Yes I do- because it's how statistics work.
If there's no substantive difference between the 35 states included, and the 15 not, in terms of vehicle quality, then it shouldn't matter to the validity of the results nationally.
That's a common misunderstanding when it comes to survey data. It's not just the subject matter (vehicle quality) that has to be the same from state to state, it's also the demographics. People in traditional auto producing states could view minor Tesla flaws as more serious than the same flaw on their Ford or Chevy (and we don't know which states were included/excluded from the survey).
Can you unpack that a bit?
To my knowledge they send the same survey to everyone.
You don't know they send the same survey to everyone. Hopefully the survey contents are the same but the packaging can greatly impact the response rate and who (and why) someone might respond. JD Powers might mention the survey is done on behalf of Ford (for Ford owners) or GMC (for GMC owners) in order to tie the survey to the brand the owner purchased and make the survey seem more legitimate (which it would if it was being done for the benefit of the manufacturer). This could increase response rates. On the other hand, it would be fraud for JD Powers to claim their Tesla survey had the manufacturer's approval.
So again there should be no difference from one person to another in that regard.... and no impact of any kind regarding who 'sponsors' anything.
Everyone gets the same set of questions.
I would like to assume that is true but JD Powers didn't address the specifics. More importantly, the presentation of the material (as mentioned above) impacts the response rate (and this in turn impacts the kind of person who responds and their reasons for responding might be motivated by different things). People have conscious biases but the unconscious ones are the hardest to control for.
That's possible. But again doesn't really change much unless you have a reason to think upset Tesla owners are somehow MUCH more likely to respond than upset owners from all other brands?
I don't have an opinion on that because I don't have enough data to form one. Neither do you and that is my original point. There is a reason they didn't include Tesla's in their final results (because they wanted to claim it was properly performed according to commonly accepted standards). Obviously, it wasn't or they wouldn't have to present the results separately.
But they don't exclude it.
The entire point of the story was that they're publishing their results based on the 35 states they were able to send surveys to.
Wrong. They exclude the Tesla data from their scientific survey. While you can claim they published the results, they are not claiming the Tesla results are statistically valid or meet the same standards.
The larger problem is counting a minor fit/finish flaw with the same weight as a reliability problem or a car that needed a new transmission in the first 1000 miles. But, of course, doing it in a more reasonable manner would not add as much value to their paying customers (all the existing OEMs).
