The American Sociologist, 9 (3) p.164-165, 1974

Letter to the Editor


Eugene Garfield

It is readily apparent that Daryl Chubin (American Sociologist, November, 1973 :187-191) did not know of the existence of the Social Sciences Citation Index (Weinstock, ISI's Social Science Citation Index: A New Comprehensive Multidisciplinary Information Retrieval System for Social Science Literature. Presentation at American Society for Information Science Annual Convention, Washington, D.C., Oct. 24, 1972) at the time he submitted his manuscript on the use of the Science Citation Index in sociology.

The availability of the SSCI covering the 1972 and 1973 worldwide sociological literature overrides Chubin’s concern that coverage of the sociology literature in SCI is inadequate. Clearly it is the SSCI, not the SCI, which should now he the reference tool of choice in evaluating sociology publications. (See Garfield, "The Information Revolution Reaches the Social Sciences" in Current Contents 2 Jan. 10, 1973:5-7; and "Current Contents/Social and Behavioral Sciences—A Component of ISI’s Total Information System for the Social and Behavioral Sciences" in Current Contents 45, Nov. 7, 1973:5-6.)

Chubin makes a fundamental error in stating that the SCI or SSCI ... distorts the citation practices of scientists because "the index magnifies visibility of senior authors by ignoring junior authors." As we have repeatedly asserted, one can easily determine the exact frequency of citation of any author, junior or senior, if the index is used properly. To measure the impact of any author, one needs to first determine that individual’s list of source publications through the Source Index or by any other means such as biographical resumes. Once you have identified every article in which an author has appeared in either senior or junior role, each such item may then be looked up in the Citation Index section of the SSCI. Chubin confirms the insignificance of the "second author effect" in aggregate studies, which is most interesting. But irreparable damage may be done to a particular individual if the artifacts mentioned above are overlooked in citation analysis.

Whether the wide availability of citation indexes will contaminate their use for sociometric purposes should not be a factor as long as scholarly publications adhere to a high standard of refereeing with particular attention paid to proper bibliographic practice.

Institute for Scientific Information
Philadelphia, Pa




I thank Eugene Garfield for his letter since it affords me an opportunity to clarify some of the issues in citation analysis, especially those addressed in my note (American Sociologist, 8 1973:187-191).

In advocating the use of the SSCI, Garfield underscores the evaluation of "sociological publications." Precisely! All ISI indexes are literature retrieval tools, not citation ledgers. Garfield— and the Coles ("Reply to letters on citation analysis" Science 183, 11 January, 1974:32-33) for that matter— deplore the use of citation counts to evaluate individuals. Nevertheless, aggregate SCI counts that correlate highly with several prestige measures (Cole and Cole, Social Stratification in Science, Chicago: University of Chicago Press, 1973) reflect more on the atypical samples studied (e.g., eminent or professorial rank scientists) than on the generalizability of the citation measure to most of the scientific population. Generalizability has yet to be demonstrated. Thus, my argument is not with Garfield, but with the potential misusers of his tools.

As for the SSCI, its Source Index contains complete citations to articles appearing in 1000 "core" journals for the current year only. Although two years are presently indexed, I have been informed that the prospect is good for retroactive coverage of the social science literature (one year back for every year henceforth). I welcome that. However, I doubt that coverage will extend beyond the late 1960’s, thereby rendering the SSCI’s utility in longitudinal research questionable at best.

Still, creation of the SSCI means that mere counting of citations will proliferate in sociological research. And this will simply not do.

Weinstock ("Citation Indexes" p. 19 in Encyclopedia of Library and Information Science, New York: Marcel Dekker, 1971) in presenting the ISI rationale for citation indexing, specifies fifteen reasons for citing the work of others, from "paying homage to pioneers" to disputing priority claims of others." This acknowledgment that citations are of various types and put forth in various contexts points to analyses that ISI has ostensibly left for others.

Such analyses —something more exacting and substantive than counting— should indeed be our concern. Research has finally begun, for example, on the classification of citations in theoretical high- energy physics (Moravcsik "Measures of Scientific Growth" Research Policy 2, 1973 :266-275; "Studies of the Nature of Citation Measures I: Some Results on the Function and Quality of Citations, 1974, mimeo); though Cloyd and Bates (Homans’ Transactional Theory in Footnotes" pp. 93-101 in H. Turk and R.L. Simpson eds., Institutions and Social Change, 1971, Indianapolis: BobbsMerrill) had earlier content analyzed citations to Homans’ works. I find this approach much more satisfying and fruitful for the sociology of science than enumerating citations as a measure of quality, visibility, impact, or whatever.

Garfield’s parting thought on refereeing assumes that "proper bibliographic practice" is a major factor in the editorial decision. Although I reaffirm his hope, data on the reasons for manuscript rejection in one sociological journal (Smigel and Ross, "Factors in the Editorial Decision," American Sociologist 5, February, 1970:19-21) fail to corroborate this assumption.

Cornell University
Ithaca, NY