Questionable validity of Poisson assumptions in a combined loglinear/MDS mapping model

✍ Scribed by Gleason, John M.

• Publisher: John Wiley and Sons
• Year: 1993
• Tongue: English
• Weight: 315 KB
• Volume: 44
• Category: Article
• ISSN: 0002-8231
• DOI: 10.1002/(sici)1097-4571(199301)44:1<57::aid-asi6>3.0.co;2-v

✦ Synopsis

The use of a Poisson model as a generating process in the case of journal citations is not justified in the absence of an empirical data analysis which supports its use. The underlying assumptions of the Poisson process are discussed with respect to characteristics of the citation process. The citation process appears, a priori, to violate all of the Poisson assumptions.

Due to the apparent violation of the Poisson assumptions, the use of a Poisson process in the recent development of a combined loglinear/MDS model for mapping journals by citation analysis does not appear to be justified, and this raises serious questions regarding the credibility of the results of that analysis. A more comprehensive discussion of the assumptions on which the recently proposed model is based, along with an empirical analysis which supports the use of the Poisson generating process, would better permit the reader to evaluate the credibility of the Poisson-based model.