VodkaPundit

With Polls Like These I'm Supposed to Make Predictions?

Lots of chatter at American Thinker and elsewhere over that IDB poll showing Jews breaking for Romney. Well, cut it out. It just ain’t so. Or if it is so, you can’t tell from this poll.

Here’s IBD’s own breakdown:

Want to know what the fine print says after the asterisks? “*Small sample size. Interpret with caution.”

Interpret with caution? “Don’t read anything into this at all,” would be more like it. IBD talked to 909 likely voters. Assuming their samples are representative of the nation as a whole, that means they spoke with about 13 Jews. What conclusions can you draw about Jewish sentiment across the nation by doing phone interviews with 13 random members of the Tribe?

None. No conclusions. Your sample size is so vanishingly small as to be meaningless. And that’s before we even get to the 3.5% margin of error for the entire sample of 909 people. You know how high that shoots up when you whittle it down to 13?

Well, I don’t know, either. But it’s very, very high. And Jewish sample is very, very meaningless.