On Fri, 30 Mar 2007 23:17:11 -0700, SMS <scharf.steven@geemail.com>
wrote in <460dd2b7$0$27237$742ec2ed@news.sonic.net>:

>karlkrandall@sbcglobal.net wrote:
>> On Fri, 23 Mar 2007 16:43:19 GMT, John Navas
>> <spamfilter0@navasgroup.com> wrote:
>>
>>> In fact I've pointed to JDPowers as the best available data, which shows
>>> relatively small differences between carriers, on the order of the
>>> sampling error.
>>
>> Not nationally. Only in certain markets.
>
>The advantage of the Consumer's Union data is two-fold. First, they used
>a much larger sample size, hence the margin of error is lower. Second,
>they broke down the data into much smaller geographic units than J.D.
>Power. With J.D. Power, they ranked Qwest and Verizon significantly
>better than the other carriers in the West region, but the West region
>is huge, literally half the area of the lower 48, with many variations
>within the region.
>
>Even though the Consumer's Union survey only targeted Consumer's Reports
>subscribers, any differences between subscribers and the population at
>large cancelled out, since subscribers were only rating their own
>carrier's quality.

Nope. Statistics don't work that way. CU surveys (as I've explained
several times) suffer from serious flaws:
* Non-representative population
* Self-selected sample
* Small sample size when broken down by region
* Improper methods (e.g., lumping together D-AMPS and GSM,
ENS and non-ENS, iDEN and CDMA2000)
* Poor methodology
The tarnished credibility of Consumers Union is also apparent from the
child seat and other recent debacles:
<http://www.detroitnews.com/apps/pbcs.dll/article?AID=/20070119/AUTO01/701190403/1148>

--
Best regards, SEE THE FAQ FOR CINGULAR WIRELESS AT
John Navas <http://en.wikibooks.org/wiki/Cingular_Wireless_FAQ>