Competition between academic journals for scholars’ attention: the ‘Nature effect’ in scholarly communication

Here, we study readers’ choice in a context in which scholar’s attention is drawn to salient attributes of academic papers such as importance or accessibility. An article’s attribute is salient when it stands out among the paper’s attributes relative to that attribute’s average level in the choice set. In our model, scholars may attach disproportionately high consideration to salient attributes of academic articles. This paper shows that, depending on the writing complexity in determining article importance, scientific communication in some research fields exhibits accessibility–salient equilibria in which scholars are most attentive to accessibility and less sensitive to article importance. Generalist disciplines (the social and human sciences) with an abundance of multidisciplinary journals which publish research in several fields can be described in this way. In other academic disciplines, scholars are attentive to article importance and are to some extent insensitive to differences in accessibility. There, journals compete on article importance, which can be over-supplied relative to the efficient level of a scholarly paper. One academic discipline with an abundance of highly-specific journals within a sub-field of physics/mathematics/engineering can be described by such equilibria. We also explore the possibility of radical change in scholarly communication when the use of writing complexity in determining article importance changes drastically, whereby a journal acquires access to a revolutionary system of determining articles whose importance is at a much lower writing complexity that its competitor journal. There, when the marginal complexity in determining importance is low, a large improvement in importance entails only a small decrease in accessibility. This allows the academic journal to set a salient high article importance and to win the scholars’ attention, which is named as the ‘Nature effect’.