A Different Kind of Sampling Error

    Jim Geraghty at National Review Online pointed out some puzzling results yesterday from a recent CNN poll:

Among those who describe themselves as “Tea Party Supporters,” 86 percent have a favorable opinion of the Tea Party, 4 percent have a negative opinion of the Tea Party, and 9 percent have no opinion on the Tea Party.
If you describe yourself as a Tea Party supporter, why would you have a negative opinion of it? Or how could you say you don’t have an opinion when you just told the pollster you support it?

Geraghty goes on to suggest professional help for those self-hating Tea Partiers.  But another kind of professional help might be in order for another demographic subgroup uncovered by the poll.  Page 16 of the poll records the results of the question: "We'd like to get your overall opinion of some people in the news. As I read each name, please say if you have a favorable or unfavorable opinion of these people -- or if you have never heard of them... The Democratic Party."  Responses are grouped in various ways, but the most interesting is the 1% of respondents who are listed as Democrats who "Never heard of" the Democratic Party:

So, how does that conversation with the pollster go?

Pollster: What is your party affiliation?
Respondent: Democrat.
Pollster: Do you have a favorable or unfavorable opinion of the Democratic Party?
Respondent:  Never heard of 'em.
Pollster: Oooooookay.  Thanks for your time...

Maybe "Sampling Error" has more than one application.