Unfortunately, there's a problem with your survey assumptions.  You
wrongly assume that a red Cadillac is somehow related to specific
attributes of the buyer.  There are many factors to automotive buying
that you're not considering.  It's not possible to answer your survey
without making it totally unscientific.  Market research isn't so
simple.  You haven't allowed for multiple choices, and you didn't
include an option for when the participant has no opinion.  The most
obvious thing that you should change is to instead ask if the
participant would buy something, and then collect personal information
for the statistical weights.  If that isn't the purpose of the survey,
you need to state what the intended purpose is before the participant
begins it.  It appears that you're trying to determine if there exists a
general stereotype against Cadillac buyers, but stereotypes are also
caused by many factors which you're excluding.