I agree.. the sample is too small to be of any significance - esp on the AT&T / VZ customer polling.
I'd recommend somewhere in the area of 10,000-25,000 to at least make a sampling. Having 7 or 11 people as a valid number for AT&T or VZ is not enough for any meaningful statistics. 1/10 our of 10,000 samples means something. 1/10 out of 10 samples is too small of a sample base. -- Canada = Hollywood North
