The example I'm giving is when I am on cellular data (not on a WiFi network). At lunch, in a so-so coverage area, I lost 15% in an hour. It was locked in my pocket. At home, overnight, I lose about 3%.
Another reason why I'm waiting 'til the end of 2013 to get the iPhone 5S.
More than the actual battery life I'm suprised at the discrepancy between Apple's claim and real-world testing. Didn't Apple boast at a keynote a few years ago that its battery life claims were based on realistic usage while competitors were disingenuous with their claims?
Most "real world" claims are based on optimal signal. All cell phone battery usage estimates are based almost solely on impossible circumstances: Most people (at least in the suburbs) are in situations where they have all bars for even a single charge. Can't speak for apples testing methods, but the numbers they boast seem to fall in line with the same testing methods.
I'm also willing to bet that those claims are based on cellular service only, without GPS or any other antennas running at the same time.