Well sure. The whole value in these sorts of studies is that pet owners' intuitions are untrustworthy, as they have a very strong inclination to anthropomorphize their pets.
I'm not saying pets are exactly the same as humans, but I do think that you're going to be hard-pressed to find a better heuristic for dog behavior than thinking about what a non-verbal human would do in their situation.
(I know you didn't make a particularly strong statement here; I'm just responding to other objections I've heard).
It's a justifiable heuristic, but it is still a heuristic, and it's we choose because we are very biased toward human conceptualizations. So we shouldn't say it is a good heuristic, and we should be wary when relying on it to make decisions.