Investigating predatory publishing in political science: a corpus linguistics approach

This article explores the application of corpus linguistics methods in dealing with an underexplored area concerning predatory publishing, with a focus on lexical bundles and formulaicity. Using a comparative approach, the study employs two corpora of more than 1,6 million words, consisting of 220 research articles drawn from two comparable journals in the field of political science, one predatory and one top-ranking. The results show that writers publishing in the top-ranking journal use a more limited range of lexical bundles with a higher frequency, giving further evidence for the highly formulaic nature of the genre. The two groups of writers also display different preferences for lexical bundles with particular functions and/or forms. While the top-ranking journal articles feature more disciplinary-specific bundles with noticeable variation across the main sections of the research article, the predatory journal articles highlight in particular a set of common-core lexical bundles typical of general academic language use. Our findings also demonstrate the potential of lexical bundles in revealing the amount of scientific information research articles contain as well as the level of scientific literacy of the authors.