Google Scholar and Economics Nobelists
Stephen J Bensman
notsjb at LSU.EDU
Fri Jun 12 09:48:23 EDT 2015
We are engaged in research here at Louisiana State University on Google Scholar and economics Nobelists. A preliminary working paper on this research has been posted on arXiv at the following URL:
1. arXiv:1411.0928<http://arxiv.org/abs/1411.0928> [pdf<http://arxiv.org/pdf/1411.0928>]
Power-law distributions, the h-index, and Google Scholar (GS) citations: a test of their relationship with economics Nobelists
Stephen J. Bensman<http://arxiv.org/find/cs/1/au:+Bensman_S/0/1/0/all/0/1>, Alice Daugherty<http://arxiv.org/find/cs/1/au:+Daugherty_A/0/1/0/all/0/1>, Lawrence J. Smolinsky<http://arxiv.org/find/cs/1/au:+Smolinsky_L/0/1/0/all/0/1>, Daniel S. Sage<http://arxiv.org/find/cs/1/au:+Sage_D/0/1/0/all/0/1>, J. Sylvan Katz<http://arxiv.org/find/cs/1/au:+Katz_J/0/1/0/all/0/1>
Subjects: Digital Libraries (cs.DL)
An interesting phenomenon has emerged concerning Eugene Fama, who is one of the ten most popular economics Nobelists and probably has had the most important impact on society. He won the prize for developing the efficient market theory, laying the foundation of index funds, which dominate modern investing. Most retirement plans use these, and I myself have become economically secure by investing according to his principles. He has the most Google cites but the lowest R^2, which we hypothesize as a measure of how completely new findings have been integrated into the knowledge corpus of a given discipline. As the table below shows, Fama is also the only one who is on the borderline of the lognormal and the Gaussian domain. Inspection of his histogram below shows he is extremely deficient in low-cited publications and excessive in high-cited publications particularly at the right tip of the horizontal asymptote This deficiency at the low end causes the trendline to rotate counter clockwise, reducing the slope and the concomitant Lotka exponent. This indicates that his works are more broadly read. It also disproves that the R^2 is broadly applicable as a measure of knowledge integration. It also seems to indicate that Smolinsky, Lercher, and McDaniel (2015) are right: you cannot have a power-law distribution if you are deficient on the low end of the horizontal asymptote, where the low cited works concentrate, and the Yule-Simon model does a piss-poor job in estimating the upper end of the horizontal asymptote. This finding should be emphasized. One can compare Fama's histogram to that of Krugman, who is also one of the ten most popular economics Nobelists but more academic without the huge, practical impact on society. Krugman comes nowhere near the lognormal and the Gaussian domain. It thus seems that Fama is more broadly read, whereas Krugman's influence comes mainly from his works on the extreme right end of the asymptote, which contain the works for which he was awarded the prize. It does show that the power-law distribution is a wrong but useful scientific model, with which one has to be very careful. This may raise more questions than it answers-a low exponent leading to the Gaussian domain!?-but I thought that I'd throw these thoughts before you.
Respectfully,
Stephen J Bensman, Ph.D.
LSU Libraries
Lousiana State University
Baton Rouge, LA 70803 USA
E-mail: notsjb at lsu.edu
Smolinsky, L., Lercher, A. and McDaniel, A. (2015), Testing theories of preferential attachment in random networks of citations. Journal of the Association for Information Science and Technology. doi: 10.1002/asi.23312
Tests for Whether the H-Index Tuncated Belong to the Lotkaian or Gaussian Domains
Prize Winner
Index of Dispersion Test for Contagion*
Shapiro- Wilks Test for Normality
Index of Dispersion
Type Distribution
P Value
Lognormal Hypothesis
Name
Year
October 2013 Download
Robert William Fogel
1993
400.4
Contagious
P = 0.005
Rejected
C. W. J. Granger
2003
7604.7
Contagious
P < 0.001
Rejected
Thomas J. Sargent
2011
694.6
Contagious
P < 0.001
Rejected
Lloyd S. Shapley
2012
1529.1
Contagious
P =0.002
Rejected
Eugene F. Fama
2013
4682.1
Contagious
P = 0.001
Rejected
March 2015 Download
Robert William Fogel
1993
426.9
Contagious
P < 0.001
Rejected
C. W. J. Granger
2003
9217.4
Contagious
P < 0.001
Rejected
Thomas J. Sargent
2011
581.8
Contagious
P < 0.001
Rejected
Lloyd S. Shapley
2012
1854.2
Contagious
P = 0.001
Rejected
Eugene F. Fama
2013
5613.7
Contagious
P = 0.009
Borderline
* The index of dispersion is the variance divided by the mean. If it is significantly above one, then the distribution is contagious.
** If P is significantly below either 0.01 or 0.05, then the null hypothesis of the lognormal distribution is rejected.
[cid:image001.png at 01D0A4CD.73796190]
[cid:image002.png at 01D0A4CD.73796190]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20150612/492cefae/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 19294 bytes
Desc: image001.png
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20150612/492cefae/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 26903 bytes
Desc: image002.png
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20150612/492cefae/attachment-0001.png>
More information about the SIGMETRICS
mailing list