Wednesday, August 29, 2018
Correcting for the problem of multi-author articles cited as "John Smith et al." in citation studies
Ted Sichelman (San Diego) contacted me with a proposed solution to the problem of undercounting multi-author articles in citation studies, a problem that washes out at the school level, but not necessarily at the level of individual authors, as noted before. Prof. Sichelman here explains what he did:
Using the HeinOnline citation data, which does not suffer from multi-author limitations, I examined every 3+-author article with a large number (70+) of all-time citations in HeinOnline (as of late 2016), because articles with fewer citations are very unlikely to have any sizable effect on the most-cited lists (which I confirmed for a medium-sized sample set). For each second and later-listed author of these highly cited articles, I estimated a high number of "missing" citation counts in Westlaw's "law review & journals" database for the period 2013-17. Then, using this high-estimate missing count, Brian examined the raw data from Sisk et al. to determine if any author in the Hein-generated list might make a most-cited list (who didn't originally) or might substantially move up a list in ranking. Next, for each of the authors Brian flagged, I searched Westlaw to determine the missing (Westlaw) citation count (using the Sisk et al. methodology) for the period 2013-17 in two phases (providing information to Brian in each phase), generating accurate counts of missing citations for flagged authors. For completeness, these final counts included missing Westlaw citations not only for highly cited articles, but all articles in HeinOnline with more than 10 citations (as of late 2016) published since 1995 (other than for Mark Lemley and Cass Sunstein, because the additional citations for their less highly cited articles would not materially increase their cite counts). If you have any questions on the methodology, please feel free to email me ([email protected]).
I am grateful to Prof. Sichelman for undertaking this and sharing the results with me. I will be updating some earlier rankings (no dramatic differences, but some slight ones) and incorporating this data into the specialty area citation rankings to come.