Also bear in mind, Google Analytics is completely dependent on Javascript in order to function. Some users do disable this technology. However, that should not account for a 1/5 variance in traffic reported. If anything, you might expect GA's numbers to be off by up to 10%. Generally, if you run a local analytics program that analyzes your internal logs, which counts everything including bots and users without javascript, and you run GA which counts everything but bots and users without javascript, your actual visitors should be somewhere in the middle, probably closer to GA then the local analytics program.

__________________
The best way to learn anything, is to question everything.