Birthers vs. Truthers, Ctd

A reader writes:

You've twice (here and here) referred to this survey of "American attitudes toward presidential conspiracy theories" without noting the unusually large margin of error.

The sample size for this poll is only 621, with means that the overall margin of error is +/- 3.9 %. However, only 35 percent of respondents identified themselves as Republicans, which means that the margin of error for this subgroup is significantly higher (approximately +/- 6.65%). The pollster also notes that "other factors, such as refusal to be interviewed and weighting, may introduce additional error that is more difficult to quantify." So we can assume that nonresponse bias adds another dimension of uncertainty.

Pollsters often tell you margin of error for the entire sample, but uncertainty grows when you do cross-tabulations because you're breaking the sample into smaller subgroups.  

If you look at some of the other cross-tabulations for this survey, they're very difficult to believe. For example, according the poll, 22% those between the ages of 18 and 29 believe President Obama is the Anti-Christ. This compared to only 5% of those ages 30 to 45; 9% of those ages 46 to 65; and 7% of those over the age of 65.

Since when are younger people more predisposed to believe President Obama is the Anti-Christ? This contradicts every polling trend we've seen.

I think we need to be more careful when we interpret the results of these kinds of surveys. The degree of uncertainty is often larger than you expect.