Science Confirms U.S. Scientists Lie More Than Any Other Scientists

Casey Hendrickson & Heather Kydd on KDOX ^

[AndrewC]

Americans publish 30% of all scientific papers.

Do you mean "publish" or "author"?

[allmendream]

My point is that the standards in blogging are low, and mathematical competence is at an all time low. Neither the author of the “study” (he did pubmed searches) nor his peer reviewers in the “Medical Ethics” publication he published in had the mathematical competence to figure out what BETTER bloggers (links previous post), myself and another FR poster immediately figured out.

It isn't the total number... it is the RATE.

Politics does tend to trump all, and a bit of Anti-Americanism seems to sell well to a particular segment - as does anti-science to a different sort.

[AndrewC]

My point is that the standards in blogging are low, and mathematical competence is at an all time low. Neither the author of the “study” (he did pubmed searches) nor his peer reviewers in the “Medical Ethics” publication he published in had the mathematical competence to figure out what BETTER bloggers (links previous post), myself and another FR poster immediately figured out.

Well, I don't accept my science from blogs. I accept it from multiple journals and my own experiences. That said, numbers are easy to make up on blogs. You realize, of course, that this is a blog and you have already denigrated it as a source of mathematical accuracy. You can "fortify" whatever numbers you provide by giving a reliable source for them. In any case, I sincerely doubt your conclusion as to the mathematical competence of those you have belittled. All I know is that the author of the paper went through a process which ostensibly provides evidence of some reliability. Your analysis has not gone through that process. However, I would believe your analysis if you can provide the raw data. I suspect you may be correct, but my suspicions and a buck will buy a cup of coffee.

[SaraJohnson]

I read they are studying the public to determine why Americans don’t believe the “smart” experts when they tell us b.s. like they do in Europe.

Socially and culturally, we are like the Soviet Union now!

[SaraJohnson]

Are bloggers not allowed to post on Free Republic? I hope they are because I enjoy the interesting information they find and bring here to share with us.

[AndrewC]

To illustrate my misgivings, here are numbers I received from the links in the references you provided.

Medline trend for 

  Number Year     %000
  852172 2009 100000.000
  817533 2008 100000.000
  773972 2007 100000.000
  737724 2006 100000.000
  692746 2005 100000.000
  631508 2004 100000.000
  588951 2003 100000.000
  559159 2002 100000.000
  541365 2001 100000.000
  527130 2000 100000.000

The total was 6722260.

This was the result of just pressing the build trend button which provides "•To find out just how many papers have been indexed by PubMed every year, enter an empty query (simply press 'Build Trend');"

And finally, this warning is given....

WARNING: Counting papers with a given feature is a very gross bibliometric method. Sometimes, the results are relevant, sometimes they require extensive checking, but they must always be interpreted very carefully.

[allmendream]

The figure of 30% seems fairly accurate and is based upon the EXACT same analytical method as the “study” in question, looking at pubmed results; and when you NORMALIZE to find the RATE, you get a much more RELEVANT measure.

The conclusion as to total numbers is about as relevant as pointing out that INDY 500 drivers are going to need to replace more tires than your average driver.

More miles = more tires.

More publications = more rejections.

The overall rejection rate is quite low, as is the fraud rate; and Americans seem quite honorable among the community of scientists and surely shouldn't be besmirched by mere virtue of being the colossus that bestrides the Earth.

But Anti-Americanism sells to a particular segment, and anti-science sells to a different segment.

[rockinqsranch]

An informative post. Thanks for the reply.

[AndrewC]

The figure of 30% seems fairly accurate and is based upon the EXACT same analytical method as the “study” in question, looking at pubmed results; and when you NORMALIZE to find the RATE, you get a much more RELEVANT measure.

I'm not arguing with that. I'm arguing with the accuracy of the numbers to begin with. I went to the site and entered a query which gave the total of publications indexed on pubmed. I then added the totals for 2000-2009. I got 6722260. The numbers you used included 6111156 for the same range. There is something wrong here.

In any case, you can analyze the statistics in many ways. 

84/197 = 42.6% of the fraudulent papers are U.S. authored while 

1819543/6111156 = 29.8% of the total papers were U.S. published.
What does that mean? Well, it looks like U.S. authored fraudulent papers are disproportionate.

[AndrewC]

Let’s turn that fraudent number on its head. Instead consider it as the number of U.S. Nobel prize winners versus total Nobel prize winners during those years. Obviously, they would be “included” in the set of published papers.(or so I say just to support my argument) ;^)

[dragnet2]

Here in places like Summerlin for example would be akin to Westlake, Simi, Thousand Oaks, Encino, Sherman Oaks and so on.

Until ya look over the wall, and everything else appears to be an abandoned nuclear test site...lol

I'm just banging on your bars Ron....

[SunkenCiv]

...global warming no longer exists, and never was the result of man-made activity.

· join · view topics · view or post blog · bookmark · post new topic · subscribe ·

[allmendream]

The idiocy of comparing just the total number is so obvious that I cannot believe I am having to point out just how huge an error it is.

Saying that American scientists are more likely to commit fraud is meaningless unless you calculate the RATE of fraud, not the total number of fraud papers from America.

The RATE of fraud papers from America is in line with Japan and less than India, China and South Korea, and more than the UK and Germany.

China and India are both more than twice as likely to have a paper retracted for fraud.

Keep in mind that the worse offender, India has a rate of fraudulent retractions of 0.018% or 18 papers out of 100,000. In America it is 5 out of 100,000.

[Las Vegas Ron]

I'm just banging on your bars Ron....

Touche FRiend...hang tight, the ride is just starting!

[AndrewC]

The idiocy of comparing just the total number is so obvious that I cannot believe I am having to point out just how huge an error it is.

It is not an error depending upon what information you are trying to convey. It is shown by the calculations I performed, using your numbers that the proportion of fraudulent papers involving Americans is disproportionate to the proportion of the total papers written by Americans. That is akin to discovering that a certain metal is involved in an extremely rare illness. The numbers may not be large but if a certain metal is present out of proportion to its presence in the general populace, then "Lucy's got a lot of splainin to do".

[AndrewC]

Here are some real statistics drawn from "Cancer Query Systems: Delay-Adjusted SEER Incidence Rates". The rate per 100000 puts these numbers in line with the fraud numbers. Are these just numbers to discard or is there something in there? Did something happen in the 1989-1990 range? Graph the points.

SEER Incidence Delay-Adjusted Rates, 1975-2007

Selections:

Statistic type = Delay-adjusted Rate;

Registries = SEER 9 Registries (1975+);

Site = Thyroid;

Sex = Male and female;

Age at diagnosis = All Ages;

Results:

1975

All races	4.9
White	4.7
Black	3.0

1976

All races	4.8
White	4.7
Black	3.4

1977

All races	5.4
White	5.3
Black	3.7

1978

All races	5.1
White	4.9
Black	2.9

1979

All races	4.5
White	4.4
Black	3.2

1980

All races	4.3
White	4.3
Black	2.4

1981

All races	4.4
White	4.4
Black	2.4

1982

All races	4.6
White	4.6
Black	3.3

1983

All races	4.7
White	4.7
Black	3.0

1984

All races	4.8
White	4.7
Black	3.0

1985

All races	5.1
White	5.1
Black	2.9

1986

All races	5.3
White	5.2
Black	3.4

1987

All races	5.0
White	5.0
Black	2.8

1988

All races	4.9
White	4.9
Black	2.5

1989

All races	5.4
White	5.5
Black	2.5

1990

All races	5.5
White	5.6
Black	3.8

1991

All races	5.5
White	5.7
Black	3.2

1992

All races	5.9
White	5.9
Black	4.2

1993

All races	5.6
White	5.8
Black	3.4

1994

All races	6.1
White	6.2
Black	3.8

1995

All races	6.2
White	6.4
Black	3.1

1996

All races	6.5
White	6.7
Black	3.9

1997

All races	6.8
White	7.1
Black	4.0

1998

All races	7.0
White	7.2
Black	3.9

1999

All races	7.4
White	7.6
Black	4.6

2000

All races	7.6
White	8.1
Black	4.6

2001

All races	8.3
White	8.7
Black	4.9

2002

All races	9.2
White	9.7
Black	5.5

2003

All races	9.6
White	10.3
Black	5.3

2004

All races	10.1
White	10.8
Black	6.3

2005

All races	10.8
White	11.8
Black	5.9

2006

All races	11.2
White	12.1
Black	6.5

2007

All races	12.2
White	13.1
Black	7.3

Table Variables:

Page = Year of diagnosis;

Row = Race;

Notes:

• Source: SEER November 2009 submission.  
  
• The SEER 9 registries are San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah and Atlanta.  
  
• The SEER 13 registries are the SEER 9 registries, San Jose-Monterey, Los Angeles, the Alaska Native Registry and Rural Georgia.  
  
• Rates are per 100,000 and age-adjusted to the 2000 US Std Population (19 age groups - Census P25-1130).  
  
• Rates are for invasive cases only unless otherwise stated.  
  
• Statistics are provided by the SEER Program for research purposes only.

 

Alternative Formats:

Delimited Data

In table form.

	All	`White	Black	Blk/Wht
1975	4.9	4.7	3	63.83%
1976	4.8	4.7	3.4	72.34%
1977	5.4	5.3	3.7	69.81%
1978	5.1	4.9	2.9	59.18%
1979	4.5	4.4	3.2	72.73%
1980	4.3	4.3	2.4	55.81%
1981	4.4	4.4	2.4	54.55%
1982	4.6	4.6	3.3	71.74%
1983	4.7	4.7	3	63.83%
1984	4.8	4.7	3	63.83%
1985	5.1	5.1	2.9	56.86%
1986	5.3	5.2	3.4	65.38%
1987	5	5	2.8	56.00%
1988	4.9	4.9	2.5	51.02%
1989	5.4	5.5	2.5	45.45%
1990	5.5	5.6	3.8	67.86%
1991	5.5	5.7	3.2	56.14%
1992	5.9	5.9	4.2	71.19%
1993	5.6	5.8	3.4	58.62%
1994	6.1	6.2	3.8	61.29%
1995	6.2	6.4	3.1	48.44%
1996	6.5	6.7	3.9	58.21%
1997	6.8	7.1	4	56.34%
1998	7	7.2	3.9	54.17%
1999	7.4	7.6	4.6	60.53%
2000	7.6	8.1	4.6	56.79%
2001	8.3	8.7	4.9	56.32%
2002	9.2	9.7	5.5	56.70%
2003	9.6	10.3	5.3	51.46%
2004	10.1	10.8	6.3	58.33%
2005	10.8	11.8	5.9	50.00%
2006	11.2	12.1	6.5	53.72%
2007	12.2	13.1	7.3	55.73%

[allmendream]

But the information doesn't match the headline.

It is as if you declared “Americans are less safe drivers than Peruvians” and you based that entirely upon the number of accidents per driver. But when you pointed out that Americans drive, on average, ten times the amount of miles per driver, the “safe driver” aspect is accidents PER mile, not number of accidents per lifetime. Otherwise the more you drive, the more likely you are to rack up an accident, and thus the less “safe” you are. People who never drive are, by this analysis, the most safe drivers in the world.

This is such a simple concept that I am again amazed that the author, the blogger, and now you seem to be having such a difficult time with it.

American scientists have a RATE of fraud that is HALF that of Chinese or Indian scientists.

The actual RATE of fraud bears no relationship to the headline.

[allmendream]

“As it turns out, you are far more likely to be given inaccurate (or fabricated) scientific data by American scientists, than any other group of scientists in the world.”

As it turns out, the person who wrote this sentence is WRONG WRONG WRONG, and so mathematically incompetent he doesn't even know he is wrong, and even after correction, still might labor under the delusion that he is correct.

You are more than TWICE as LIKELY to receive a fraudulent paper from a Chinese or Indian scientist than from an American scientist.

[Lazamataz]

Last time I was in Vegas, I looked like Mexico..

You did? Where on your body is your Yucatan?

[AndrewC]

This is such a simple concept that I am again amazed that the author, the blogger, and now you seem to be having such a difficult time with it.

Can't you read English? I understand your viewpoint, but I also understand that selecting one marble from a bowl of 6 million(6,111,156) marbles of which all are black except for 197 red ones and of those red, 84 are marked with "USA" when the selected marble happens to be red it will most often have a "USA" on it. That occurs despite the fact that selecting a red marble is an extremely rare event. Well how do you put that outcome in words? I assert that it is proper to say,"If it is a red marble you select then it will most likely have a "USA" on it. Likewise you can say, "If a paper is fraudulent then it was authored by an American." Well, that outcome is not surprising since Americans produce most of the papers. That is your point. But my point is that 42.6%(the proportion of the fraudulent papers written by Americans) is significantly larger (assertion) than 29.8%(the proportion of all papers written by Americans). So something is going on.

Finally, the word "safe" is a word with relative meaning as you are using it. So how would you characterize the safety of Martian probes? They travel a heck of a long distance.