cc345 at DREXEL.EDU
Thu Aug 19 21:25:26 EDT 2010
We have some preliminary results that seems to share some characteristics of Ronald's scenario.
When we measure the diversity of a paper's references with respect to the structure of its relevant network (we used a co-citation
network prior to the publication of the new paper), we found that the extent it alters the structure of the existing network
*sometime* ends up as the best predictor of citations in subsequent years.
Secondly, the cross-topical references, implemented as connections across co-citation clusters, appear
to be also a detectable predictor, but tends to be weaker than the first suspect.
The work is still at an early stage. Statistically significant effects are found in some datasets but not in others.
I intend to include the relevant functions in the new releases of CiteSpace by the end of the year.
You can find where we came from in the following paper:
Chen, C., Chen, Y., Horowitz, M., Hou, H., Liu, Z., & Pellegrino, D. (2009). Towards an explanatory and computational theory of scientific discovery. Journal of Informetrics, 3(3), 191-209.
Chaomei Chen, Ph.D., Associate Professor
College of Information Science and Technology, Drexel University
Editor in Chief, Information Visualization
ChangJiang Scholar, Dalian University of Technology, China
From: ASIS&T Special Interest Group on Metrics [SIGMETRICS at LISTSERV.UTK.EDU] On Behalf Of Pikas, Christina K. [Christina.Pikas at JHUAPL.EDU]
Sent: Thursday, August 19, 2010 12:12 PM
To: SIGMETRICS at LISTSERV.UTK.EDU
Subject: Re: [SIGMETRICS] References-Citations Relationship
Seems like you could test whether a document covered multiple subfields by clustering the references. They would probably cluster into methods and content clusters, but would the content (related work) citations cluster into subfields? Would you do it on journal names, words in the title or by pulling the full text? So then you would bin the article into one and more than one subfield bins and test the citedness means.
Yes, the variation can be explained as Phil says, but I think Ronald's scenario is plausible and could be tested. (maybe it has been already?)
Christina K Pikas
The Johns Hopkins University Applied Physics Laboratory
Christina.Pikas at jhuapl.edu
(240) 228 4812 (DC area)
(443) 778 4812 (Baltimore area)
From: ASIS&T Special Interest Group on Metrics [mailto:SIGMETRICS at listserv.utk.edu] On Behalf Of Ronald Rousseau
Sent: Thursday, August 19, 2010 2:11 AM
To: SIGMETRICS at listserv.utk.edu
Subject: [SIGMETRICS] References-Citations Relationship
I believe in the following scenario. An article that deals with
several topics (is related to several subfields) has a higher
probability of being useful (i.e. being cited) to at least one
subfield than an article that is related to just one subfield. It is,
moreover, probably richer in ideas.
Moreover, an article related to several subfields has on average a
longer reference list than an article dealing with one topic (or
related to one subfield).
Hence, there might be a relation between longer reference lists and
receiving more citations, although the length of the reference list
itself is not the cause of this relationship.
Who proves or disproves this conjecture?
President of the ISSI
KHBO - Association K.U.Leuven
Industrial Sciences and Technology
Zeedijk 101 - 8400 Oostende, Belgium
Professor associated to K.U.Leuven
Guest Professor Antwerp University, IBW
Honorary Professor Henan Normal University (Xinxiang, China)
Adjunct professor of Shanghai University
Guest Professor at the National Library of Sciences CAS (Beijing)
Guest Professor at Dalian University of Technology
Honorary researcher at Zhejiang University, Information Resources Management
E-mail: ronald.rousseau at khbo.be
web page: http://users.telenet.be/ronald.rousseau
There is nothing more practical than a good theory (Hilbert)
This message was sent using IMP, the Internet Messaging Program.
More information about the SIGMETRICS