Comparative analysis of epistemic stance in abstracts of published biomedical research and associated National Institutes of Health funding applications (1985–2020)

Abstract Research funded by the US National Institutes of Health (NIH) exerts considerable influence over the trajectory of biomedical science and healthcare policy and practice. Here, we extend previous research by assessing the relationship between the expression of epistemic stance (i.e. confidence in propositions) in successful NIH funding applications and the subsequent research publications. Analysis of 140 stance features (modal verbs, hedges, and boosters) in all PubMed abstracts describing NIH-funded research during the period 1985–2020 identified trends that broadly mirrored those previously reported for abstracts of the associated funding applications. We argue that trends, in part, indicate that investigators adopted a stance that became less cautious and less tentative (e.g. consistent declines for appear, seem, and probably), and increasingly confident, assertive, and empirical (e.g. consistent increases for highlight, likely, and typically). Discussing our results in relation to changes in the biomedical research system, including growth in promotional writing strategies, we suggest that increasing salesmanship in the NIH research system is in part a downstream effect of funding mechanisms. We also infer methodological and pedagogic implications for analysis of metadiscourse.