Republican Jewish Coalition executive director Matt Brooks wasn’t happy with my post yesterday. It cast doubt on the reliability of a poll showing that New Jersey Republican Gov.-elect Chris Christie received 38 percent of the Jewish vote on Election Day. One of the biggest problems, I wrote, was the sample size of just 72 voters, which resulted in a margin of error of plus or minus 11.5 percent.
Brooks had a legitimate argument: Why do you treat those presidential exit polls of the Jewish vote as so definitive every four years, when their sample sizes aren’t that large either? I responded that while they’re not perfect, they are conducted in the same way every four years — so at least we can compare "apples to apples" and observe trends in the Jewish vote even if the exact numbers are perhaps given too much importance.
Brooks retorted that if the margin of error were too big to be reliable in those exit polls, then they shouldn’t even be used for comparison’s sake — since they wouldn’t be accurate. I thought that was a fair point, so I decided to check out the sample size and margin of error on those presidential exit polls.
In fact, the sample sizes for the presidential exit polls are all larger than for the New Jersey gubernatorial poll — although some years larger than others.
In 2004, according to this Solomon Project analysis by Democratic pollster Mark Mellman, 268 people were surveyed for a margin of error of plus or minus 6 percent. While not as low as the three to four percent margin of error in some national polls, it’s still low enough — and a big enough sample size — for pollsters to feel perfectly comfortable using.
Last year’s election was a little different. According to a Mellman memo prepared for the National Jewish Democratic Council, just 104 Jewish voters were polled in the national exit poll sample which gave 78 percent of the Jewish vote to Barack Obama — for a margin of error of plus or minus 8 percent. But Mellman also cites an aggregation of 28 individual state exit polls which found a similar 76-22 Jewish vote for Obama, and had a large sample of 825 Jews for just a 2.9 percent margin of error.
Brooks, though, did e-mail to me another poll, taken by GOP pollster Neil Newhouse showing similar results to the RNC poll the RJC released Tuesday. The Newhouse poll also had a small sample size of 78 voters, and had Christie garnering 34 percent of the Jewish vote. Newhouse suggests combining the two samples, which would amount to 150 Jewish voters giving the Republican 36 percent of the Jewish vote.
Of course, we don’t have numbers for Democrat John Corzine’s defeat of Republican Doug Forrester in 2005, so it’s not clear if, as Brooks argued, this shows Jewish independents turning Republican or is simply a typical breakdown of Jewish voters in gubernatorial races in New Jersey.
But in Brooks’ favor, I will say that one pollster I spoke to said that a poll taken the night of and day following the election is, in his opinion, just as reliable as an exit poll taken all on the day of the election. While, as I noted, voters tend to want to identity with the winner in post-election polls and would have thus been more likely to say they voted for Christie, such an effect usually takes more than a day to develop.
So what have we learned? I’m not sure, except that I’ve probably spent way too much time analyzing the New Jersey Jewish vote.
