The influence of editorial decisions and the academic network on self-citations and journal impact factors

Abstract There are many means by which editors can inappropriately manipulate journal impact factors, but questions remain as to whether these potentially inappropriate behaviors actually influence these scores to an empirically meaningful degree, and which academic disciplines are most culpable. In this manuscript, we propose a game-theoretic/information-asymmetry model that suggests manipulation is reinforced by a feedback loop that creates incentives for manipulation to spread and for disciplines to specialize in the type of manipulation used. We empirically investigate these hypotheses for four different manipulation strategies; coercive citation, self-serving review articles, editorials, and online queuing. Results show that all four of these techniques are effective, they inflate JIF scores and the h-index, and a significant part of that effect is due to inflated self-citations. We also find journals within disciplines tend to specialize in which technique they most frequently employ. Moreover, we show that disciplines are also interconnected, tied together by a journal cross-discipline content network and that disciplines that share more content also tend to rely more heavily on the same JIF influencing behaviors. Effective policy needs to change the editorial decision calculation by removing the benefits of manipulation; removing self-citations from journal metric calculations drastically reduces those benefits.