Enriching the Impact Regression Equation

Loet Leydesdorff loet at LEYDESDORFF.NET
Tue Jan 18 02:20:03 EST 2005

Dear colleagues,

Some of you may have seen the Pajek files that I made available for mapping
the journals of the Science Citation Index at
http://www.leydesdorff.net/jcr03/cited and at
http://www.leydesdorff.net/jcr03/citing . I am considering to extend this
representation by giving the vertices (nodes) the size of the
citations/publications (c/p) ratio as a measure of (average) impact in this
local environment. Since the matrix is at most places nearly decomposable,
this local impact is probably more meaningful than the global impact over
the file. (The threshold for inclusion in the journal environment was set to
1% of total cited/citing.)

I first considered using the impact factor, but thinking it over I realized
that the impact factor combines two normalizations: (1) the size effect and
(2) a kind of normalization over the time axis by using the two previous
years. However, it seems to me that the whole archive is always cited and
not just the last two years. The citation window and obsolescence of
publications can be expected to vary among disciplines. Thus, this
normalization over the time axis is not meaningful. Therefore, it seems as
clear (or unclear) to use just the c/p ratio. Would you agree? The c/p ratio
has an intuitive interpretation.

Since I can vary the size independently in the horizontal and the vertical
direction, I am considering to use the c/p ratio for the horizontal
dimension and perhaps total citing/total cited for the vertical one. Any
suggestions among this esteemed audience to do it differently (e.g., total
cited/total citing)? Would this be useful also for librarians? Or would it
be confusing? I am not sure how clear this will be in the eventual pictures,
but I'll let you know in due time when everything is in place.

With kind regards,

Loet  _____

Loet Leydesdorff
Amsterdam School of Communications Research (ASCoR)
Kloveniersburgwal 48, 1012 CX Amsterdam
Tel.: +31-20- 525 6598; fax: +31-20- 525 3681
loet at leydesdorff.net ; http://www.leydesdorff.net/

More information about the SIGMETRICS mailing list