This survey only took the answers of readers of the magazine, and it is what is called a "voluntary response survey" which makes it useless, because it is taking data on people based on their voluntary responses, and probably the only people who would bother to respond were those really pissed at microsoft for selling them a broken 360, which skews the results
A survey is only accurate when it is a random sample, and the data taken is not based on voluntary response (also they only took the answers from THEIR readers, who are probably hardcore gamers who play more than the average 360 owner, thus giving the system more wear and tear)
don't get me wrong, the 360 failure rate is ludicrously high, much higher than it should be
but don't say this survey is accurate, because it is not, I have taken a number of college level courses in statistics, and this sample is biased because it does not accuratly (and randomly) represent every group of the population it is taking data on (it's data is inferring the failure rate of 360's, and it is only taking data on 360 owners who read and are willing to respond to a game informer survey)
shame on game informer for publishing this like it is fact