James Russell Bellamy [I give permission for use in LIS Learning Showcase web server.]
May 2003
Dr. Coleman
International Journal of Geographical Information Science 1997-2001
ISSN: 1365-8816 (Paper) 1365-8824 (Online) ACCESSION NO. 4021277
www.tandf.co.uk/journals/tf/13658816.html
CULTURE
IJGIS is both a print publication and an electronic journal. Its aim is “to provide a forum for the exchange of original ideas, techniques, approaches and experiences in the rapidly growing field of Geographical Information Systems…intended to interest those who design, implement and use systems for monitoring, planning and policy-making. Research covers applications of GIS in natural resources, social systems and the built environment, and developments in computer science, cartography, surveying, geography and engineering…”
Publisher: Taylor and Francis Group
Head Office:
Editor: Peter Fisher—Dept. of Geography,
North American Editor: Harvey Miller—Dept. of Geography,
Western Pacific Editor: Prof. Brian Lees—Dept. of Geography, Australian
National University
Book Review
Editor: Dr. Nicholas Tate—Dept. of
Geography,
Contents pages of online issues can be browsed free of charge. However, in order to view full-text articles online, one must have purchased an online subscription, with its accompanying ID number, or likewise through an institution. For individuals, the price is $423 per year; for institutions, $901.
IJGIS publishes eight issues per year: Monthly, except February, May, August and November.
IJGIS was begun in 1987 under the former title of International Journal of Geographic Information Systems (ISSN: 0269-3798). It was changed to the current title in 1997.
The
U of A Main Library has print volumes 1-10 (from inception in 1987 through the
title change in 1996. All were in the
library as of
Circulation Policy: U of A Print Circulation Policy—Bound journals circulate for one day and are not renewable by phone or online. (Unbound do not circulate; must be used in the building.)
BIBLIOMETRIC ANALYSIS
PRODUCTIVITY
Publication/Articles IJGIS publishes eight issues per volume (year), 3 to 6 articles per issue. From 1997 through 2001, IJGIS published 244 articles, 84 book reviews, one correction, 16 editorials, one item about an individual, and 6 reviews. Because of this variance in types, the length of articles varies greatly. Articles can be anywhere from two pages to as many as 28 pages, with the average length somewhere between 10 and 12 pages.
Authors In 2001, more than 70 different authors contributed to IJGIS (counting co-authorship). The average number of authors is about 9 per issue. In the January 2001 issue, there were 15 different authors (again, counting co-authorship). Fifteen articles in 2001 had a single author; 33 were collaborations. Only two authors wrote for more than one article (3) in 2001 and these were in collaboration. Comparing the authors from 1997 to 2002 and the first half of 2003, we find just five authors still contributing to IJGIS. One of them is the much-respected colleague of Dr. Keith Clarke’s at UC Santa Barbara, Dr. Michael F. Goodchild. Thus, the criterion of “author turnover” for an assurance of fresh ideas and research, to avoid stagnation, seems to be well met.
Countries of origin IJGIS
published 48 articles in the year 2001.
Ten of these originated in the
Authors cited The most frequently cited authors during that period were: Briggs, DJ, Collins S, Elliott P, et.al. with 29 cites, from one article; Takeyama M and Couclelis H with 19 cites, from one article; Clarke KC and Gaydos LJ with 17, from one article (and Clarke alone from one book review); Stockwell D and Peters D with 16, from one article; Devogele T, Parent C, and Spaccapieta S with 16 from one article; and Wu FL with 16, from three articles, counting a collaboration with Webster CJ.
QUALITY—Impact, Value, Influence
For
a small “citivity” analysis, I chose IJGIS issue
Barr S, Barnsley M “A Region-based, graph-theoretic Data Model…”
Cited References: 52
Keywords: Land-use Classification, Contextual Classification, Systems
Janknowski P, Nyerges TL, Smith A, Moore TJ, Horvath E “Spatial Group Choice: A SDSS Tool…”
Cited Ref.: 47
Keyword: Systems
Stefanakis E, Theodoridis Y, Sellis T, Lee YC “Point Representation of Spatial Objects…”
Cited Ref.: 42
Keyword: Algorithm
Mean number of references: 47
Median number of references: 47
Most articles in IJGIS do not list as many references, though some list over 50 and one had 119! However, the average is more in the neighborhood of 22 citations per article with book reviews (usually having only one or two citations) included. When book review citations are not counted, the average rises to 30 per article.
Impact factor (as of 2001): 0.905 The “Impact Factor Trend Graph” in Journal Citation Reports shows a consistent and steady range, from a high of 1.093 in 1997 to a low of 0.769 in 1998, for an average Impact Factor of 0.939 for those five years—very respectable, as is 2001’s 0.905.
Immediacy Index: 0.171
Sources of References: From 1997 through 2001, works from The International Journal of Remote Sensing were cited 13 times in IJGIS articles; Geoinformatica 12 times; Geography Journal 10 times; Photogrammetric Engineering & Remote Science 7 times; Computer GeoScience—UK 7 times.
Cited Half-Life: 6.3 Citing Half-Life: 7.4
Citations in Other Works: Remembering to subtract out the citations made to other IJGIS articles, in ’97 other journals cited articles from IJGIS 40 times; ’98—56 times; ’99—39 times; 2000—23 times; 2001—only 5 times. The trend is clearly downward. The explanation may be that the field is no longer emerging (“exploding”) and maybe has leveled off, awaiting new technology.
Journal Co-citations: Articles that cited works from IJGIS contained citations to works in other well-respected GIS and geographical journals, e.g. Geography, The International Journal of Remote Sensing, Geoinformatica, The Annals of the Association of American Geographers, Geophysics Journal International, Geographical Journal, The Journal of Photogrammetry and Remote Sensing and Photogrammetric Engineering and Remote Sensing.
CONCLUSION
On the face of it, evaluating the quality of a journal would seem easy enough: lots of high-quality articles by well-respected writers/researchers per issue, articles that are frequently cited by others (Impact Factor), articles that are cited soon after publication (Immediacy Index). However, as Stephen Cole points out, it is not quite as precise as it might seem and there are some caveats. A good peer review system (“referees”) is needed. “Social evaluation” is an important factor for maintaining high quality. Of course, high quality authors submit to the most prestigious journals, but each article must be judged on its own merits. The prestige of the journal itself does not ensure a significant article. There is a “halo effect”, that is, articles in prestigious journals often being cited. However, Cole maintains that it has a small effect in the sciences. Those articles that are cited frequently in the sciences, he maintains, are “those that report a valuable piece of work that will need to be read by all workers in that field.” Conversely, few citations means the topic was too narrow or was of little importance. Thusly, Impact Factor will be an important factor in this evaluation. The goal here is to measure productivity in terms of output and impact, i.e. productivity and quality.
The International Journal of Geographical Information Science has the highest Impact Factor of the purely GIS journals (0.905) in 2001. The only journals covering GIS with a higher Impact Factor are Annals of the Association of American Geographers (1.855), but GIS is only one of many other subjects covered by this journal; and Journal of Photogrammetry and Remote Sensing (0.963), which is not as comprehensive as IJGIS. Likewise, the Photogrammetric Engineering and Remote Sensing journal, closely behind IJGIS with an Impact Factor of 0.841, has a more narrow focus. Annals of the Association of American Geographers received 1,119 total citations compared to 531 for IJGIS, but not all of these were in the GIS discipline. Even if half of them were (doubtful), IJGIS holds its own very well. In terms of output, IJGIS published 41 articles in 2001 (compared to 24 for AAAG) and all 41 were in the GIS subject area. Another journal listed as a GIS publication is Geographical Journal, which deals with many other subjects. It has an Impact Factor of 0.523 and published just 21 articles in 2001. Though a little lower in each, IJGIS compares well in Cited Half-life (9.9 to 6.3) and Citing Half-life (8.3 to 7.4) to AAAG. The research in IJGIS seems to remain relevant.
From the data I was able to gather, I would say The International Journal of Geographical Information Science is the premier journal focusing on GIS, in terms of productivity and quality, as well as staying power.
PERSONAL BIBLIOMETRIC PROFILE Keith C. Clarke (CLARKE KC) Geography professor, UC Santa Barbara
His specialties are cartography, remote sensing and GIS.
Dr. Clarke is also a former editor for IJGIS.
[NOTE: There are three other “CLARKE KC”’s listed in the Web of Science database—one is an economist, another is a medical doctor and the other is also a geographer!—to add to the confusion.]
Dr. Clarke of UCSB has nine articles listed in Web of Science written with no collaborator, 5 of which were strictly in the GIS discipline.
More than 60 authors have been cited by Dr. Clarke [Citation Identity]. They are: ACEVEDO W, ALBIN FA, ANSELIN L, BATTY M, BOATMAN J, BURROUGH PA, CHRISMAN N, CLAR S, COUCLELIS H, COX NJ, CRAIN IK, CRAWFORD J, DECKER D, DENTON RV, ESTES JE, FOURNIER A, GIBSON W, GOODCHILD MF, GOTTMANN J, HAGERSTRAND T, HARBAUGH JW, HARLEY JB, HARWIG P, HENDERSON BW, HEWISH M, HORGAN J, HSU ML, JACOBSON L, JENKS GF, KIERNAN V, KIRTLAND D, KOLCUM EH, KRAMER J, MANDLEBROT BB, MASKE AA, MARK DM, MONMONIER M, MORRISON JL, MULLER JC, NORDBECK S, PARK S, PENTLAND AP, PEUCKER TK, PEUQUET DJ, PRATT J, RHIND DW, ROBINSON AH, SAMPSON RJ, SHELBERG MC, SHEPARD D, SNYDER JP, SWEETMAN B, TABLER WR, TAKEYAMA W, TOBLER WR, VITOUSEK PM, WAGNER DF, WALDROP MM, WALTERS RF, WHITE R, WILSON J, WILSON RA, WOLFRAM S and WONG D.
Authors cited more than once by Dr. Clarke are: BATTY M, DECKER D, GOODCHILD MF, MARK DM and TABLER WR.
Between
1997 and 2001, WOS lists 65 cited
references for Dr. Clarke. [There were 76 listed in the search results, but 11
were for other Dr. Clarkes.] Authors who cited him more than once (recitation) were:
ANDERSSON C, ARMSTRONG MP, CARPINTERI A, CHANG KT, CHRISMAN N, EMERSON
CW, FISHER PF, MARCIANO R, MESSINA P, PERRY GLW,
Dr. Clarke has had 27 different articles cited in journals. One of his articles, “Computation of the Fractal Dimension of Topographic Surfaces”, has been cited 51 times! The five papers listed in WOS in the GIS discipline were cited 29 times in total. One of the papers was cited 17 time by itself. NOTE: Dr. Clarke refers back to his own articles quite often, i.e. self-citation.
Dr.
Clarke cited Nicholas Chrisman twice and Chrisman has cited him three
times. (In fact, Dr. Chrisman appears on
all three lists.) [see below] Dr. Chrisman is a professor of Geography at
the
[Citation Image] More than 225 authors have been cited in works where Dr. Clarke was cited (co-citation). There are 30 authors that have been co-cited in more than one work: ALBRECHT J, ARMSTRONG MP, BATTY M, BURROUGH PA (10 co-citations), CHOU YH, CHRISMAN N, COUCLELIS H, EGENHOFER MJ, FISHER PF, FRANK AU, GARDNER M, GOODCHILD MF (12), HAGERSTRAND T, LANDIS J, LANGRAN G, LAURINI R, LI X, MAGUIRE DJ, MARK DM, OPENSHAW S, PEUCKER TK, PEUQUET DJ, PORTUGALI J, RAPER JF, TAKEYAMA M, WEBSTER CJ, WHITE R, WORBOYS MF, WU FL, and YEH AGO. Goodchild and Burrough lead the way with 12 and 10 co-citations, respectively. One could surmise that this group provides a picture of the areas of expertise of Dr. Clarke, as well as his collegial peers, with Drs. Goodchild and Burrough having the most in common.
SUBJECT KEYWORD EXAMINATION
A topic search in the Web of Science database for just “systems” turned out to be too broad, of course, yielding 143,724 documents. Adding a term to narrow the search and access the discipline of GIS, such as “land” (50 documents), “mapping” (86 documents) or “geographic” (4 documents, though one was not in the GIS discipline) is necessary unless the searcher has a great deal of time.
Searching “land use classification” yielded 75 documents. Fifty-six (75%) were on the subject of GIS.
A search for “contextual classification” yielded 52 documents, but 16 (31%) were not related to GIS, therefore adding a more narrow descriptor would be necessary.
The topic “algorithm” likewise needed delineation. The search yielded 104, 456 documents, but few were related to GIS. Most were in Math, Engineering and Medical Physics.
Searching the general subject term “geographical information systems”, yielded 1,896 documents (though a few were not truly in GIS), still unwieldy, but better than over 100,000.
Geography is a multi-faceted, interrelated discipline, therefore information systems dealing with it, using modern technology, i.e. computers, satellites, are the modern tools that will facilitate researchers arriving at conclusions and applications. There is much quality research being done these days.
THE USEFULNESS OF BIBLIOMETRIC METHODS
Bibliometrics is a “family of techniques used to identify relationships in the published literature.” (von Ungern-Sternberg) The focus of which is to search for regular patterns in the characteristics of published literature and documents. Bibliometrics can be used to evaluate both Quality and Productivity in scientific communication. From what I gleaned from my readings, I compiled an outline for BIBLIOMETRIC ANALYSIS as follows:
I. Productivity
A. Publication
1. Frequency of publication
2. Issues per volume/year
3. Size/length of articles
4. Types of articles
B. Authors
1. Number of authors per volume/year
2. Number of authors per issue
3. Number of authors per article (co-authorship)
4. Frequency of authorship
5. Turnover
C. Articles
1. Number per volume/year
2. Number per issue
3. By country of origin
II. Quality—Impact, Value, Influence
A. Citivity/Citation Analysis
1. Number of references per article
2. Number of citations in other works
3. Authors most cited
4. Co-citation analysis
5. Impact Factor
B. Information Growth and Ageing
1. Sources of references
2. Immediacy Index (early recognition)
3. Age of references
a. Citation Half-Life
b. Median Citation Age
c. Price Index (<5 yrs. old/ >5 yrs. old)
Bibliometric methods can provide a description of patterns of publication within a given field or body of literature. (Palmquist) Citation tracing and analysis can help determine the influence of a single writer, describe relationships between two or more writers and even describe relationships between two or more works. (Palmquist) As with almost every statistical methodology, however, there are some caveats.
Some
of bibliometrics’ basic assumptions and laws have
slight flaws, e.g. Lotka’s Law,
Citations vary by types and degree. Thus, not all citations are equal. Some articles are cited for negative reasons. Early in a work, articles are cited merely as background material. Sometimes a citation is a “hat tip” to a colleague or a friend and there is even self-citation, calling into question the motives behind the citation. The expected or seminal works in a field are frequently cited, perhaps over-cited. Some of the best works can be overlooked, so a citation doesn’t always reflect high-quality. In addition, a citation does not take into account how extensively the cited document was used. These days there may even be diluted value of citations due to a current trend toward more and more citations.(von Ungern-Sternberg)
Dr. von Ungern-Sternberg and Dr. Blaise Cronin both feel the methodology can be criticized for incompleteness and lack of coverage in indexes such as ISI. Dr. Cronin has been pushing for years, he says, to get ISI to include monographs. He would also like to see acknowledgments added, e.g. those critiquing a paper, data analysts. There are some technical limitations in indexing and bibliographies, also, e.g. authors with the same last name (causing confusion or omission), synonyms, homonyms, etc.
Cybermetrics or Webometrics attempts to apply Bibliometric principles and techniques to the evaluation of possible conceptual linking of cyberlinks/hyperlinks on the World Wide Web, in order to determine the most useful and influential websites. This “scientific mapping” is based on the number of times a site is hyperlinked to another. This method, in its relative infancy, is fraught with problems and uncertainties, however. For starters, search engines do not index the entire web. The best search engines (Google, Yahoo) provide access to less than 60%.(SearchEngineWatch.com, 2002) Searching is inexact due to a lack of standardization of formats. Often titles or headings are not provided, or the author’s name doesn’t appear. Dr. Cronin expresses concerns about the “pedigree” of information over the ‘Net. Some has doubtful origins, lack of verification or supporting information.
The Internet is volatile and dynamic, making reliability an issue. URL’s come and go (even in the same search engine), so a search engine can lose information. Documents disappear and/or undergo changes, so that the search engine index doesn’t always reflect its contents. Also, there is often lag time in updating information. Dr. Cronin believes that these “are serious issues as to the reliability of the search engines.” He also believes that a new “tool” must be developed for evaluating
Internet links, “beyond acknowledgements and citation counts.”
Is a hyperlink as noteworthy as a citation? Most bibliometricians think not. Journal citations are not equal in value and hyperlinks are even more so. It is not always known why people link their pages to other ones. “Web linkages are different in nature from scholarly citations, and currently most links are to the journals themselves, rather than to the articles contained in them.” (Alastair Smith) Dr. Smith believes there needs to be more research in the nature of links and why they are made. Thus, usual bibliometric measures such as Impact Factor do not directly translate to Webometrics. New thinking and methods are needed.