these two basically discount any ability to draw the conclusion you wish to draw.

The book Betrayers of the Truth by Broad and Wade offer more that damaging statistics on scientific fraud. It is a critique of science and the scientific process. The book was published in 1982. Science fraud has substantially increased since. So one can even say their estimates that for every major case of fraud that comes to light, perhaps 100 go undetected were very conservative. It was recently estimated there is a tenfold increase in scientific research papers retracted for fraud.
What would a tenfold increase of jt512 OMG! 200% of science is fraudulent! amount to?
 
Science fraud has substantially increased since.
Provide evidence.
What would a tenfold increase of jt512 OMG! 200% of science is fraudulent! amount to?
jt512 was being sarcastic. If your analysis leads you to that conclusion, there's something wrong with your analysis - and if you think that the conclusion can or should be ten times as high, there's something seriously wrong with your understanding of percentages!
 
Last edited:
justintime said:
The book was published in 1982. Science fraud has substantially increased since. So one can even say their estimates that for every major case of fraud that comes to light, perhaps 100 go undetected were very conservative.
Translation: They were just speculating and throwing numbers out there. I'm going to speculate that those imaginary numbers were low. Then I'm going to pretend that this is somehow rational.

Is it any surprise that someone who thinks THAT is a rational process fails to understand science?

It was recently estimated there is a tenfold increase in scientific research papers retracted for fraud.
Once you get below a certain level, you're dealing with background noise. At that point, even fifty percent increases are insignificant.

Out of several hundred thousand if not several million person-hours worked last year my company experienced about 10 OSHA reportable incidents, and about 3 actual injuries (defined as anything requiring more than first aid). If there are 4 injuries next year, it's a 33% increase in reportable incidents. Out of hundreds of thousands if not millions of person-hours. And this tremendous rise consists of one person getting stung by a bee, or getting into a minor accident, or getting a blister on their foot (I've seen all three of them cited as OSHA reportables, and filled otu the paperwork for the first and last myself). This shows how small numbers yield large increases even with very, very little actual, real-world change.

Scientific fraud is the same. Out of hundreds of thousands of papers, less than 1% of 1% are fraudulant. A ten-fold increase raises it to less than 10% of 1% of published papers, and consists of a very, very small number of additional cases of fraud. This is background noise, not a breakdown of the system; and as such, even minor variations can be painted as dramatic upswings.

I defy you--or anyone else decrying science as broken--to find ANY other activity where less than 10% of 1% of the work done involves fraud.
 
Translation: They were just speculating and throwing numbers out there. I'm going to speculate that those imaginary numbers were low. Then I'm going to pretend that this is somehow rational.

Is it any surprise that someone who thinks THAT is a rational process fails to understand science?

Once you get below a certain level, you're dealing with background noise. At that point, even fifty percent increases are insignificant.

Out of several hundred thousand if not several million person-hours worked last year my company experienced about 10 OSHA reportable incidents, and about 3 actual injuries (defined as anything requiring more than first aid). If there are 4 injuries next year, it's a 33% increase in reportable incidents. Out of hundreds of thousands if not millions of person-hours. And this tremendous rise consists of one person getting stung by a bee, or getting into a minor accident, or getting a blister on their foot (I've seen all three of them cited as OSHA reportables, and filled otu the paperwork for the first and last myself). This shows how small numbers yield large increases even with very, very little actual, real-world change.

Scientific fraud is the same. Out of hundreds of thousands of papers, less than 1% of 1% are fraudulant. A ten-fold increase raises it to less than 10% of 1% of published papers, and consists of a very, very small number of additional cases of fraud. This is background noise, not a breakdown of the system; and as such, even minor variations can be painted as dramatic upswings.

I defy you--or anyone else decrying science as broken--to find ANY other activity where less than 10% of 1% of the work done involves fraud.

I don't think there is any profession that would brag when 2% of their researchers/workforce admit to committing fraud and knowingly acknowledge the actual frequencies of misconduct are much higher but go undetected.
 
justintime said:
I don't think there is any profession that would brag when 2% of their researchers/workforce admit to committing fraud and knowingly acknowledge the actual frequencies of misconduct are much higher but go undetected.
This is yet another lie on your part. I never said that they'd brag about it. I asked you to find another industry with a fraud rate of less than 1% of 1% (you know, what the actual data say, rather than your speculative fantasy numbers).

Are you able to do it? Are you WILLING to do it? Or are you afraid to actually look to see if fraud rates in science are higher or lower than those in other walks of life?

Personally, my money's on you being afraid. You're to scared to even look at the actual data, prefering to make up random numbers with no evidence supporting them in order to give your arguments the apperance--but not the substance--of rationality. I think you're going to continue in that vein.
 
I don't think there is any profession that would brag when 2% of their researchers/workforce admit to committing fraud and knowingly acknowledge the actual frequencies of misconduct are much higher but go undetected.
The red and blue bits contradict each other. How can we know what the frequencies of undetected misconduct are?
 
The red and blue bits contradict each other. How can we know what the frequencies of undetected misconduct are?

We can't, but such things have never deterred Justintime from scurrilous accusations in the past.
 
We can't...

Not true. We can provide rigorous estimates, as I've discussed previously in the thread. The issue is that no one's done the leg work yet, so while in theory we could get a very good sense of this information at present it remains an unknown.
 
This is yet another lie on your part. I never said that they'd brag about it. I asked you to find another industry with a fraud rate of less than 1% of 1% (you know, what the actual data say, rather than your speculative fantasy numbers).

Are you able to do it? Are you WILLING to do it? Or are you afraid to actually look to see if fraud rates in science are higher or lower than those in other walks of life?

Personally, my money's on you being afraid. You're to scared to even look at the actual data, prefering to make up random numbers with no evidence supporting them in order to give your arguments the apperance--but not the substance--of rationality. I think you're going to continue in that vein.

You have not read the last survey I posted that provided numbers.

How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0005738

A pooled weighted average of 1.97% (N = 7, 95%CI: 0.86–4.45) of scientists admitted to have fabricated, falsified or modified data or results at least once –a serious form of misconduct by any standard– and up to 33.7% admitted other questionable research practices. In surveys asking about the behaviour of colleagues, admission rates were 14.12% (N = 12, 95% CI: 9.91–19.72) for falsification, and up to 72% for other questionable research practices.

However, it is likely that, if on average 2% of scientists admit to have falsified research at least once and up to 34% admit other questionable research practices, the actual frequencies of misconduct could be higher than this.
 
You have not read the last survey I posted that provided numbers.

Oh, I read it. However, I understand the issues with such surveys, unlike you. They're little more than anecdotes.

And I note that you continue to ignore the fact that your 1 in 100 or 1 in 10 numbers or the like are pure fiction; that other data contradicts your arguments; and that a rigorous methodology has been proposed to determine rates of undetected fraud. You want to play make-believe, not actually engage in a serious discussion.

I also note that you continue to ignore phrases like "...it is likely...", which prove that the statements following are nothing but fantasy with no rigorous data to support them. After all, if the author had the data, why not provide it? Answer: Because they don't have the data to support it.
 

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom