With permission from Iris Young
originally available here
Personal take on how to give negative feedback that doesn’t attack a person, as a self-professed extremely critical reviewer:
First, remind yourself as many times as it takes that you do not know this person. You only know what they’ve presented of their work, and that’s limited by their presentation of it, context, etc. It’s a tiny shard of their actual existence.
With that in mind, approach everything as if (1) you could be the one misunderstanding (because you could!), and (2) there could be very good reasons for any actual oversight or shortcoming in the work on the page (because there can be!) It sounds simple, but I rely on using “I statements” to accomplish this. Examples:
This is a communication process and you’re listening.
Double and triple check that every fact you put down is really a fact. A value statement is not a fact. An assessment of the best way forward is not a fact. Speculation about the person behind the manuscript is definitely not a fact.
Once you’ve culled errant unsupported assertions from your text, look at what’s left and decide how to shape it in a helpful way. This is your primary responsibility – your goal should be to help the authors and editors bring the paper to publication. Somehow, somewhere. It’s still legit to say it doesn’t belong in X journal or that it’s not ready without Y experiment, but that, too, can be framed as guidance down the path toward eventual publication. You’re a scientist – name what would convince you to accept the paper. I’ve recently tested my limits on this and invoked a plea to bring in an additional reviewer, as I felt the author and I were talking at cross purposes and one or the other of us was badly misunderstanding the situation. I need someone with X expertise to resolve this standoff (I only recommend this in the direst circumstances, as it’s going to cost even more people even more time, and worse, delay the authors getting a decision one way or another.)
Finally, when polishing and adjusting overall tone, I take off my reviewer hat and put on a human being hat. How is a human being going to receive this? Have I written something that they can take in? If they’re expecting to be on the defensive, how does this read? I review not as a service to journals but as a service to authors. Every review is a professional interaction with colleagues, and I owe them respect and kindness, because they are colleagues and people. It has nothing to do with the quality of their work.
P.S. signing your reviews is a great way to force yourself to take responsibility for how you make people feel.
It’s not quite up to date but you can read a bunch of mine here if you want more specific examples: http://irisdyoung.github.io/reviews/