Obama Losing Support Among New York Jews? Or, More Statistical Noise?
In addition to the Public Policy Polling Poll of North Carolina purporting to show 20% of African-Americans supporting Mitt Romney, which I wrote about earlier today, another poll that is receiving a lot of attention today is a Sienna College poll of New York State. The poll itself not surprisingly shows the President winning the Empire State by some 24 points, but it’s the level of support among Jewish voters that is exciting conservatives:
President Obama’s support among Jewish voters in the state of New York has dropped 22 percentage points in only a month, according to the results of a just released poll.
The poll, conducted by Siena College, finds that currently President Obama has the support of 51 percent of Jewish voters, while 43 percent are opposed to him. Five percent are undecided. That means, Obama’s lead among Jewish voters is at 8 percentage points.
Previously, in Siena’s May poll, Obama had the support 62 percent of Jewish New Yorkers, while 32 percent opposed him. That means, last month, Obama’s lead among this group of voters was at a strong 30 percentage points.
Those polled were responding to this straight forward question: “If the election for President were held today, who would you vote for if the candidates were [Barack Obama, Mitt Romney, or Don't know]?
Interestingly, this Daniel Halper piece at The Weekly Standard doesn’t link to the Sienna poll, but having found it on my own, I can confirm that this is indeed what the poll says. It’s understandable why this would be arousing interest on the right. Conservatives have been hoping that their hyperpartisan support for Israel and Israeli interests, as well as the President’s supposed “abandonment” of Israel over the past three years, would bring Jewish voters, who have historically voted Democratic for uncountable decades, into the Republican fold. Clearly, there has been some success in that regard among Orthodox Jews, but that group still represents a small percentage of the Jewish vote as a whole, even in New York, so the impact on elections has been pretty much non-existent. Indeed, considering that Jewish voters comprised 2% of the national electorate in 2008, it’s unclear even a significant shift in Jewish support from one party to the other would have a major impact on an election, even in a state, city, or Congressional District. Nonetheless, it’s been a point of pride you might say for conservatives to try to attract more Jewish support, so it’s understandable that this poll is getting some attention on the right side of the blogosphere. And, a 22 point shift in support in one month, even among a sub-group, is something that raises an eyebrow at the very least.
But, is this a reflection of actual trends, or another example of statistical noise?
The answer, as with the North Carolina poll, is that it’s too early to tell.
One of the first things worth noting is that, in the Sienna Poll, Jewish Registered Voters constituted 10% of the total sample. This is significantly different from the 2008 Exit Poll of New York State which showed Jews representing 3% of the electorate (in New York the breakdown of the Jewish vote was 3% of the electorate statewide and went 78% to 21% for Obama. So, the Jewish vote is over represented in this poll, most likely because they wanted to over-sample that particular demographic, which is understandable given that this was a poll of 807 Registered New York Voters, and 3% would have meant sampling the opinion of just 24 self-identified Jews. At the same time, even with the oversampling this poll only represents about 81 people who were actually polled, meaning that the margin of error for this demographic subgroup was plus or minus approximately 10.3%. Even then, were also dealing with a 95% confidence ratio, meaning that there’s a 1 in 20 chance that this particular poll of New York Jews is completely wrong.
When the General Election season really began after Rick Santorum dropped out of the Republican race, I made note of a Nate Silver post about how to read the many, many polls that will be coming out over the ensuing months. One particular piece of advice from Silver seems entirely appropriate both with respect to this Siena poll, and the PPP poll I linked to earlier:
10. Don’t abuse demographic cross-tabs. The sample sizes on subpopulations in a poll — like Hispanics, young voters or evangelical Christians — are much smaller than for all voters as a whole and therefore contain much larger margins of error. For instance, a poll that surveys 600 respondents, of whom 75 are Hispanic, has a margin of error of about plus or minus 11 points on that subgroup. And that is under ideal circumstances; in practice, some subgroups (including Hispanics) are harder to get on the phone than others.
It’s easy to write the “Candidate X has problems among Group Y” stories, but very often they are just weaving narratives from statistical noise. Unless the demographic patterns are clear and consistent across several different polls, these stories are usually worth ignoring.
What we’ve seen today in both of these cases strikes me as being a perfect example of people ignoring this advice.
There’s one final thing worth noting. Even if you accept for the sake of argument the accuracy of the Siena numbers about how Obama’s support among New York State Jewish residents has declined in just a short people of time, Obama’s poll margin in New York State actually widened by four points over the same period. That seems to me the best reason of all to consider that one little bit of data to be nothing more than statistical noise.