An Interesting Corpus Linguistics Study (Millar et. al. 2012)

This is a study which produced a corpus of about 1,200,000 words from Randomized Control Trials from 5 of the top medical journals. The methods described in the study provide a practical framework for how to build a written corpus of high quality medical English and compare it with general English in order to gain insights into the characteristics of medical English.

‘Use the active voice whenever possible’: The Impact of Style Guidelines in Medical Journals

  1. Neil Millar1,*,
  2. Brian Budgell2 and
  3. Keith Fuller3

Author Affiliations

  1. 1Department of English, University of Birmingham, 2Canadian Memorial Chiropractic College, Toronto and 3Institute of Communication, Culture and Information Technology, University of Toronto
  1. *E-mail:


Medical writing is sometimes criticized for excessive use of the passive voice. The purpose of this study is twofold: (i) to provide quantitative descriptions of how the passive voice is used in medical journals and (ii) to assess the impact of style guidelines encouraging use of the active voice. From a corpus of 297 primary research articles published in the top five medical journals, we extracted 19,691 passive constructions. Analyses show that guidelines have a significant effect on use of the passive voice, and that this is highly localized in the ‘Methods’ and ‘Results’ sections. Analyses also identify a core set of verbs which are strongly associated with the passive voice, and which play a central role in structuring the discourse. We argue that current guidelines influence author’s linguistic choices, and that although paraphrasing a sentence in the active voice may be possible, a passive alternative is sometimes preferable. Findings demonstrate the need for formative guidelines which better reflect the reality of conventionalized usage.

Applied Linguistics:


Methodologies used in this study have wide ranging applicability. A study examining dermatological case studies is currently is in the planing stages. That study will examine commonly used dermatological expressions such as nodule, papule, lesion etc. and how they are used in dermatological case studies.

This entry was posted in Corpus linguistics, Journal articles, Research methods. Bookmark the permalink.

Leave a Reply