Review Articles, Paradigms, and the Impact Factor

[Stephen J Bensman]

I have just had the following article published:

 

"Distributional Differences of the Impact Factor in the Sciences vs. the
Social Sciences: An Analysis of the Probabilistic Structure of the 2005
Journal Citation Reports," Journal of the American Society for
Information Science and Technology 59 (July 2008): 1366-1382.

 

Although I generally do not like to call attention to my own work, I
consider this article very significant.  Then, too, there is also the
consideration once expressed by John L. Lewis, president of the United
Mine Workers, in his King James English that "he who tooteth not his own
horn, that same shall not be tooteth."  An executive summary of the
article is below.  The paper analyzes the relationship between Kuhnian
paradigms, review articles, and impact factors.  It comes to the
conclusion that a key characteristic of the journal literature of a
scientific discipline as opposed to others is that this literature is
dominated in the impact factor by review journals.  Sciences need
consensual paradigms in order to proceed, and the function of review
journals is to define such paradigms.  

 

The study is only exploratory, and its limitations are spelled out in
the last paragraph.  The hypothesis needs to be tested by further
studies, which I cannot do.  Therefore I am bringing the paper to
attention of others, who may want to investigate this question.  If
interested, please contact me, and I will provide access to the article.
I must warn you that you must love the Poisson process.  One of the
referees did not, and he thought that not only should this paper not be
published, but I should have been executed for having written it.  He
was outvoted 3 to 1.    

 

Stephen J. Bensman

LSU Libraries

Louisiana State University

Baton Rouge, LA   70803

USA

notsjb at lsu.edu

 

This paper examines the overall probability structure of the 2005
Science Citation Index (SCI) and Social Sciences Citation Index (SSCI)
Journal Citation Reports (JCR) by analyzing the distributions of their
journals by the Impact Factor.  The Impact Factor is an estimate of the
mean citation rate per citable item published in a given journal during
the preceding two years.  One major purpose of this analysis is to
determine whether there is any criterion for the macro-classification of
disciplines as to whether their journal literature conforms to a science
or a social science model.  Using the chi-squared index of dispersion
test, it was found that the Impact Factor distribution of the SCI
journals corresponded with a distribution generally modeled by the
negative binomial distribution (NBD), which arises as a result of the
stochastic processes of probabilistic heterogeneity and contagion.
Contagion is when the occurrence of an event raises the probability of
its subsequent occurrence, and it is the stochastic model of the
"Matthew Effect."  In contrast, the test found that the SSCI Impact
Factor distribution fit the Poisson distribution, which is the model for
random, rare events, although it came close to rejecting the Poisson in
favor of a distribution of the NBD type.  Both Impact Factor
distributions were positively skewed-the SCI much more so than the
SSCI-indicating excess variance.  An examination of the journals highest
in the Impact Factor revealed that in both JCRs the high-impact journals
tended to class in subject categories well funded by the National
Institutes of Health, and, therefore, it seems that the probability
structures of the JCRs are affected by the same biomedical "multiplier
effect" that influenced the development of American universities after
World War II.  It was also found that review journals play a much more
important role in the sciences than in the social sciences.  Review
articles tend to be more highly cited because of their important
function of defining the paradigms governing a given discipline, and
their importance was causal in the higher skewness of the SCI Impact
Factor distribution.  Because the sciences are generally considered to
have higher paradigm consensus than the social sciences, the greater
importance of review journals in the sciences was judged to be a
classificatory principle by which to gauge whether a discipline's
journal literature conforms to the science or the social science model.
As an example, the behavioral sciences were analyzed from this
perspective, and it was found that certain SSCI subject categories in
Psychology appear to adhere to the SCI model of dominant review
journals, thereby increasing the variance and skew of the SSCI Impact
Factor distribution.  For both the SCI and SSCI JCRs, surprisingly high
correlations were found between the journal Impact Factors and Total
Cites, indicating that the larger, older, and more prestigious journals
tend to have higher current mean citation rates per article-another sign
of the operation of the Matthew Effect..       

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20080613/3ab603a3/attachment.html>