Just a few additional ideas on the significance of the presented values
of the correlation coefficient.
For samples of size N from a bivariate normal with correlation r, its
standard deviation is approximately
StDev(R) = (1 - R^2)/sqrt(N – 1) - note that it depends on the number of
points used to calculate CC. One good reference is Hotelling, H. (1953).
New light on the correlation coefficient and its transforms. Journal of
the Royal Statistical Society, Series B, 15(2), 193-232. For the three
cases mentioned in Figure 3, fraud, error and duplicates the correlation
coefficients and their standard deviationa are respectively:
0.294 0.031 (almost 10 times above its standard deviation)
0.338 0.042 (also well above standard deviation)
0.119 0.045 (2.5 times above standard deviation, what the authors call
'slight' correlation)
A distribution of CC is not Gaussian, which is often inconvenient. One
can do a Fischer's z-transformation to obtain z-statistics
z = (1/2)ln((1+R)/(1-R))
which is approximately normally distributed with standard deviation
1/sqrt(N-3) and then do z-score tests on it. For the same three cases
the values of normally distributed z and their standard deviation are
very similar to the values obtained from the approximation above.
0.303 0.034
0.352 0.048
0.120 0.045
With best regards,
Victor
On 18/10/2012 19:52, DUMAS Philippe (UDS) wrote:
Le Jeudi 18 Octobre 2012 19:16 CEST, "Bernhard Rupp (Hofkristallrat a.D.)" <hofkristall...@gmail.com> a écrit:
I had a look to this PNAS paper by Fang et al.
I am a bit surprised by their interpretation of their Fig. 3: they claim that
here exists a highly signficant correlation between Impact factor and number of
retractations. Personnaly, I would have concluded to a complete lack of
correlation...
Should I retract this judgment?
Philippe Dumas
Dear CCP4 followers,
Maybe you are already aware of this interesting study in PNAS regarding the
prevalence of fraud vs. 'real' error in paper retractions:
Fang FC, Steen RG and Casadevall A (2012) Misconduct accounts for the
majority of retracted scientific publications. Proc Natl Acad Sci U S A
109(42): 17028-33.
http://www.pnas.org/content/109/42/17028.abstract
There were also a few comments on related stuff such as fake peer review in
the Chronicle of Higher Education. As not all may
have access to that journal, I have put the 3 relevant pdf links on my web
http://www.ruppweb.org/CHE_Misconduct_PNAS_Stuft_Oct_2012.pdf
http://www.ruppweb.org/CHE_DYI_reviews_Sept_30_2012.pdf
http://www.ruppweb.org/CHE_The-Great-Pretender_Oct_8_2012.pdf
Best regards, BR
-----------------------------------------------------------------
Bernhard Rupp
001 (925) 209-7429
+43 (676) 571-0536
b...@ruppweb.org
hofkristall...@gmail.com
http://www.ruppweb.org/
-----------------------------------------------------------------
