I've always considered The Mirror Test a deeply flawed experiment because it relies on a number of assumptions that are difficult (in my non-expert opinion at least) to guarantee.

For examples, babies are tested with a sticker, it's assumed that a baby's lack of motivation to remove the aforementioned sticker is proof that they baby is not yet self aware. Yet that's assuming that the child isn't just curious about his or her own appearance (as at that age they might not be familiar with their own image) so that the sticker is inconsequential to them.