With any survey-based research, it is important to remeber that the question and answer format limits the accuracy and scope of such research.
In addition, the sample size has to be sufficiently large to be representative and relevant. Both J.D. Power and CR provide the total sample size, however they do not provide a breakdown by automaker. Considering that makers such as Toyota and GM have much larger sales volumes, presumably makes the ratings more accurate (bodes well for Toyota/Lexus and not so well for Ford and GM). Further, none of the big surveys provides a definition of how answers are weighted and how attributes are positively or negatively affected. Finally, people who do send back surveys could be considered a biased group (one that likes to fill out surveys or has an excessive need to overcome buyers remorse or has a bone to pick with the carmaker, etc). I suppose once you start doubting there is no end to it...
IMHO, CR is the least reliable survey, due to the fact that CR readers are geographically focused and may already be predisposed by the reviews of CR (positive/negative CR review forms positive/negative impression) to answer one way or another. But then again one can extrapolate this point and apply it to the general public being biased due to media exposure.
I guess it's time for my medication... :P
EDIT: From a global perspective, I consider the What Car survey to be one of the best surveys. The ratings are in part based on actual warranty claims submitted through an extended warranty provider called Warranty Direct. It also provides the average mileage and age for each manufacturer.
I'd welcome to hear comments from any carspyshots members out of the UK in regards to this survey
Modified by evomk8 at 7:28 PM 1/24/2006
Modified by evomk8 at 7:29 PM 1/24/2006