An Active Study

Dec 13 2010 Published by under [Information&Communication]

I ran across an intriguing article in the current issue of AMWA Journal:  Use of the passive voice in medical journal articles. Amdur, Kirwan, and Morris.

The paper begins with a quote from Anitra Sheen:

“Passive voice is the bane of medical writing. It pervades medical literature with the haze and heaviness of stagnant air. Writers sometimes use passive voice in an attempt to make their work sound scholarly and scientific, when actually they are perpetuating a writing tradition that is fraught with ponderous and obscure language."
Breathing Life into Medical Writing: A Handbook. St. Louis: Mosby; 1982:21-22

The authors then discuss the real facts: no data exist on the use of passive voice in medical writing and how it compares to other forms of written communication! They set out to fill this data void by calculating the use of passive voice construction in 3 types of articles from 3 major medical journals. They selected JAMA, New England Journal of Medicine (NEJM), and The Lancet because they have high impact factors and may model medical writing style for other journals. They also have different publishers and editorial staffs, and each issue addresses a variety of topics from a wide range of fields of medicine. From each of the 3 journals they selected 10 articles of each of 3 types, namely opinion papers, review articles, and original research reports. Thus, 90 articles from 2006 issues of these journals were selected randomly for inclusion in this study. For perspective, 30 articles from The Wall Street Journal (WSJ) provided a nonmedical control group. They chose WSJ because articles seemed similar regarding detailed analyses of complex subjects. Titles, headers, abstracts, references, and quotations were not included in the analysis.

Now for the fun part: each sentence in every article was hand-coded for passive or active construction:

Specifically, one of us (RA) evaluated every sentence in each article to determine the number of sentences with passive voice construction. For the first 20 articles, another one of us (JK) repeated the process so that both of us independently recorded the voice as active or passive in every sentence in the article. We then discussed any discrepancies and made a final determination that we both agreed on. After doing the 2-person analysis on 20 articles, there were no differences in our calculations in the next 10 articles, so, for the remaining 90 reviews, only one of us determined passive voice frequency without a double check by the other.

That makes taking daily vaginal swabs on mice seem glamorous! I am really happy that they went to the effort, for they then determined percent passive voice using Microsoft Word's grammar check. Other analyses included use of first person pronouns, and two forms of passive voice construction:

  • Passive voice, doer mentioned:  “Data were collected by a member of our research team from patients with pneumonia.”
  • Passive voice, doer not mentioned:  “Data were collected from patients with pneumonia.”
  • Active voice:   “A member of our research team collected data from patients with pneumonia.”

Figure 1: Passive Voice Frequency by Journal

As shown in the figure, no differences in passive voice use was demonstrated among the 3 medical journals, all of which were substantially higher than WSJ with a median value of 3% passive voice. Use of passive voice did not differ with article type in medical journals. When research articles were studied by section, differences were found among sections, yet all displayed more passive voice than the WSJ sample (next figure).

Figure 3: Passive Voice Frequency for Original Research Article Sections

So what's an author to do? Do we have to hire an English major to fix our writing? Lucky for us, the tools in our word processor estimate passive voice frequency pretty well as shown below:

Microsoft Word vs The Humans

The authors discuss many reasons for high frequency of passive voice use in medical writing, and they make some recommendations:

Use the active voice in place of “doer mentioned” passive constructions.

Limit the use of the passive voice to the Methods section of an Original Research Report because this is a section where describing who did the action distracts the reader from what was done methodologically.

Use passive voice frequency in the overall paper as an endpoint for evaluating the quality of the writing.

Medical journal editors should make a passive voice frequency of ≤10% a publication requirement for all types of articles.

All types of articles in all journals showed plenty of articles meeting the 10% limit, so the authors feel this goal is readily achievable. They also recognize that excessive passive voice construction is not the sole problem with medical writing; however, it is one we have the tools to fix quickly and easily.

Every writing course from middle school on up encourages the use of the active voice. Should this be a journal standard?

*By the way, I ran my text for this blog through the Word grammar check and I logged 9% passive voice sentences. W00T!

7 responses so far

  • Hmm, I use the passive voice because that is what I was taught for science, without exception, in college. Is verb tense in medicine different than other parts of science? Or are the times a-changing?

    • Pascale says:

      I think medicine and science share roughly the same degree of passivity. Grammatical types recommend active voice as a stronger construction, even in scientific and grant writing courses. I believe it is possible to use active voice and say exactly what you want in most scientific papers.

  • Bob O'H says:

    How do the authors justify going from "is" to "ought"? Judgement of style is a subjective thing, and I'm not convinced that changing from passive to active voice is going to improve writing: it's another arbitrary rule that we'll be made to follow, and which will get in the way of actually writing. There are more important problems with the way scientists write: overly long sentences with too many clauses, an inability to write coherent paragraphs, overuse of jargon etc.

    Even worse, people who learned grammar from Strunk & White won't be able to follow this rule, because they won't even know what a passive voice is.

    • Bob O'H says:

      Ah, they don't go from "ought" to "is". this is how they introduce their recommendations:

      However, writers who are concerned about using the passive voice too frequently need more detailed instructions. We recommend the following.

      IOW, the recommendations are only for those who want to reduce the use of the passive.

      • WhizBANG! says:

        They do make those recommendations for authors "who wish to reduce passive voice" but end on a stronger note, specifically directed at editors and journals:

        The real question is not whether
        authors of medical journal articles
        use the passive voice too frequently,
        but why prestigious medical journals
        routinely publish articles that would
        be shorter, clearer, and easier to read
        if the author or an editor revised the
        manuscript to minimize passive voice
        constructions. Several veteran editors
        attribute the problem to a change in
        the role academic publications play in
        modern society and the workload of
        journal editors:

        “It seems that many
        medical articles are written to be
        published and cited, but not to be
        read.”2 In addition,
        Editors say that they try to keep
        things simple, but are overwhelmed
        by the sheer volume of articles written
        in bad English. To stem the flow
        many of them write editorials urging
        simpler English, or run training
        sessions teaching likewise. But
        nothing changes, which means that
        we need to look further than what
        appears to be the current assumption,
        which is that pompous medical
        prose is a kind of infection that
        can be cured with a quick dose of
        ‘common sense.’15

        Excessive use of the passive voice
        is not the only problem with modern
        medical writing, but it is a well-defined
        problem with a simple solution: medical
        journal editors should make passive
        voice frequency a standard for publication.

        I found the study interesting, simply because every writing instructor in my life has fiercely advocated the active voice, and I heed this advice when I write. Is the passive voice inherently bad, even in biomedical and scientific writing? Hmmm, I can think of ways to test that hypothesis...

        • Bob O'H says:

          It would be interesting to test that. Hm, I wonder if we can find some test subjects - I suspect we could get some linguistic assistance from Language Log.

          I don't find the passive intrinsically bad, but I can see how it can lead to boring writing. From the passage you quote it looks like the authors are arguing that something should be targeted because it's easy to target, not necessarily because the targeting will lead to better writing.

          (BTW, that last sentence isn't great, is it?)

  • [...] word. The first readability indicator, the percentage of passive sentences, has been addressed in earlier posts. Flesch Reading Ease purports to measure just that, with higher scores being easier. The [...]

Leave a Reply