Assessing Obliteration by Incorporation: Issues and Caveats

Empirical studies of obliteration by incorporation (OBI) may be conducted at the level of the database record or the fulltext citation-in-context. To assess the difference between the two approaches, 1,040 articles with a variant of the phrase “evolutionarily stable strategies” (ESS) were identified by searching the Web of Science (Thomson Reuters, Philadelphia, PA) and discipline-level databases. The majority (72%) of all articles were published in life sciences journals. The ESS concept is associated with a small set of canonical publications by John Maynard Smith; OBI represents a decoupling of the use of the phrase and a citation to a John Maynard Smith publication. Across all articles at the record level, OBI is measured by the number of articles with the phrase in the database record but which lack a reference to a source article (implicit citations). At the citation-in-context level, articles that coupled a non-Maynard Smith citation with the ESS phrase (indirect citations) were counted along with those that cited relevant Maynard Smith publications (explicit citations) and OBI counted only based on those articles that lacked any citation coupled with the ESS text phrase. The degree of OBI observed depended on the level of analysis. Record-level OBI trended upward, peaking in 2002 (62%), with a secondary drop and rebound to 53% (2008). Citation-in-context OBI percentages were lower with no clear pattern. Several issues relating to the design of empirical OBI studies are discussed.