I'm not sure you do see my point! The results for the other emergency service personnel would be irrelevant, but by including them and not revealing the intention of the survey you are making it more accurate.
What I mean is people are likely to be better at identifying who is real and who is not when they're asked "Right, who is real and who isn't?". It's like being asked "Who's the odd one out?" - people may not have noticed any difference at all if they were not prompted that there were differences...
I'm somehow stumbling over my words here, but what I'm trying to say really is that the question and the content in the survey by its very nature are going to raise the chances of people being able to identify differences in comparison to an every day situation.
Apologies if you actually get that already, it's just your post makes it out that you don't
