Do The Test.

[Rummy]

I would actually say this case is more out of the ordinary than that, in that you are specifically filtering out an attribute that the bear has, it's colour. Colour is quite a primitive and high up method/mode of selection, and I have no doubt this experiment would be much more different if the bear was in pink, brown, yellow or pretty much any colour but black. In fact, I'm fairly sure I was told it was replicated, but with the black team doing the passes that you're asked to count, and in that condition alot more people see the gorilla. While it's technically selection, I'd rather call it exclusion, you're specifically filtering out black because you've already concluded any black you see is the other team's players. When you're reading a book, or watching TV, you've selected to pay attention to the TV, but not to exclude anything specific, if you get my point?

The only other interesting effect I could offer that contradicts this is the name/cocktail party effect, where other people saying your name can have interesting effects on what you're doing. It's incredibly hard to ignore your name apparently.