You're talking about a subsample of 101 respondents; there's a huge margin of error for a sample that small at 95% confidence, and even then 1 out of 20 results (5%) will be outside the margin of error.
Newt is getting about 20% in total - so in a sample of 101 he'd get 20 supporters. But if he's not doing so well with younger voters, maybe he's at 5%, which would put him at 5 voters in an average subsample. In any particular subsample he's likely to get 0 to 15 respondents supporting him (or thereabouts).
People get 0% in subsamples all the time. Look at some polling cross tabs. In this poll, Perry gets 0% of the 209 respondents in age group 30-44, Paul and Santorum get 0% of the 22 blacks in the sample.
Is it really unbelievable that Newt doesn't pull much support from younger voters? After all, the Republican Congress ran him out of town on a rail when these folks weren't even in high school, and he was the most reviled politician in America.
I have to admit, I've never seen a subset in a poll where a candidate got 0%. I suppose they exist, but I haven't seen one before, even though I've seen the breakdowns for a lot of polling.
I have to admit, I've never seen a subset in a poll where a candidate got 0%. I suppose they exist, but I haven't seen one before, even though I've seen the breakdowns for a lot of polling.