The relationship between readability and scientific impact: Evidence from emerging technology discourses

Abstract

• Readability analysis of 135,502 abstracts pertaining to twelve technologies. • The language of the abstracts has become more complex over time. • More complex texts are more likely to receive at least one citation. • Highly-cited articles of larger discourses tend to be more complex to read. • Scientists may have an incentive to (artificially) make their abstracts more complex to read. This study examines how the readability of scientific discourses changes over time and to what extent readability can explain scientific impact in terms of citation counts. The basis are representative datasets of 135,502 abstracts from academic research papers pertaining to twelve technologies of different maturity. Using three different measures of readability, it is found that the language of the abstracts has become more complex over time. Across all technologies, less easily readable texts are more likely to receive at least one citation, while the effects are most pronounced for comparatively immature research streams. Among the more mature or larger discourses, the abstracts of the top 10% and 1% of the most often cited articles are significantly less readable. It remains open to what extent readability actually influences future citations and how much of the relationship is causal. If readability indeed drives citations, the results imply that scientists have an incentive to (artificially) reduce the readability of their abstracts in order to signal quality and competence to readers—both to get noticed at all and to attract more citations. This may mean a prisoner dilemma in academic (abstract) writing, where authors intentionally but unnecessarily complicate the way in which they communicate their work.