Ah, I see what's going on here. The thing is that stating a confidence level only works for quantifiable matter; there's no factual distiction between
satisfied and
unsatisfied in terms of numbers. If I look at my own case, I'm partially dissatisfied (say: 40% because there's no new track included in the patch, but I'm also happy with the game content in the first place and the driving physics), am I then
satisfied or
unsatisfied?
Also, a standard deviation is the 'margin of error' and 5% means a value A can be between 0.95A < A < 1.05A (one sigma). If you talk about the 99% confidence interval, that means 3 sigma, i.e.: you can say for 99% certainty that all of your answers are between 0.85A < A < 1.15A.
Linking the information from the two paragraphs above equates to saying that people are unsatisfied with a confidence (interval!) of 99%, is therefore unfortunately not correct.
What we do know, is that most members on GTPlanet are dissatisfied (apparently). Extrapolating that to the whole community is just a tad more difficult, if not impossible. The only way of getting more accuracy is a larger sample...but that would only do so much. Just remember that statistics is all about manipulation.