Possibly. However, I've found that it's fairly easy to identify who is saying what based on writing style, grammar and punctuation habits. Any HR person will tell you that's how they identify respondents to "anonymous" employee surveys.
This is a good point. To date I have not had an anonymous survey in a small group that hasn't bitten me in the ass afterwards.
I also find some anonymous surveys interesting when they are placed in my mailbox and come with a unique identification number. When asking about this uid, I am always told it is for quality control purposes or such. Yeah right.
Anyway, the proposed idea is interesting. I could see it being abused though. Someone from management, A, who wants to know what you really think of them could set up a fake discussion using several other sign ins from coworkers B C and D (and of course management has their passwords, that's required per the corporate contract), puppets who then ask "What do you think of boss A?" followed by puppet accounts B-D saying A is a jerk and then listening to see your response. This is the sort of thing that happens with the sort of jock management that infects many workplaces.
Perhaps you can try translating the message to another language and back. Then present the resulting text to the user for corrections. The resulting grammar and punctuation could be very different than the original.
You could probably strip all punctuation and capitalization from any user input so that the style remains fairly obfuscated. IMO it's punctuation and capitalization that usually gives away the IDE identity especially on a chat medium.
I suppose it depends on the size of the company, but I can't imagine any HR person at a co with >20 ee's that could identify someone based on their writing style unless it was particularly idiosyncratic.
I imagine the answers would allow you to shrink the sample down to a few select individuals. E.g. This anonymous survey complains a lot about a supervisor, let's start with his/her direct reports.
How about each message has to be rephrased by another person in the room in order to scramble the original author. It suffers from the telephone problem and general misunderstanding however.
Agreed that you can probably find commonalities within the individual comment threads, but I feel like you'd be hard pressed to do it within a large group.
Though I guess those semantic specialists from HR could do some damage.
Could you elaborate on this with some real world examples? Unless it is something very obvious like using "..." or "-" a lot while nobody else does, I am wondering how this can be done reliably and efficiently especially considering that what HR knows about your writing style is most likely only from business emails, your application and other business correspondences where one would likely use a different writing style than in an internal anonymous discussion.
Good point - then again, how many 66-word sentences do I usually write and is that enough to "identify" me? And see bugsy's reply, there is a looong sentence there too.
