No, I'm reiterating what's already been said. These figures here can't possibly be anywhere near accurate, and the only way to get accurate figures would be from Apple, whether they choose to be honest about it or not.
I also don't think they're accurate, but it is worth noting that they published numbers on previous generations of Apple notebooks using essentially the same methodology. So, aside from systematic changes in the way they delivered the survey or the characteristics of the "responder" sample within the overall sample (because the sampling is clearly not representative), the difference between the current and past rates should still be interpretable.