On the other hand, doing well at a non-name-brand school doesn't mean another person can't perform well, it just means that you can't predict quite as easily how well they'll do.
If you give an easy exam and everyone scores 90% or better you don't know much more about the person who scored 100% than you do about the person who scored 90%. If, on the other hand, you give an extremely difficult exam that results in a curve with the 90% cut-off at 80% and the 10% cut-off at 20% you know much more about the abilities of those who did well compared to the abilities of those who did poorly. I think people tend to view a name-brand college as a the very tough exam and a non-name-brand college as the easy exam. Whether or not this is a safe assumption is another question.