Is McCain really up 22 points among Jews in New York? (UPDATED)

What should you do when a poll contradicts everything you thought you knew about the campaign and the Jewish vote? Some pollsters suggest you might not want to take it too seriously.

A Siena (College) Research Institute survey released Monday has Barack Obama up just five points, 46-41 percent, in New York – suprising enough in a state that Obama is expected to carry easily. More shocking is the poll found John McCain leading the Democrat by 22 points, 54-32 percent, among New York Jews. That’s despite Obama leading McCain nationally in the recent polls of American Jews by about 30 points.

Is that number reliable? First of all, Siena poll spokesman Steven Greenberg said, the Jewish sample size was 77 respondents, which translates to a margin of error of plus or minus 11 points.
GOP pollster Frank Luntz said that margin of error makes the poll “unreliable” and said the “press shouldn’t report it.” He added that McCain is running very well among Jews for a Republican candidate, but that means a possible high-water mark of 40 percent – not a lead on Obama.

One Democratic pollster who did not wish to be identified criticizing another poll also said that only interviewing 77 Jews is a “very small sample size.”

“I wouldn’t make any decisions for a candidate based on 77 interviews,” he said.

But even before examining the Jewish results, Democratic pollster Mark Mellman said, the fact that Obama is only leading by five points in the overall poll raises a red flag. He noted that every other non-Siena poll in the state over the last two months has Obama up anywhere from 13-21 points, and believes there must be some kind of “fundamental flaw” in the Siena survey. He added that if one poll differs so significantly from every other poll, there needs to be some explanation for the outlier.

One pollster who did not want to be identified said he believes that Siena made a “fundamental violation of polling procedure” by asking respondents their preference for president at the very end of the survey–after a series of questions about how the candidates rate on issues–instead of near the beginning. The pollster said that such questions, which include queries on such topics as which candidate is more patriotic or has more integrity, are likely to influence a voter’s thinking by the end of the survey.

Greenberg defended the Siena poll by noting that it used a similar sample size in June for Jews, when Obama led 59-21 percent, and in August, when he was up 50-37. But it’s hard to imagine what could have caused Obama to lose 27 points in the Jewish community in just the last three months.