If you can’t be kind in peer review, be neutral

Expressive language can make for better reading, but pruning it from peer reviews might create a kinder research culture, say Rebekah Baglini and Christine Parsons.
Teacher grading and correcting work.

Small tweaks to language can make sure feedback is constructive rather than hurtful.Credit: Getty

Like most scientists, we are both regularly asked to serve as peer reviewers: to rank proposals and candidates and to identify flaws that should prohibit a publication of a manuscript or provision of grant support. Journals encourage reviewers to ‘plainly state’ their opinion of a manuscript, and reviews of manuscripts and grants land in authors’ inboxes unfiltered and unedited.

However, some reviews are more hostile than others. Rejection is always difficult, but reviews that use emotive or sarcastic language are often the hardest for recipients to deal with, particularly if they are early-career researchers.

As a computational linguist (R.B.) and a psychologist (C.P.), one of us has experience in dissecting the specifics of language, and the other has a professional grounding when it comes to human emotion. Together, we hope to show how even small tweaks to language can make words wound or express disdain, rather than offer constructive feedback. Here’s how to identify and avoid this type of language.

How you say it matters

Most words convey information dispassionately. For example, the next sentence is a simple statement of fact.

“This project proposal didn’t fulfil the requirements stated in the call.”

But some phrases, punctuation marks and intonations convey an ‘expressive’ meaning: they communicate the review writer’s emotional state. For example:

“Disappointingly, this project proposal didn’t fulfil the requirements stated in the call.”

“This project proposal didn’t even fulfil the requirements stated in the call.”

“This ‘project’ proposal didn’t fulfil the requirements stated in the call!”

“This project proposal didn’t bother to fulfil the requirements stated in the call.”

Reviewers are often asked to write subjective judgements on matters such as a project’s scientific value. In writing opinions, reviewers should try to resist the natural temptation to include expressive words that convey emotion, however subtly. For example, the first sentence below provides a neutral statement of the reviewer’s recommendation, whereas the second equates a missing reference to failure:

“The submission should cite Doe, 2020.”

“The submission fails to cite Doe, 2020.”

Other examples of expressive words include:

• Speaker-oriented adverbs such as surprisingly, obviously and disappointingly. These represent the reviewer’s attitude towards the truth of the statement that follows.

Subjective adjectives such as careless, uninteresting, simple and poor. These reflect personal feelings, tastes or opinions.

• Depreciatory modifiers such as least, (didn’t) even, barely and just. These reflect the reviewer’s intent to minimize the main claim. For example, “this is just a hypothesis” minimizes the status of the hypothesis.

• Implicative verbs2 such as happen to, manage to and remember to. These allow the reviewer to imply something about what was required for an event to take place, without stating it directly. For example, “the applicant didn’t bother to do a thorough literature review” implies that the literature review is lacking because the applicant didn’t take the needed time or care.

In addition to word choice, an unnecessary expressive layer can be added to reviews through stylistic and typographical devices, such as:

• Narrativizing, where the reviewer narrates their experience of reviewing (for example, “At this point, I almost stopped reading”).

• Rhetorical questioning (for example, “Did the authors even read the submission guidelines?”).

• Universalizing (for example, “As anyone/everyone/any expert knows”).

• Speculating (for example, “I bet the outlier observations were omitted”).

• Expressive punctuation, such as exclamation marks or scare quotes (for example, “This is not correct!”).

Each of these stylistic choices shows the subjective stance of the reviewer, without adding to the content of the review itself. Although they might make for more interesting reading, they undermine the goals of objectivity and courtesy in the peer-review process. Emotive language — be it positive or negative — can unfairly bias an evaluator’s decision because it is hard to separate the content from the expressive noise. If a submission is rejected, it should be on the grounds of its problems, not how much it annoyed the reviewer to read it.

Helpful advice on how to bounce back from rejection is plentiful. But there is much less advice for those doing the reviewing. Providing more-detailed language guidelines and feedback could improve peer-review communication, and perhaps reduce bias in the review process.

Reviewers, your task might be thankless and burdensome, but writing in a neutral manner is a hallmark of good science communication. If you can’t be kind in your review, be neutral.

This is an article from the Nature Careers Community, a place for Nature readers to share their professional experiences and advice. Guest posts are encouraged.


  1. 1.

    Ernst, T. Nat. Lang. Linguist. Theory 27, 497–544 (2009).

  2. 2.

    Karttunen, L. Language 47, 340–358 (1971).

Download references

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.