Citing reprinted material.

Azar, Ofer H.

Occasionally articles from peer-reviewed journals are reprinted as chapters of edited books, and authors sometimes cite the book chapter instead of the original journal article. Because the edited book cites the original articles that it reprints, usually the author should not find it difficult to cite the original articles instead of the book. Nevertheless, often the book is being cited, possibly because authors do not realize that citing the original article is more appropriate. I claim that in general, the original article should be cited, for several reasons.

First, the reader who is interested in the cited material is much more likely to be able to find a journal article than a book because academic libraries subscribe to most of the important journals, but purchase only a fraction of available books. Second, journal articles are often available online, thus saving the reader the time required to go to a library and find the book, allowing him to work also in locations in which he does not have library access, offering him the option to save a copy of the article on his computer and to search electronically in the full-text, etc.

Third, citing the journal publication gives the reader important information that can help him to decide whether he wants to spend time to obtain the cited article and to read it. The quality of the journal in which the article appeared, for example, is a signal about the article's quality. Similarly, the date of publication of a journal article gives a good idea about how updated the material is (a book's date of publication, however, does not indicate when the article was first published, and is therefore not particularly useful in this respect).

Finally, it is fairer that whoever published the material first (i.e., the journal) will enjoy the credit when it is cited. This is important not just because of fairness issues per se, but also because of the incentives that the citing practice creates. Journals that find that they lose many potential citations to books that reprinted their articles may be reluctant to permit such reprinting in the future, because the number of citations a journal receives affects its perceived quality. Most rankings of journals are based on the number of citations that the articles published in the journal generated. (2)

The number of citations that a book, which is based on previously published articles, can "steal" from the original journals can be quite large. For example, Richard Thaler's book, "Winner's Curse: Paradoxes and Anomalies of Economic Life" (Thaler, 1992), was cited in the journals indexed by ISI Web of Science 202 times (as of December 2005). Since this book's chapters are all articles that were previously published in the Anomalies section of the Journal of Economic Perspectives (over the years 1987-1991), the JEP lost a substantial number of citations by allowing the book to reprint the material (assuming that many authors who cited the book would have found and cited the original article if the book had not existed). Because reprinting previously published articles in books is desirable from a social perspective (it helps to disseminate knowledge further), it is important to adopt a citing practice that encourages journals to permit such reprinting, thus providing another justification for the proposed practice to cite the original article rather than the book chapter.

If it is so desirable to cite the original article, why do authors often cite the reprinting book? There are several possible reasons. First, authors might be unaware of the advantages of citing the original article. Second, journals do not require any specific practice in this respect. I have read over the years style guidelines of dozens of journals, some of which were very long and detailed, but I do not remember seeing even once an instruction that says whether the author should cite the original article when he encounters a reprinted version of it.

Third, certain people might want to publicize the book. For example, the book's editors, as well as authors of additional chapters in the book (other than the chapter cited), have an interest to cite the book in their papers so that readers become more familiar with it. Finally, sometimes authors might find it easier to cite the book (assuming this is where they found the material). This is the case, for example, when the journal article cannot be easily located because it is an old article not available online, or because it appeared in a journal to which the author's institution does not subscribe. It is also easier to cite the book when the book chapter is a revised version of the original journal article, because citing the book in this case saves the author the time required to verify that the points he wants to cite based on the book chapter also appeared in the original article.

The above are potential reasons why authors might cite the book rather than the original article, but they do not mean that from a social perspective citing the book is welfare increasing. This is obvious with respect to the first two reasons. The third reason suggests that the author might derive a benefit from citing the book, while the fourth reason implies that he might incur a cost when citing the journal article. This means that from a social welfare perspective we have a trade-off between the welfare of the reader and that of the author. However, with respect to the last reason (the costs associated with citing the journal), it seems reasonable that the benefit to many readers outweighs the one-time costs to the author, and therefore if the third reason does not apply, citing the journal article is probably still the welfare-maximizing solution. (3)

There are, however, certain circumstances in which citing the book entails certain benefits even for the readers. If these benefits are large enough, they can outweigh the advantages to cite the journal article discussed at the beginning, and then citing the book might be socially desirable. One such case is when the book's chapter is a revised version of the journal article. Because the book's chapter was written later, presumably it was improved compared to the original article, for example by adding new material. It might also be the case that because journals are more space-constrained, the journal published a shortened version of the paper, while the book published the full-length version. If only the journal article is cited in these cases, the reader might be unaware of the book chapter, which he may in fact prefer to read. A similar case happens when the book chapter is a translated version of the original article, in particular when the book is in English. Citing the English book can improve the ability of most readers to read the material.

Another case in which readers might prefer a citation to the book is when the journal article is not posted online and is harder to find than the book. This is sometimes the case with particularly old articles. Finally, it might also make sense to cite the book when it contains a collection of articles by the same author, and several of these articles are relevant to the citing paper. Then it may be more appropriate to refer the reader to the single book where he can find all the articles, rather than to several different journal articles.

I hope that this article will encourage authors to cite the original articles and not their reprinted book versions in the vast majority of cases, with the few qualifications mentioned. This will improve significantly the usefulness of these references to the article's readers. Journal editors can help in making citations more useful by asking authors to cite the original articles except in the specific cases mentioned, in which citing the book is more appropriate.

References

Kalaitzidakis, Pantelis, Theofanis P. Mamuneas, and Thanasis Stengos (2003). "Rankings of Academic Journals and Institutions in Economics," Journal of the European Economic Association 1(6): 1346-1366.

Thaler, Richard H. (1992). Winner's Curse: Paradoxes and Anomalies of Economic Life. New York: Free Press.

by Ofer H. Azar *, (1)

* Ben-Gurion University of the Negrev

Notes

(1.) Ofer H. Azar, Department of Business Administration, Guilford Glazer School of Business Management, Ben-Gurion University of the Negev, P.O.B. 653, Beer-Sheva 84105, Israel. E-mail: azar@som.bgu.ac.il, Tel.: +972-86472675, Fax: +972-8-6477691. I thank Michael Szenberg for helpful comments.

(2.) Several adjustments are often made in order to create a more sophisticated ranking that presumably reflects quality better. For example, the quality of the citing journal might also be taken into account. Sometimes, rather than computing the total number of citations, a journal impact factor is computed by counting only citations of articles published in a certain period, and dividing the number of citations by the number of articles published in the journal during that period (a recent ranking of economics journals and a description of the methodology used appears in Kalaitzidakis, Mamuneas and Stengos, 2003). Regardless of the exact ranking methodology, however, as long as its basis is citations, any citation (during the relevant period) in which a book was cited instead of the original journal article lowers the ranking of that journal.

(3.) It is hard to make a similar argument with respect to the third reason, because in this case, the author's benefit from citing the book is proportional to the number of readers of the paper, so the point that there is one author (or a few) and many readers does not necessarily imply that the welfare-maximizing solution is to cite the journal.