Does anyone have any good advice for obfuscating one's writing style when reviewing papers? I have a fear that I would inadvertently use the same turn of phrase in a referee report as in some published work, thus revealing my identity, and I'm curious if there is a good general practice to follow to mitigate this as much as possible.
It's not worth it. Few waste time trying to unblind their reviewers to take retribution for bad reviews. It's just not worth it. The average journal submitter doesn't have enough power to do damage to a random reviewer.
If you need a strategy, trying passing your text through Google Translate to Spanish and back to English. It'll probably be garbage afterwards, but it'll definitely lose any idiomatic turns of phrase that might identify you.
Assuming that you really want to do this, here are some strategies that should not strongly diminish the readability of your review:
- If you are from an English-speaking country: Use the spelling conventions of another English-speaking country (e.g., British instead of American English).
Switch to a strongly different punctuation convention, e.g., the French one:
On page 2, the authors write : « We could not find any evidence for this. »
If you typically tend to write long sentences, split sentences up as much as possible. If you tend to write short sentences, make them slightly longer than you’re comfortable with.
Try to use certain words particularly often, e.g., particularly. On the other hand, if you are particularly fond of certain words, avoid them entirely (of course, the main challenge is to become aware of this in the first place).
If your field has two competing notations for something, use the other one.
Unless this is a feature of your native language, drop articles, in particular if this does not inhibit understanding your review. On the other hand, if your native language is not particularly fond of articles or does not have them at all, use an article wherever possible.
After you wrote your review, try to replace every rarer, non-technical word by a synonym found in a thesaurus.
To make you seem German, capitalise some arbitrary nouns, but do not capitalise adjectives derived from proper names:
I disagree with the use of bayesian Statistics.
To add to Bill's answer, most scientists would agree that it is not productive to try figuring out who a reviewer may be, and consciously avoid thinking about it too hard. I make that a point when talking to my students and postdocs. I'm not sure most are that explicit about it, but I've never spoken to anyone who actively tried to find out who a reviewer might be, but I've spoken to many who agree that it's the wrong thing to do.
The whole point being: nothing good can come of it if you know who your reviewers are. But you can violate the spirit of peer review and alienate colleagues if you try too hard. So just let it go.
For you, this means: don't try too hard to obscure your identity.
I would suggest that you forget your current train of thought. Openly sign your review and write it as if you would tell it to the author's face. If your colleague wrote a bad piece, then help him by pointing out how it could be improved, in a way that does not make him lose his face. If you are working in a falsifiable field and he made an objective error, then explain that error in an objective way (stating facts etc.).
Your new problem is then to figure out how to write your reviews like that, and there are plenty of techniques for that (there should be plenty of resources out there about how to put out criticism without the hurt factor). This is much more healthy and probably also easier than obfuscating your identity.
If you are really serious about maintaining your anonymity, you should keep in mind that not just the style of your writing, but also the content of what you write, contain major clues as to your personality and hence (given how small your field is) to your identity.
I therefore suggest choosing a random subset of your recommendations to the author(s) and flipping them: e.g., if you were going to suggest making the paper longer, tell them to make it shorter instead; if you thought of suggesting to collect more data, tell them they have too much, etc.
Finally, for extra safety, also flip the accept/reject bit. With this technique, your anonymity will be virtually guaranteed.
If you really want to do this, become a better writer. Seriously. Take writer training classes, including the ones aimed at fiction writers. Read books on writing style, non-fiction as well as fiction. Study writing styles. When you read something, take some time to note the phrases, grammar, etc., not just the content. Understand how words and grammar bring meaning across.
Also try some role playing. Go to impro theater classes, or play a RPG. Every now and then, imagine being somebody else, and talking and writing like them.
Once you have developed a sufficient understanding of prose to identify how exactly your style differs from others, you will be able to create one or more personas with a distinct writing style. Write your reviews as this persona.
This requires quite a bit of commitment and you will need to learn considerable skills in that area. On the plus side, you will not only be able to hide your identity, but you will also gain a lot of mastery in writing, which is a very valuable skill for a person judged on the quality of his written publications.
It is up to you to decide if the considerable effort is worth it.
To review in a less personal way, I avoid the "I", and use the third person writing. I thus write "the reviewer suggests", "to the reviewer". Accordingly, I write "in the paper", "The authors". This helps me take a more distant look at the work under review.
I learned this from a colleague, received a review in this style. Thus, at least three reviewers use this technique (I won't be detected so easily).
Anonymity is ultimately about separating yourself from your speech. Use a standard and concise style that avoids all biases, irrelevance to the purpose of the work under review, and unprofessional language or conduct.
Doing so will additionally elevate any intellectual endeavor and need not be done for privacy purposes alone.
You could use an automated paraphrasing tool such as Spinbot.
The text will need a little bit of editing afterwards... here's what it made of your question:
Benefits anybody have in any way guidance for jumbling one's written work style when assessing papers? I have an apprehension that I would accidentally utilize the same turn of expression in a ref report as in some distributed work, consequently uncovering my character, and I'm interested if there is a decent broad practice to take after to alleviate this however much as could reasonably be expected.
Many years ago my colleague was asked to review an article written by a well known person in our field of research, who was familiar with the writing style of my colleague.
There were serious issues with the paper and my colleague intended to write a very negative review but wanted to maintain anonymity.
We discussed the issues and I wrote the review. Problem solved.
The article was rejected, but later published in another journal.