Please don't think I want to give you a hard time but... it is not elementary how you measure this and I would like to ask your method. 34% is a big deal after all (I am a homeowner in S.Texas and paying 17 cents/kwh). One of the pitfalls I see is adjusting for weather, average conditions could be different in the before vs. after test periods. It is a rare electricity user who even goes as far as you do.
Myself, I got a 17.5 SEER combo installed at the beginning of summer 2007, and tried to track my own usage. Thought it would be good to attempt to normalize it by KWH per degree-day. My usage was *way* down but so far that I suspect my methodology was not good enough. That summer was actually mild and I think KWH is not linear with respect to degree-days. Still saving a lot of money, just not sure how much until another summer or two goes by.
Best wishes -- Pstu