Subject: calculating confidence rates in samples. (Statistics)

Never had a straight answer on this. Can you help?
Assuming that I want to know how many people have a red car in the U.S., 
how many people would I have to sample in order to be 90% confident of my 
answer?
I suspect that the formula must incorporate a variable that depends on the 
answer that my sampling returns. I.E. If I sample 100 people and only 1 
person has a red car, then my confidence of accuracy must be much lower 
than if 80 people had a red car. I just need the formula. Thanks for your 
help.
Roy Feinson

Re: calculating confidence rates in samples. (Statistics)