I really don't know how Harris can call this a valid poll at all.
"a nationwide sample of 2,513 U.S. adults, 611 of whom follow college basketball."
So, the nation is being summed up by 2500 people? Of which 24% don't even follow the sport they are polling about? I don't even know how that can be a valid sampling. "Do you follow any of these sports?"
Base: All adults
2007
2008
%
%
FOLLOW COLLEGE SPORTS (NET) 41
41
College football
33
34
Men’s college basketball
24
23
Women’s college basketball
8
7
College baseball
6
8
College hockey
4
3
Other college sport
8
7
I follow none of these sports
59
59
Note: Multiple-response question.
Basically, the Harris Poll had 59% of the 2513 polled didn't follow sports. I don't think this random sampling truly reflects the nation. More than likely, it is name recognition (if that). 
