The key word there is "can". You can get a reasonably accurate figure from a relatively small sample size, but only if all of those people are telling the truth, and only if the group of people you're surveying are representative of the 360 owner base as a whole. Surveys have been manipulated to skew all kinds of statistics in the past. Whether that's happened here or not is open to debate, but those that are adamant that these figures are 100% accurate clearly have an agenda of their own.
Do you think that if there was a survey on here about 360 failure rates, that maybe, just maybe, some PS3 owners that have never owned an Xbox 360 might vote to say that they've had multiple failures? Oh no, that's completely beyond the realms of possibility! lol.
I haven't read the full thread, so forgive me if I'm repeating things that have already been said, but not denying something is not an admission of guilt. If MS had said "no, the failure rate is not 54.2%" it means nothing. It could be 54.3%, but they'd still have told the truth. Unless they come out and say "the failure rate of the 360 so far has been x%", which they will never do, then it's always going to be speculation. There's enough anecdotal evidence out there to show that the failure rate of the 360 has been (and possibly still is) unacceptable, but trying to put an exact figure on it is a waste of time.
Also, without knowing which hardware revisions failed, it's a pretty pointless statistic as there's little context to it. If 54.2% of Jaspers fail, well that's not particularly good news for anyone wanting to buy a 360 now, but the numbers could be (and almost certainly are) skewed in favour of old, long discontinued models.












