A corpus study of the term evidence in open peer reviews to research articles in the British Medical Journal

Abstract

The linguistic study of peer-review discourse has focused principally on pre-publication occluded referee reports. However, there are few studies on post-publication open peer reviews of research articles. To address this imbalance, we analyse a type of open peer review, Online Rapid Responses (ORRs) to articles, in the British Medical Journal (BMJ), which is the leading medical e-journal. Using a corpus-based approach, we focus on the term evidence owing to its importance in scientific discourse. We compiled an ad-hoc corpus of 875 ORRs (260,651 tokens) and analysed it using Wordsmith Tools 6 to ascertain the frequency of evidence. We then compared its frequency in our corpus with the British National Corpus (BNC), the Corpus of Contemporary American English (COCA), the COCA academic subcorpus, the Cambridge Academic English Corpus (CAEC) and the sub-corpus of reviews in the Lancaster-Oslo-Bergen Corpus (LOB-C). We also performed a keyness analysis of our corpora to ascertain the position of evidence and obtained the contexts in which it appears. Our analysis reveals that evidence is more frequent in our corpus of ORRs than in general and academic corpora, which highlights its importance in the evaluation of research. Our exploration of its contexts of use show that it reflects the concern of the medical academy for evidence appraisal in state-of-the art medicine.