"pkbrandon" <pkbrando@hickorytech.net> wrote
> CR's user surveys might not be of publishable scientific
> level,

CR's automotive survey is as valid as most any study
published in science journals, and probably moreso given the
size of the database yada. For crying out loud, studies with
a mere 20 data points are usual in many fields. Total crap,
but it gets published in referreed journals.