Bibliometric Study: Remote Sensing of Environment

The On-line Medical Dictionary defines bibliometrics as “the use of statistical methods in the analysis of a body of literature to reveal the historical development of subject fields and patterns of authorship, publication, and use.”  Thus, bibliometric methods can provide a description of the publication patterns within a subject field or other body of literature.

Bibliometric methods such as those that analyze the publication history of a journal, or more accurately the journals within a subject area, are useful in determining the growth patterns of the subject area itself.  If patterns exist with many journals within a subject area such as gains in size or changes in scope throughout their history—assumptions can be made about the scope, growth, and importance of the subject area.  If many journals split or combine into different publications over a period of time, then assumptions can also be made about the narrowing nature of the subject area.

Analyzing the authors of a journal or even the authors within the specified subject area can be used to make similar comparisons and correlations within the data sets.  If many authors are published within the genre, a researcher could come to the conclusion that the field isn’t dominated by one group of elite people.  If papers are commonly written with a large amount of co-authors, assumptions can also be made about the socialization of the researchers in the subject area.  Many co-authors would indicate a large amount of teamwork or research teams within the subject area.  A good mix of different authors from different geographic areas would indicate that the journal and the subject area are of global concern.

Of course, there are limitations to bibliometric studies and their usefulness in gaining an understanding of the patterns of scholarly communications.  Maybe the biggest limitation to citation studies is that the thoroughness in the collection of data is limited to the database used to collect the data.  In an ideal situation, this could be looked at as sample data (much like the sample data collected for the Census), but the database representations for journals in relation to subject area and geographic affiliation (of both authors and country of publication) is not always a good representation of what is actually published.  Complete runs of journals are also not always available.  This lack of coverage can distort the actual publication history of the journal, authors, and publishing countries.  Correcting the distortion would be a lot of work if done hand.  Problems also exist with author names.  Misspellings, name changes, and the use or non-use of the middle initials can affect the retrieval results for the authors.  This is especially disadvantageous to women researchers who elect to take their husband’s last name at some point during their career as a researcher.  Multiple searches need to be done for those who are not consistently listed with the same name throughout their career.  Knowledge of the actual subject area is almost definitely needed to accurately perform the gathering of much of the bibliometric data.  By knowing the history and current culture of the subject area, the collector of data could make educated guesses as to the relevancy of some of these inconsistencies within the data.

Another potential problem with practiced bibliometric methods is that they study how many times an article is cited.  This count does not take into account how the author or article is cited.  If an article or author is cited in a negative manner (i.e. people disagree or refute what the author or study is reporting), then the actual importance or prominence of the author is misrepresented.  An author could have been cited hundreds of times for a piece of faulty information that other researchers are analyzing (and citing).  This bad citation could misrepresent not only the prominence of the author, but also effect the author’s citation identity, set and image—falsely grouping the author with (hopefully) more credible researchers.  This counting of citations does not even always accurately depict the importance of an author within a field.  There is the rare occasion in which someone as prominent as a Nobel laureate has a low amount of citations compared to a peer whose work wasn’t as important or influential (Cole 2000).  Although one would think this would change, post-honor, over time.

            Webometrics attempts to apply the principles of bibliometrics to the hyperlinks of the World Wide Web, in order to examine how the Internet and its pages are searched by users.  Currently, webometrics is capable of being useful in tracking many aspects of scholarly communications.  Webometrics can track the popularity or unpopularity of particular links.  This tracking could indicate subject trends in research--this would allow for certain hot topics to be expanded.  For academic journals, the editors may look at this data and solicit (or at least pay closer attention to) papers related to this area.  This would allow for a more user-centered demand publication.

            Webometrics is also subject to some of the same reliability risks as traditional bibliometric methods.  Coverage would be a major concern to those conducting webometric studies.  In the case of the web, different search engines do not index the same pages.  In addition, different engines need to be searched differently (Borgman 2000).  This is much like the coverage and author issues that exist in searching ISI, except ISI dilemmas would seem to be easier to fix.

            It seems for webometrics to be really useful in gathering information about patterns of scholarly communications; it needs to be invasive upon the users utilizing the resources.  The gathering of usage logs is useful in measuring the traffic of a website, but doesn't give much insight towards the understanding of the patron’s behavior patterns.  This type of information gathering could be comparable to gathering in house use statistics within a library--just because a book is out of place (on a table, on a re-shelving cart, etc.) doesn't mean it has been utilized.  The book could have actually been used as a prop for a laptop.  Future webometrics data could track a multitude of behaviors such as whether or not the accessed resource was saved to a disk, printed from a workstation, emailed, or if nothing at all was done to the resource after it was accessed.  Furthermore, it could show for how long the resource was accessed.  The length of access could show that the scholar read the article at the time and place of access.  The use of monitoring tools could provide such details by gathering the necessary webometric data.

            This sort of data--how people physically use the information they access could then be examined in correlation and/or comparison with citation data.  Researchers could examine if the information accessed is the same information that is cited within the data's literature.  Researchers would then be able to better assess what kind of literature is used as merely a jumping off point for scholarly studies and what kind of resources are utilized as more of the backbone of the literature.

            The digital environment provides for many opportunities to apply bibliometric and webometric methods so a better understanding of scholarly communication can be attained.  These methods and studies all have their limitations in accuracy and reliability.  It is important to consider these limitations while gaining the unique insight that these methods have to offer.