[Riverheart]

Not nearly on the level as what is being suggested but my company has had several anonymous surveys and I started thinking about writing style when taking them. If you're prone to certain phrases, words, use of contractions or lack thereof, especially when the pool of people is small and you're providing critical (but needed) criticisms, you could potentially be identified by your immediate supervisor. Introducing typos and avoiding phrases you commonly say, adjusting your "tone" is a lot of effort when you can just disengage entirely and/or behave like everything is public (which it may as well be at this point).

[Agentlien]

Most "anonymous" surveys I've been asked to take through work require listing more than enough information for unique identity. One assured I would be anonymous, then asked me to fill in the name of my manager, my team, and job title.

[Riverheart]

Fortunately mine have not but at a certain point they're useless because no matter no low the scores go nobody in their right mind wants to provide long-form feedback to identity actionable fixes because product teams are usually small even if there are a lot of developers in the pool your pain points will be unique to what your working on.

[Agentlien]

Yes, this is usually my experience as well. What makes sense for you to bring up identifies who you are. Hardly anonymous. But sometimes I've also been asked to explicitly identify myself as mentioned yet it's still supposedly anonymous.

[ziml77]

We can give anonymous feedback about others where I work. We can submit it at any time about anything, positive or negative. I have never touched it despite knowing that HR doesn't get my name. It's not hard to figure out who's submitted a piece of feedback from their writing style and the specific situation you're writing about. Like if I were to give feedback related to working on a project with one other person, any sort of specifics about the project would make it very obvious that it was me writing the feedback.

[notsoanonynous]

I am open to ideas for how to mitigate this remaining vulnerability even further

[nisegami]

One idea I've seen is running through translation services. IE, convert to spanish and then back to english. But unless we have good offline services, it defeats the point.

[sndwnm]

Maybe not very practical, but to combat targeted writing analysis on the internet you could try running such analysis software on your own writing to find out what makes it stand out. Work to make the writing as "bland" as possible, perhaps with aid of software translators or filters.

[yosame]

Maybe you could run all your communications through a translator twice (e.g. English -> French, French -> English), and fix any typos?

It would hopefully keep the sentiment while changing the words.