What's the margin of error on this survey? My guess is the one tenth of a point drop in satisfaction is inside it.

A more accurate title (for the original article) would be something like: "iOS 6 Does Not Show the Same Increase in Customer Satisfaction as Previous Releases". And even there, there could be multiple causes which might not have anything to do with how users see iOS 6 (bias in the users sampled, compression of responses as users get closer to the maximum satisfaction, etc.).