Tuesday, June 21, 2011

Dept. of Misleading Polls

NOM released a poll that shows 57% of New Yorkers in favor of one man-one woman definition of marriage, contrary to an April Sienna poll that shows 58% of New Yorkers in favor of marriage equality. Dig a little deeper though, and you'll see the topline numbers for the NOM poll is completely full of it.

First we need to look at the poll's sample size. The Sienna poll has a sample size of 777 with a Margin of Error of 3.5%. This is a standard sample size and MOE. The NOM poll by contrast has a sample thats less then half of Sienna's, coming in at 302 with a MOE of 5.6. Thats a disturbingly large MOE, not surprising considering the relatively tiny sample size. Further, there does not appear to be a date given for when the NOM poll was taken, in fact they only say it is a "flash" poll, while the Sienna poll was taken over three days. This leads me to believe that the NOM poll was a single day poll. Good pollsters try to conduct surveys on multiple days in order to make sure that short term fluctuations are counteracted. From this alone we should view NOM's survey with a skeptical eye.

Next lets dig into the cross-tabs of the poll. NOM's sample is composed of 69% of people 50 or older. That is a hugely disproportionate number of older people, and old folks are much more likely to oppose same-sex marriage then their younger counterparts. Sienna by contrast is more evenly distributed, with 22% 18-34, 37% 35-54 and 37% 55+.

Because of the older population in the sample, we also get a much more conservative one. I seriously doubt that more then 50% of New Yorkers would characterize themselves as fiscally "very conservative" or "somewhat conservative" or that there are more people who characterize themselves as very or somewhat conservative then there are people who are very or somewhat liberal. This doesn't pass the smell test at all.

Finally we can look at the wording of the questions. Question 5 asks if the issue should be decided by "legislators in Albany" or "directly by the voters of New York." This is a silly, silly question on face. The question can be translated easily to would you prefer a public referendum on same sex marriage or a vote in the legislature? Further "legislators in Albany" conjures up images of corrupt, slimly politicians. Albany is after all home to one of the most dysfunctional legislatures in the nation. The voters of the New York implies you, me, mom and apple pie. And who doesn't want apple pie? Plus the point is completely moot since unlike many other states, New York has no meaningful public initiative and referendum system. Attempts to introduce one have gone nowhere fast. This is a post in and of itself.

The money question, question 6, asks "do you agree marriage should only be between one man and one women?"Contrast this to Sienna's wording: do you support "making same sex marriages legal in New York State?" NOM's wording presents a default position, and asks the person to support it. Sienna asks if you would support a change. While this might seem like semantics, the wording is incredibly important. People naturally flock to a default position, regardless of what it may be. Right now New York is debating if we should change the system, not if we should keep the status quo. That is the thrust of the debate, and of the bill being discussed.

Between the tiny sample size, the short time period, the bizarre cross-tabs, and the loaded question wording, this NOM poll should be discounted as the hack partisan poll it is. So far the media doesn't taking the bait on this, so good for them for a change.

No comments:

Post a Comment