I give permission for this project to be made available through the LIS Learning Showcase.
Bibliometric Study of the International Journal of Geographical Information Science
and One of Its Top Cited Authors, Fulong Wu

Gary F. Daught

School of Information Resources and Library Science, University of Arizona, Tucson, AZ
IRLS 589 Scholarly Communication
Professor Anita Coleman
 May 7, 2004
 

Introduction
Bibliometrics, is a word derived from a compound of two Greek terms, biblios "book" and metrikos "a measurement, or a measuring." But the "measurement" that is sought from the "book" is not its physical dimensions. Rather, bibliometrics is interested in measuring the life and impact of a discrete (traditionally written) form of communication within an intellectual community or the broader society. Borgman (2000) alludes to two classic definitions of bibliometrics: 1) the "counting and analyzing various facets of written communication" to shed light on the development of a discipline (Pritchard, 1968); and 2) the process of "assembling and interpreting statistics relating to books and periodicals...to demonstrate historical movements (Raisig, 1962)" (p. 144). As applied to scholarly communication, bibliometrics help researchers discern (through quantified data) the development of a discipline, and the patterns of communication among scholars and researchers.

The bibliometrician's primary data are citations. Cronin and Atkins (2000) quote Eugene Garfield (1979, p. 1) who writes: "Almost all the papers, notes, reviews, corrections and correspondence published in scientific journals contain citations. These cite--generally by title, author, and where and when published--documents that support, provide precedent for, illustrate, or elaborate on what the author has to say. Citations are the formal, explicit linkages between papers that have particular points in common" (p. 2, italics added). This idea is captured nicely by Joshua Lederberg, who termed reference citations as marking the "parent-offspring relationships of publications" (Cronin and Atkins, 2000, p. 3).

This paper will present a short bibliometric study of the International Journal of Geographical Information Science (IJGIS), one of the first scholarly journals dedicated exclusively to the interdisciplinary field of Geographical Information Systems (GIS), and one of IJGIS's top cited authors, Dr. Fulong Wu. Following, the paper will evaluate the usefulness of bibliometric methods in gaining an understanding of patterns of scholarly communication in a chosen discipline, and speculate about the potential of 'webmetrics' for analyzing Web-based scholarly communication.

 
International Journal of Geographical Information Science (IJGIS) [ISSN: 1365-8816]
Publisher: Taylor & Francis, London
Journal web page <http://www.tandf.co.uk/journals/titles/13658816.asp>
Current annual subscription rates (combined print and online): Institutional US$987; Individual US$463

Editor: Professor Peter F. Fisher, Department of Geography, University of Leicester
North American Editor: Professor Harvey Miller
Western Pacific Editor: Professor Brian Lees
Book Review Editor: Dr. Nicholas J. Tate

35 Member Editorial Board with broad international representation

IJGIS has been published continuously since 1987. The founding editors were J. Terry Coppock (Department of Geography, University of Edinburgh) and K. Eric Anderson (Eastern Mapping Center, U.S. Geological Survey). It was one of the first journals dedicated to exclusive publication of research articles in GIS (Geographical Information Systems) as it was emerging as a distinct field of study. Originally named International Journal of Geographical Information Systems [ISSN 0269-3798], the journal included research articles, a News Section (to post information on meetings and developments in GIS), Reviews, and frequently an Editorial addressing some current aspect of GIS.

The News Section was eventually dropped and the number of editorials were reduced. However, publication frequency steadily increased in the first ten years, allowing for a greater number of research articles to be printed: Quarterly (4 issues/year) 1987-1991; Bimonthly (6 issues/year) 1992-1995; and then 8 issues/year 1996-present. These trends indicated a settling-in of the journal's purposes, a formalization of the journal's readership network, and a maturing of the GIS field itself. While more recent Editorials are intended to "check-in" with readership about trends and developments or to introduce special issues (often with guest editors), earlier ones sought to "educate" the readership as to what to expect from the journal. Dropping the News Section probably reflected the impracticability of trying to keep-up with happenings in such a diverse and expanding field (other communication channels being more efficient). Meanwhile, increasing the number of research articles would be a clear indication of the maturing of the technical and scientific basis of the field.

In the very first issue of IJGIS (January 1987), the editors offered an extensive Editorial review (pp. 3-11), outlining historical developments in GIS, and proposing the objectives, intended subject areas, format and editorial policy of the journal. Their comments included these words:
Geographical information systems (GIS) represent a rapidly developing field lying at the intersection of many disciplines--among them cartography, computing, geography, photogrammetry, remote sensing, statistics, surveying and other disciplines concerned with handling and analysing spatially-referenced data. They are also of interest to a wide and increasing range of users, such as land and resource managers, market researchers, planners, property assessors and those responsible for utilities, to say nothing of administrators and policy makers in departments of government at all levels.…It is this very diversity and communality that has indicated the need for a journal that will be a source of information on advances and experiences and a meeting place for researchers, developers and operators of systems, and users of such systems, actual and potential.…We intend to take a broad view of the term GIS and provide a forum for the exchange of ideas, techniques, approaches and experiences, with a strong emphasis on applications. (p. 1)

[I]t seemed to many of those working in the field that a journal devoted to GIS was needed. Relevant developments were either not being reported publicly or were appearing in a wide range of disciplinary journals not normally seen by many of those interested in GIS… (p. 5)

IJGIS began with a clear sense of the interdisciplinary and international nature of this emerging field of study. A particularly strong emphasis is placed on the GIS practitioner and the user. In 1997 the name of the journal was changed to International Journal of Geographical Information Science. The rationale for the name change was explained in an Editorial by Peter Fisher in the January 1997 issue (Volume 11, Issue 1):
Over the 10 years of the publication of IJGIS the field to which it contributes has evolved considerably, and particularly the meaning of the 'S'.…Indeed, it is apparent that, in the 10 volumes of IJGIS to date, articles have contributed not to the understanding or development of systems as such, but to the science which underpins and exploits the systems.…Arguably, a natural system is there to be understood and researched, but an artificial one is built through the construction of the inter-related bits and is not a matter for research. Understandings of the issues which underlie the systems, on the other hand, comes about through research. The consequences of the whole and the best way to build them are therefore the subject of research, and more correctly described as science.…Therefore from Volume 11 the International Journal for Geographical Information Systems is to be known as the International Journal for Geographical Information Science.…The intention in renaming the journal is primarily to reflect better the nature of the material carried in it, as so to inform better potential readers and reviewers. (pp. 1-3)

The name change reflected a confidence (by the Editor and Editorial Board) that geographical information systems had evolved a strong enough theoretical base to be appreciated as a science unto itself, with a greater emphasis on research. A happy coincidence allowed the journal to keep the abbreviation IJGIS unchanged! But who decides when/if a field of study has matured sufficiently to claim the coveted mantle of science? On the one hand, it is observed that this assessment came lately (after only 10 years of publication) and by a proponent within the field. Might the interdisciplinary nature of GIS be creating a perception that the field of study still lacks a sufficient core, notwithstanding its own claim to the contrary?1 On the other hand, the field has clearly matured, and research activity has expanded since IJGIS began publication in 1987. This is reflected within the journal itself by its steady increase in the number of issues published yearly (from 4 to 8), and outside by a proliferation of other GIS journals.2 Too, appreciation for greater collaboration across disciplines seems to be increasing generally in the sciences. Whatever the answer in absolute terms, the determination will undoubtedly become clearer as journals like IJGIS continue to publish the results of GIS researchers.3 Indeed, some of the trends observed above--particularly the increased frequency of publication, and the accompanying increase in theoretical research--do suggest what Koehler (2001) (following the rubric of Derek J. de Solla Price) described for Information Science, as a transition in GIS from "small science" to "big science."4

Productivity Analysis of International Journal of Geographical Information Science Covering Years 1998-2002
The following tables summarize findings of a productivity analysis of IJGIS for the years 1998-2002 utilizing the ISI Web of Science databases.

Table 1: Document Types and Frequency: Article, Book Review, Correction, and Editorial Materials
Year
 Article
 Book Review
 Correction
 Editorial Materials
 Total Documents/Year
2002
 40
 19
 0
 1
 60
2001
 41
 4
 0
 3
 48
2000
 43
 14
 0
 2
 59
1999
 36
 10
 1
 3
 50
1998
 44
 27
 0
 3
 74
5 Year Totals
204
 74
 1
 12
 291
Mean (rounded)
 41
 15
 < 1
 2
 58

IJGIS increased publication to 8 issues/volume/year in 1996. Findings here reveal a stable editorial format focusing on research articles (average of 41/year), book reviews (averaging 15/year4), and the occasional Editorial (average 2/year).


Table 2: Five Top Cited Authors/Articles in the years 1998-2002
Top Author(s)
 Times
Cited
Stockwell D, Peters D
The GARP modelling system: problems and solutions to automated spatial prediction
INT J GEOGR INF SCI 13 (2): 143-158 MAR 1999		 41
Clarke KC, Gaydos LJ
Loose-coupling a cellular automaton model and GIS: long-term urban growth prediction for San Francisco and Washington/Baltimore
INT J GEOGR INF SCI 12 (7): 699-714 OCT-NOV 1998		 30
Devogele T, Parent C, Spaccapietra S
On spatial database integration
INT J GEOGR INF SCI 12 (4): 335-352 JUN 1998		 19
Wu FL
SimLand: a prototype to simulate land conversion through the integrated GIS and CA with AHP-derived transition rules
INT J GEOGR INF SCI 12 (1): 63-82 JAN-FEB 1998		 19
Bishr Y
Overcoming the semantic and other barriers to GIS interoperability
INT J GEOGR INF SCI 12 (4): 299-314 JUN 1998		 17


Table 3: Number of Documents (all types), and Number of Research Articles published in IJGIS (1998-2002) by authors named in the Five Top Cited Authors/Articles above.
Author
 Documents
 Articles
Bishr Y 1
 1
Clarke KC 2
 1
Devogele T 1
 1
Gaydos LJ 1
 1
Parent C 1
 1
Peters D 1
 1
Spaccapietra S 1
 1
Stockwell D 1
 1
Wu FL 3
 3
 

Table 4: Citivity and Subjects (Keywords) in a sample issue of IJGIS (Volume 16, Issue 8 [December 2002])
#
 Author(s)
Title
(Page Numbers)
 Document
Type
 References
Cited
Subjects (Keywords)*
1.

 van Oosterom, P; Maessen, B; Quak, W
Generic query tool for spatio-temporal data
(pp. 713-748)
 Article
 56
GIS: error, accuracy, quality, legislation (data processing)
2.

 Stefanidis, A; Agouris, P; Georgiadis, C; et al.
Scale- and orientation-invariant scene similarity metrics for image queries
(pp. 749-772)
 Article
 20
DATABASES; GIS: methodology (image analysis)
3.

 van der Poorten, PM; Jones, CB
Characterisation and generalisation of cartographic lines using Delaunay triangulation
(pp. 773-794)		 Article
 18
Cartography
4.

 Wu, FL
Calibration of stochastic cellular automata: the application to rural-urban land conversions
(pp. 795-818)		 Article
 29
SAN-FRANCISCO, MODEL, GIS, SIMULATION, DYNAMICS, INTEGRATION, EVOLUTION, GROWTH, FORM; Cartography (GIS)
5.

 Zhou, QM; Liu, XJ
Error assessment of grid-based flow routing algorithms used in hydrological models
(pp. 819-842)		 Article
 43
DIGITAL ELEVATION MODELS, TOPOGRAPHIC INDEX DISTRIBUTION, DRAINAGE NETWORKS, DIRECTION ALGORITHMS, REPRESENTATION, EXTRACTION, ACCURACY, TOPMODEL, SLOPE, AREAS; Meteorology and Climatology (error analysis)
6.

 Mather, PM
Manual of geospatial science and technology
(pp. 843-844)		 Book Review
 2
No keywords given
7.

 Chandler, J
Digital photogrammetry
(pp. 844-846)		 Book Review
 1
No keywords given
Research Articles = 5; Total References Cited = 166; Mean = 27.67 References/Issue; Median = 29
Book Reviews = 2; Total References Cited = 3; Mean = 1.5; Median = n/a

*Sources ISI KeyWords Plus, GeoBase descriptors

Selection of a single recent issue (Volume 16(8), December 2002) confirms the format emphasis (Table 1) on research articles (5) and book reviews (2) in virtually the same proportions as the yearly average for the entire period (41 articles/15 book reviews; 8 issues/year). The average (mean) for references cited per article was 27.67 and per book review was 1.5. Subjects and keywords underscore the interdisciplinary nature of the GIS field.


Table 5: IJGIS Impact Factor and Immediacy Index derived from ISI Journal Citation Reports for IJGIS (2002)
Impact Factor 0.821
"The journal impact factor is a measure of the frequency with which the 'average article' in a journal has been cited in a particular year. The impact factor will help you evaluate a journal's relative importance, especially when you compare it to others in the same field."
Immediacy Index 0.024
"The Immediacy Index is a measure of how quickly the 'average article' in a journal is cited. The Immediacy Index will tell you how often articles published in a journal are cited within the same year."

Given the Impact Factor for the journal as a whole, approaching 1 cite per "average article" published in 2000-2001 (0.821), cites from the five top author(s)/articles in the period 1998-2002 (Table 2) are most remarkable. For example, the top authors, Stockwell D, Peters D (1999), had 41 cites! These are not "average" articles, marking them as especially strong with a continuing influence on other research. Indeed, a survey of the citing titles range across many fields of study, and over time from August 1999-February 2004. In general the top authors (Table 3) during this period contributed only one research article to IJGIS each (with the exception of Wu, FL who contributed 3 articles, and Clarke, KC who also contributed a book review).


Table 6: IJGIS Impact Factor and Immediacy Index Rankings compared to other journals in broad subject categories. As absolute measures, it strikes me that although the Impact Factor for IJGIS is pretty good (approaching 1 cite per "average article" in 2000-2001), the Immediacy Index seems quite low (only 1 cite out of 41 published articles in 2002). To get some appreciation for what the numbers might mean in a comparative context, I conducted a journal search by Subject Category in both the JCR Sciences Edition (2002) and JCR Social Sciences Edition (2002).

Total Journals
 IJGIS Impact Rank
 IJGIS Immediacy Rank
JCR Science Edition (2002)
Computer Science (Information Systems)
 77
 28th
 54th
Geography (Physical)
 31
 21st
 28th
JCR Social Sciences Edition (2002)
Geography (Human)
 35
 21st
 25th
Information Science & Library Science
 55
 17th
 35th

These subject categories are quite broad. Still, as to Impact, IJGIS holds up pretty good especially in the Information Science & Library Science (17th of 55) and Computer Science (Information Systems) (28th of 77) subject categories. Although IJGIS's Immediacy ranking is low (consistent with the absolute measure), the only really dismal rank appears in the Physical Geography category (28th of 31). A more accurate picture of IJGIS's performance would arise by comparing it to other GIS journals specifically. For example, the results for IJGIS could be compared to the results gleaned from GIS journals by other students in the course. The picture is further obscured by the fact that ISI Web of Science doesn't index all of the GIS journals, so such information should be used with caution.

Online access and services.
The IJGIS web page enables non subscribers to browse contents (title, author with position title, address and email address, and page numbers) for Volume 10, Issue 1 (February 1996) through Volume 18, Issue 2 (2004), and read article abstracts (without references) for Volume 11, Issue 1 (January 1997) through Volume 18, Issue 2 (March 2004). The web site also includes a list of the services in which the journal is indexed and abstracted. Persons can subscribe to a free emailed Table of Contents alerting service (SARA: Scholarly Articles Research Alerting), and following a simple registration procedure, non subscribers can view a sample full text issue online.

Online full text for Volume 11, Issue 1 (January 1997) through Volume 18, Issue 2 (March 2004) are available to subscribers directly from Taylor & Francis' web site. EBSCOhost Academic Search Premier Publications, a third party ejournal subscription service to which the University of Arizona subscribes, has full text from Volume 12, Issue 5 (July 1998) through Volume 17 Issue 2 (March 2003). (Due to publisher restrictions there is a 12 month delay for full text, although article and review abstracts are available through current issue.)
Keyword Analysis
The use of keywords can assist in analyzing the foci of a given discipline. I looked at only a very narrow slice of this through a subject/keyword survey of a single issue of IJGIS (Table 4). Do the subjects/keywords appear to fit a discipline, Geographical Information Science, or the use and application of a tool, Geographical Information Systems to some problem or area of study? The most frequently cited subjects are GIS methodology and processing, and improvement of accuracy and quality. Notwithstanding Peter Fisher's contention (above) that "articles [in IJGIS] have contributed not to the understanding or development of systems as such, but to the science which underpins and exploits the systems" (italics added), the goal of the research still appears to be oriented toward the refinement and utilization of tools for analyzing spatially-referenced data, which are then applied to some problem or area of study. The strong emphasis, as at the start, is still on application and user.
Personal Bibliometric Profile of Fulong Wu
Among the top cited authors in IJGIS for the five year period I analyzed (1998-2002) was Fulong Wu.6 Dr. Wu is currently Senior Lecturer in Geography at the University of Southampton School of Geography, Southampton, England. He received his doctorate from University of Hong Kong in 1995. Previous to his current position at the University of Southampton, Dr. Wu held positions at The Centre of Urban Planning and Environmental Management, University of Hong Kong (1995), and the Department of City and Regional Planning, University of Cardiff (Wales) (1996-1998). The focus of much of his research writing is the process of rapid urbanization in China in the context of its transition to a globalized and market driven economy.

Two fascinating studies conducted by White, H. D. (2001a, 2001b) propose that a nuanced bibliometric analysis can give us much more information about a particular author than simply his/her number of cites, or how many times he/she is cited by others. The studies were based on the observed tendency of authors over the course of their publication career (in their body of work, or oeuvre) to repeatedly cite certain "orienting figures [as] recurrent points of reference" (2001b, p. 87). These names represent the core of the author's thought, with other names scattered with decreasing frequency (many that are never cited more than once). While most citation analysis is done anonymously (being interested mainly in the counts), White proposes that linking citers and citees by name gives us, using quantitative means, insight into an author's intellectual history, research style, motivations and influences. This set of cited names ranked in descending order is known as the author's citation identity. Although the author may not always be conscious of the process, "by the time they have created an oeuvre, it is highly likely that their citation identities will be as distinctive as fingerprints" (2001b. p. 88). Of course, in order to interpret the significance of this identity, a person needs a knowledge of the discipline and beyond--researched topics, persons doing the research, what contribution they have made/are making.

Citation identity is distinguished from the author's citation set (what White (2001a) calls the citation image-makers), which is a list of authors who cite the author, and the citation image, which is a list of authors who are co-cited with the author. Whereas the author creates his/her own identity, others create his/her image. As to the social dimension of intellectual relationships implicit in an author's identity or image, White (2001a) guesses that "citation is affected by a sense of intellectual seniority. Stated baldly, there is an apparent tendency to cite up [junior to senior] or across [perceived intellectual peers] the seniority chain much more than down [senior to junior]" (p 625). For example, the student would be more apt to cite his/her teacher than the other way around.

These sets are constructed from data supplied on ISI's Web of Science databases. White alerts that account must be taken for the inherent limitations of these databases. For example, there is no coverage prior to 1974, not all the journals in which an author may have published are indexed by ISI, and there are "unintentional inconsistencies in citing and data-entry practices" (e.g., homonyms--same spelling for different entities, and allonyms--different spellings for the same entity). White removed co-authored pieces from his counts, though he admits this may be "over fastidious." I did not remove co-authored pieces. Additionally, I used "Wu FL" as the search form for Fulong Wu. I didn't attempt to sort through the more general form "Wu F" (which returned 765 documents!) though I knew that some citations for my author were entered using this form (as an allonym). As it was, I had to exclude homonyms for "Wu FL"--a chemist, a physicist, and a human physiologist.

White (2001b) adds a further point about self-citation--where an author cites him/herself in subsequent writing. Some bibliometricians would remove self-citations as an indication of an egotistical tendency. He is more judicious, suggesting that rather than removing self-cites from a proportional count beyond the author's control, any charge of egotism "should perhaps be reserved for failings more clearly personal, such as citing one's own work when it is irrelevant" (p. 89). I have chosen to retain self-cites as part of the historical record.

I compiled Fulong Wu's Citation Identity (authors cited) from ISI Web of Science using "General Search." The result was 32 documents (22 sole-authored, 10 co-authored; 26 Articles, 4 Reviews and 2 Editorials), covering the years 1995 to 2004. I removed all institutional citations (e.g., census data sources).7 I compiled Dr. Wu's Citation Set (authors citing) using "Cited Reference Search." The result was 21 documents (13 sole-authored, 8 co-authored; 20 Articles and 1 Editorial), covering the years 1995 to 2003. I attempted to compile a Citation Image using a Dialog Search. However, as my knowledge of Dialog commands was limited, my search was not successful, and I was unable to generate Dr. Wu's image. The results follow in Table 7.

Table 7: Fulong Wu's Citation Identity and Citation Set

Citation Identity Citation Set
Rank
 Number
 Name
 Number
 Name
1
 21
 Wu, FL
 25
 Wu, FL
2
 18
 Yeh, AGO
 9
 Yeh, AGO
3
 15
 Harvey, D
 8
 Li, SM
4
 12
 Logan, JR
 8
 Lin, GCS
5
 11
 Bian, YJ
 5
 Huang, YQ
6
 11
 Dowall, DE
 4
 Ma, LJC
7
 11
 Knox, P
 4
 Ng, MK
8
 10
 Wang, YP
 4
 Wang, YP
9
 9
 French, RA
 4
 Wu, WP
10
 8
 Andrusz, GM
 3
 Wei, YHD
11
 8
 Ma, LJC
 2
 21 Others
12
 8
 Marcuse, P
 1
 119 Others
13
 8
 Nee, V

14
 8
 Walder, AG

15
 7
 Lin, GCS

16
 7
 Lo, CP

17
 7
 Short, JR

18
 7
 Szelenyi, I

19
 7
 Zhou, M

20
 7
 Zhu, J

21
 6
 8 Others


5
 12 Others


4
 22 Others


3
 42 Others


2
 81 Others


1
 275 Others


Since I am unfamiliar with GIS as a discipline, I am not in a position to offer any insightful interpretation of the results. I can see, however, that Dr. Wu is just now completing his first 10 years of research scholarship. It is reasonable to conclude that he has many years ahead. Coupled with the findings in Table 2 above, he appears to be making a mark in the field, especially as his research relates to the social and economic scene in urbanized China. Frequent self-citation, rather than hinting at egotism may reflect subsequent research built upon earlier. Dr. Wu also co-authored a number of articles with Anthony G. O. Yeh. They frequently cite one another, though Dr. Wu cites Dr. Yeh twice as frequently than Dr. Yeh cites Dr. Wu. Dr. Yeh teaches at the University of Hong Kong (where Fulong Wu got his doctorate), and based on publication dates, he appears to be an older contemporary of Dr. Wu--perhaps his major professor? If so, this would follow White's guess about the "seniority chain" above.
The Usefulness of Bibliometric Methods Applied to Scholarly Communication,
and the Future Potential of  'Webmetrics'
Communication is implicit in scholarly pursuit because knowledge can only accumulate as a human resource if it is shared. In scholarly communication, the goal is not only to add to human knowledge generally, but also to engage in collaborative relationships to advance knowledge within fields of study. Scholarly influence--given and received--is a form of communication. A student whose thought and career direction has been impacted and shaped by a respected teacher; friends or colleagues who share a common subject interest; an author long deceased whose written words continue to speak in a powerful way; or even a scholar whose research is vigorously disputed8 can contribute influence. Such communication is not always carried on in public, but the results of it are frequently observable to all in the products of a person's scholarship. "Influence…is recorded openly and unambiguously in the list of bibliographic references affixed to monographs and journal articles" (Cronin and Atkins, 2000, p. 4). Each reference (citation) is a sign--a form of shorthand or surrogate--pointing to and making a connection with a larger context of another's thoughts and ideas. These relationships can be traced at the personal level, revealing the origins and development of that particular scholar's thought. But the lines drawn between individual scholars quickly accumulate as each have and are impacted by their own spheres of influence. Tracing lines of reference back and forth in time, or across and between disciplines can produce a remarkably complex--a seemingly chaotic--picture. But as the lines are analyzed patterns begin to emerge. Shared influences, geographical or language clusters, knowledge cores and boundaries of given subject and discipline, etc. All this, the intellectual relationships of an individual scholar and the complex picture of interrelationships created by many scholars, is the domain of bibliometric analysis. And it all starts with a citation.

Scholarly communication is a social process. Bibliometrics is able to analyze this social process by quantitative means because citations are left as artifacts of that process. Artifacts, being real "things," can be counted. Still, to appreciate the methods and goals of bibliometrics, some appreciation must be gained for the social process of scholarly communication. J. C. Cole (2000) introduces his teacher, Robert Merton, as "the putative father of the sociology of science," who suggested that
sociologists could treat the organization of science--its social structure, its reward systems, its norms and values--as an object of study that would yield results of both theoretical and practical value--and that science was just as legitimate an object of sociological study as was the family, the economy, deviant behavior, religion, and other major social institutions. (pp. 283-284)

Sara von Ungern-Sternberg (2000) provides an excellent summary of the relevant theories of Merton, but also of Goffman's (1964) intriguing "theory of intellectual epidemics," which compares the diffusion of an idea into a scientific population to an epidemic caused by the spreading of an influenza virus in a population of people, and "the principle of cumulative advantage" formulated by Derek de Sallo Price (1976). Price states:
Success seems to breed success. A paper which has been cited many times is more likely to be cited again than one which has been cited little. An author of many papers is more likely to publish again than one who has been less prolific. A journal which has been frequently consulted for some purpose is more likely to be turned to again than one of previously infrequent use. (Quoted in Intellectual epidemics section, ¶ 3)

Of course, the sociology of scholarly communication begins with the psychology of the individual scientist or scholar. Von Ungern-Sternberg, deriving her thoughts from Garvey (1979) writes:
The psychology of scientific communication emphasises [sic] the interaction between the scientist and his environment. Each scientist brings to each situation a particular cluster of psychological attributes: personality, skills, style, experience, habits etc., which in combination with specific circumstances in the research process gives the individual scientist a predisposition to perceive and detect, assimilate and associate what is happening with his research at any given moment. As a consequence, given similar circumstances two scientists may perceive them differently, assimilate different information or apply the same information to fulfil individual needs to advance his particular research. (Quoted in introductory section, ¶ 7)

While I feel like I am out of my element when it comes to the bibliometric laws of Lotka, Bradford and Zipf, based on frequency distribution, empirical bibliometrics, especially those applied to the social and psychological aspects of scholarly communication are very intriguing to me. In this regard, I found the recitation analysis studies of H. D. White (2001a, 2001b) most interesting and insightful. The broad outlines of White's approach were applied above in the personal bibliometric profile of Fulong Wu. His key point bears repeating:
Many citation studies feature only cited authors or works and leave citers anonymous. Recitation analysis links citers and citees by name, implicating characteristics of both. [This approach] reveals orienting figures in an author's thought--other authors who are recurrent points of reference, taking that term to cover anyone from guru to sworn foe. One needs domain knowledge to interpret the list of names. If one has that, recitation analysis can be quite engaging--a source of intelligence that, unlike much in information science, is not depersonalized. (2001b, p. 87, italics added)

Indeed, although I appreciated that the personal bibliometric profile conducted above was for demonstration purposes, the most frustrating part was not being able to interpret the significance of my findings, because I didn't have domain knowledge.

The other aspect of the bibliometric study conducted above involved looking at the productivity of the GIS journal, International Journal of Geographical Information Science. If time would have allowed, the study could have been expanded to cover the entire publication life of the journal as suggested in the studies conducted by Koehler (2001) or Schubert (2002). Koehler was looking for evidence from a bibliometric analysis of the Journal of the American Society for Information Science that the field of information science was becoming "big science." This was the rubric of Derek de Sallo Price describing the evolution of science as once inhabited by serious amateurs (what he called "little science") to a formalized discipline inhabited by professional "scientists." Koehler nuances the application of terms, replacing amateur with "practitioner," and then he writes:
Little science in the Twentieth Century is the arena of both the professional scientist and of the practitioner. There is not much space for the practitioner in big science. One way to mark the transition from little science to big science for any given discipline is to document the change from practitioner to scientist among its participants. (p. 119)

Koehler proposed that by looking at the journals of a discipline it is possible to point to the evolution of the discipline. He selected six variables to document this evolution--the basis for his bibliometric analysis: 1) outside government, foundation, or other sponsored research funding, 2) increases in the number of authors per article, 3) increased content per issue and volume, 4) changes in author affiliation from "practitioners" to "scientists," 5) changes in co-authored patterns, and 6) greater globalization of authorship (p. 121).

I found Koehler's proposal particularly intriguing viewed against the editorial claim made by Peter Fisher on the occasion of the name change of the International Journal of Geographical Information Systems to International Journal of Geographical Information Science (see above). It seems that applying Koehler's bibliometric method to the IJGIS would be one way to test the claim of the editor. If the fact was already establishable that GIS was (at least) "little science" when Fisher made his remarks, Koehler's approach could offer evidence that GIS was (at least) in a transition toward "big science" in that same context.

From bibliometrics to 'webmetrics.' The growth of electronic publishing and increasing online dissemination of scholarly research over the Internet and the World Wide Web suggest a technological shift that may replace the printed journal as the preeminent conduit for scholarly communication. The benefits of speed, focus (custom disaggregation), multimedia, and global access are revolutionary for scientific and scholarly research. However, some very important functions that have been performed by the traditional print journal need to be addressed before such a replacement can be successfully effected. Von Ungern-Sternberg (2000) cites Kircz (1998), who described four functions of scholarly communication that will need to be established as "measurable units"--the basis for bibliometric analysis--in the electronic environment:
As has been discussed, citations establish the links of intellectual relationship in scholarly communication. In traditional print journals, citations are durable artifacts, and as such, they can serve as a key tool in the bibliometric analysis of scholarly communication. In adapting traditional scholarly communication and its analysis to the Web-environment ('webmetrics') a way to secure the "durability" of citations is absolutely necessary. The 5th edition of the Publication Manual of the American Psychological Association requires that Web-based citations include a retrieval date along with the URL. Behind this practical requirement is a tacit admission that durability cannot as yet be guaranteed. The next time the URL is visited, it may be broken, and the document effectively vanishes. If the link is broken so is the relationship. Current webmetrics that analyze data on mouse clicks and visit rates to digital libraries and document repositories, while useful, do not explicitly establish the intellectual relationships that are more clearly established through citation analysis. If this hurdle can be overcome, and a project like NEC's CiteSeer <http://citeseer.ist.psu.edu/cis> which provides "autonomous citation indexing" suggests that a solution may be coming soon.9 The future potential seems great.
Notes
1 The suggestion that GIS may still lack a sufficient core of theoretical research surfaced for me in a brief conversation I had with Professor Anita Coleman on March 8, 2004.

2 The growth within fields and subfields like GIS contributing to the overall increase in research article publication follows the observation made by Meadows, J (2000).

3 I am thinking here, too, of the distinction made by Stephen Cole (2000) between "core" and "frontier" knowledge. Cole writes: "The core consists of a small set of theories, analytical techniques, and facts that represent the given at any particular point in time [based on a] substantial consensus....The other component of knowledge, the research frontier, consists of all the work currently being produced by all active researchers in a given discipline. The research frontier is where all new knowledge is created....The research frontier is linked to the core through the evaluation process [i.e., peer reviewed journals like IJGIS]....[O]nly a tiny handful [of research articles] pass through the evaluation system from frontier to the core" (pp. 109-110).

4 The methodology described by Koehler (2001) could be applied to IJGIS to quantitatively test this suggestion.

5 Excepting 1998 with a high of 27 book reviews, and 2001 with a low of 4. Was the exceptionally high year due to a burst in GIS monograph publishing? Was the exceptionally low year due to limitations of journal space, or a dearth of monograph publishing?

6 Although Dr. Wu was not the top-cited author he was the most prolific--contributing 3 articles to IJGIS in the period 1998-2002 (Table 3).

7 When counting for the citation identity, it is important to count each unique author only once per citing article. When counting for the citation set (and citation image), it is important to count each unique citing article only once.

8 Cole, J. R. (2000, p. 293) refers to the work by economic historians Robert Fogel and Stanley Engerman, "Time of the cross" (1974), which received large numbers of "negative" (critical) citations but which (ironically) became the catalyst for many major works on slavery in the antebellum South.

9 It is worth noting, however, that when James Pringle, vice president for development at Thomson ISI (who is collaborating with NEC on CiteSeer), was asked whether Thomson ISI planned to archive any of the "often-evanescent research on the Web," he indicated that "the company doesn't see archiving as its role. 'Our role is navigation, building access tools'" (Quint, 2004).
Reference List
Borgman, C. L. (2000). Scholarly communication and bibliometrics revisited. In B. Cronin & H. B. Atkins (Eds.) The web of knowledge: A festschrift in honor of Eugene Garfield [ASIS Monograph Series] (pp. 143-162). Medford, NJ: Information Today.

Cole, J. R. (2000). A short history of the use of citations as a measure of the impact of scientific and scholarly work. In B. Cronin & H. B. Atkins (Eds.) The web of knowledge: A festschrift in honor of Eugene Garfield [ASIS Monograph Series] (pp. 281-300). Medford, NJ: Information Today.

Cole, S. (2000). The role of journals in the growth of scientific knowledge. In B. Cronin & H. B. Atkins (Eds.) The web of knowledge: A festschrift in honor of Eugene Garfield [ASIS Monograph Series] (pp. 109-142). Medford, NJ: Information Today.

Coppock, T. and Anderson E. (1987). Editorial review. International Journal of Geographical Information Systems, 1(1), pp. 3-11.

Cronin, B. and Atkins, H. B. (2000). Introduction: The scholar's spoor. In B. Cronin & H. B. Atkins (Eds.) The web of knowledge: A festschrift in honor of Eugene Garfield [ASIS Monograph Series] (pp. 1-7). Medford, NJ: Information Today.

Fisher, P. (1997). Editorial. International Journal of Geographical Information Science, 11(1), pp. 1-3.

Koehler, W. (2001). Information science as "Little Science": The implications of a bibliometric analysis of the Journal of the American Society for Information Science. Scientometrics, 51(1), pp. 117-132.

Meadows, J. (2000). The growth of journal literature: A historical perspective. In B. Cronin & H. B. Atkins (Eds.) The web of knowledge: A festschrift in honor of Eugene Garfield [ASIS Monograph Series] (pp. 87-107). Medford, NJ: Information Today.

Quint, B. (2004, March 1). Thomson ISI to Track Web-Based Scholarship with NEC’s CiteSeer. Information Today, Inc. Retrieved May 7, 2004, from http://www.infotoday.com/newsbreaks/nb040301-1.shtml.

Schubert, A. (2002). The web of scientometrics: A statistical overview of the first 50 volumes of the journal. Scientometrics, 53(1), pp. 3-20.

von Ungern-Sternberg, S. (2000). Scientific communication and bibliometrics. Retrieved May 6, 2004, from http://www.abo.fi/~sungern/comm00.htm.

White, H. D. (2001a). Author-centered bibliometrics through CAMEOs: Characterizations automatically made and edited online. Scientometrics, 51(3), pp. 607-637.

White, H. D. (2001b). Authos as citers over time. Journal of the American Society for Information Science and Technology, 52(2), pp. 87-108.