What happens when passive voice is banned?

I’ve said before that advice to ban the passive voice counts as a platitude. A recently published study corroborates my point. The authors of that study wrote,

[Advice to ban the passive voice] implies that the active and passive voice can be used interchangeably and, where grammatically possible, the active voice should always be favoured.

Their results show why the implications of such advice are wrong.

Some Background on the Study

The researchers used all 2005 research articles (RAs) reporting on research using randomized controlled trials in the five most influential medical journals. Two of five discourage use of passive voice in their style guides. (I’ll refer to them as the “proscriptive” journals, and the other three as “neutral.”) The researchers also compared the RAs with New York Times articles.

The statistically significant findings include:

  • RAs contained almost twice the number of passives than NYT articles.
  • All RAs included passives, with a wide range among individual RAs:  from 6 to 159.
  • RAs in the proscriptive journals contained fewer passives than RAs in the neutral journals.
  • No differences in passive use between proscriptive and neutral journals for some RA sections: Introductions and Discussions.
  • Differences in passive use between proscriptive and neutral journals for other RA sections: Abstracts, Methods, and Results, with greatest differences in Methods.
  • The frequency of passive voice was negatively correlated with the frequency of “we.”
  • RAs in neutral journals use “we” primarily in Intro & Discussion sections, where “authors choose to ‘announce their presence’ with self-reference is often linked to gaining accreditation and acceptance for their claims.
  • RAs in proscriptive journals use “we” primarily in Methods sections.

Unintended Consequences of Banning Passive Voice

One unintended and negative consequence is that transforming some passives into active voice results in syntactic ambiguity:

  1. (passive) Thirteen patients were enrolled and randomly assigned . . .
  2. (active) We enrolled and randomly assigned thirteen patients . . .

The verb “enroll” can be used with or without a direct object (it can be either transitive or intransitive). That means, in the active version of the sentence above (#2), the reader must resolve which of two potential structures is right — the one in which the writers are the enrollees or the one in which the patients are. Syntactic ambiguity increases cognitive processing time.

Another negative consequence is that transforming passives into active voice results in ineffective organization of information:

  1. (passive) . . . we randomly allocated 10008 adults with head injury and a Glasgow Coma Scale score of 14 or less, within 8-h of injury, to a 48-h infusion of methylprednisolone or matching placebo.
  2. (active) . . . 10008 adults with head injury and a Glasgow Coma Scale score of 14 or less were allocated within 8-h of injury to a 48-h infusion of methylprednisolone or matching placebo.

The noun phrase describing the “adults” includes all of the following:

  • the adjective “10008”
  • the prepositional phrase “with head injury”
  • the coordinated prepositional phrase “and a Glasgow Coma Scale score of 14 or less”
  • the prepositional phrase added to the coordinated prepositional phrase “within 8-h of injury”

While such dense noun phrases are not always easy for other researchers to understand, they are a defining feature of RAs because they are efficient. Journals publishing RAs nearly always have maximum word- or page-lengths.

In the active version of the sentence (#2), the noun phrase is split, with a long string appearing as subject of the sentence and the final additive prepositional phrase appearing after the verb. This arrangement of information increases cognitive processing over the passive version partly because the information related to the Glasgow Coma Scale is split but also because of the front-loading of information before the verb.


The point here is that there is nothing inherently good about active voice. And nothing inherently bad about passive voice. I’ll end by repeating a point I’ve made before:

Any “expert” who focuses on limiting your stylistic choices should be ignored.  Pros have many tools to accomplish their goals. It’s the same with pro writers. (I warned you that it is not easy to become a pro.) Language allows us multiple ways of saying the same thing for a reason. Every style is appropriate in some context – otherwise it wouldn’t exist.  Yes. It all comes back to rhetorical context AGAIN.

You can learn the basics about choosing between active and passive voice in this video tutorial.


Millar, N., Budgell, B. & Fuller, K. (2013) ‘Use the active voice whenever possible’: The impact of style guidelines in medical journals.Applied Linguistics, 34(4), 393–414.

Related articles

Similar Posts


Leave a Reply