E. Other Sources of Citation Analysis Data
An alternative to employing citation indexes such as the SSCI is to examine reference lists of journals, textbooks, scholarly books, and other sources within a field and count the number of citations to a given scholar, work, or journal. This technique has been used successfully by Cohn and Farrington to produce a sizable body of citation analysis research in criminology. Although it is more time-consuming, this method is arguably more accurate and avoids many of the problems that are inherent in the use of the SSCI. For example, it allows the researcher to examine all authors of a cited work, not just the first author, and to exclude all self-citations. It also provides a knowledgeable researcher with an opportunity to correct at least some of the errors found in citation lists. For example, if the researcher sees a reference to a “T. Hirsch,” he or she will be able to determine whether this citation is of a work by Travis Hirschi or that of a different author. This method also allows the researcher to attempt to distinguish between multiple authors with the same surname and first initial, because many journals list first names (not just initials) in their reference lists, or even scholars with the same full name (e.g., the Australian vs. the American Patrick O’Malley or the Australian vs. the British David Brown).
The technique developed by Cohn and Farrington is fairly straightforward, although admittedly somewhat tedious and time-consuming to carry out. For each journal that is used as a data source, they either download the reference pages for every article from an online full-text source into a word processing program or enter the pages into a computer from a printed copy of the journal using an optical scanner. For references with multiple authors, duplicate listings are made of the reference, with each coauthor listed first. Mistakes in reference lists are corrected when found. When all references for a given journal are entered into a computer file, they are sorted into alphabetical order, and this list is examined to determine the number of times each name appears. When known, citations of scholars with multiple names (e.g., Ilene Nagel/Bernstein) are amalgamated. If reference lists did not include first names or middle initials, Cohn and Farrington would, when necessary, check references against original publications to distinguish between, for example, the various D. Smiths (Douglas A., David D., David E., David J., etc.), the various J. Cohens (Jacqueline, Jacob, Joseph, etc.), and the various different scholars with the same name (David Brown, Richard Sparks, Richard Wright, Patrick O’Malley, etc.). In some journals, references occasionally specify “et al.” rather than listing all authors. In these cases, the original works were checked, when possible, to obtain the names of all coauthors and insert them into the data file. To maximize the accuracy of the data, this method requires an extensive and detailed knowledge of criminology authors and publications.
F. Concluding Remarks
Overall, citation analysis is an extremely quantitative, objective method of exploring trends in scholarly impact and prestige. The increasing availability of electronic journals and books throughout the world is helping to make this technique more accessible and may increase its use in the near future. Researchers may in the future use citation analysis to examine how changes in theoretical, empirical, and political issues affect research and scholarly influence. They may also use this method to determine patterns and trends in research topics and perhaps predict changes in key issues over time.