Actually, according to the research (carried out by the University of Groningen--but I think the link about it was posted here as well), a bit that I left out is that it seems to be the general "feeling of disorder" somewhat, the results are really interesting, but if your posters look like trash or "feel" criminal, they'd probably have the same effect. For instance, litter or discarded flyers had a similar effect. And, unfortunately, due to the common perception of graffiti in general, people don't really distinguish between 10-second "tags" and the really colourful complicated big wall urban typography, which, I presume, most of us would agree can be a lot prettier than just a dead wall.
On the other hand, it's all about the feeling it provokes.
It's like the big average of tiny subconscious feelings of hundreds of passers-by. So no use complaining they react "wrong" to something (pretty graffiti) because they don't "understand", because understanding takes no part in it. You are communicating with a part of the Machine, not people. And it's a dumb beast, and it can't be better.
However, I would guess that the "lost cat" poster variations might have a very different effect. I mean, actual "lost cat" posters might convey a sort of warm feeling of neighbourhoodness/connectedness? Maybe? Or does it just tell people "this is an area where pets get lost / under a car"?
I think if you subvert a certain theme, like a "lost cat" poster, you don't really change the "averaged subconscious reaction" of the crowd, so to say.
Similarly, subverted "official" looking posters again have a different effect. Also depending on their location of course. A random alley would be rather out of place.
(And also, Cram! You're online again, got my PM?)