Doug Miller wrote:

> CR's surveys are entirely self-selected, which right off the bat makes them
> completely invalid from a scientific standpoint.

No, it simply makes them not double-blind surveys. Unless you think that
a Saturn driver is more likely to complain about defects than a Toyota
driver, the surveys are valid even though they're not double-blind. In
fact you could argue that a Toyota owner would have higher expectations
than a GM owner and be _more_ likely to complain about problems.

> "Often" is not the correct spelling of "sometimes".

Because often is the proper word to describe the reality.

It's easy to understand why some people get mad at the CR surveys and
attempt to dismiss them simply because they are not the "perfect" survey
where they would call 100,000 people at random and ask them all the
questions in the survey, but it doesn't make their survey "completely
invalid."

And of course other publications and organizations reach largely the
same conclusions with their own surveys, i.e. J.D. Power. It's not like
CR is telling anyone anything that numerous other organizations haven't
confirmed to be true.