Dartmouth needs to stay out of the polling business
May 13, 2009So the smarty pants folk up at Dartmouth College put together their first poll (in a decade?) and they come out with something that only samples 400 people, comes out way too late to be relevant, and has a sample size that would make even Andy Smith at UNH chuckle.
They find 45 percent of the state supports gay marriage and 41 percent opposite it. So a few things. First, statistically a 5 percent margin of error could mean that more oppose gay marriage than support it. Second it also implies that 14 percent are UNDECIDED? Seriously? They must have oversampled New Hampshire governors!
UPDATE: Apparently they didn’t over sample Lynch’s house. The poll [found here] says Lynch’s favorability is just 54 percent.
This polling at Dartmouth idea is proving to be even worse than when St. A’s tried it and got laughed at. Let’s hope it was a one time thing involving a class that won’t be taught again.
UPDATE 2: I guess it is their second annual poll. I guess last year’s poll didn’t have enough impact for me to remember. Here’s hoping this one fades just as quickly.
steve vaillancourt
May 13, 2009
If this poll is scientific, I have no problem with it. It’s the kind of poll like Kevin Smith and Cornerstone that is a real traveesty. The size of the sample is not the problem, it’s that it’s scientifically structured. Andy Smith at UNH has been good recently-he had the Obama-McCain difference exactly right in NH and had the Shaheen-Sununu race right on as well. Thus, when Andy’s results shows that gay marriage is supported by a 55-39 percent margin, Republicans ought to take note. Even if his numbers are off slightly, it’s still in the margin of error. By the way, the same poll shows good news for Republicans, that parental notification is supported by a two to one margin. You can’t pick and choose and accept the polls which support your point of view and damn the others. I’m not so sure about this Dartmouth poll, but I respect Andy Smith and UNH tremendously. (And by the way, his results on seat belts are the opposite of what I’d like to see-I’m against them, but I’m not about to claim his numbers are wrong just because I don’t like them).
Steve V (proudly one of seven NH House Republicans to vote for HB436 and happy to be with the 55 perccent of NH voters who support gay marriage)
Norris Cotton
May 13, 2009
So Steve you would have no problem with this poll showing that more Granite Stater’s oppose Gay Marriage? Or at least not the margin that the UNH poll would lead you to believe?
At least its not ARG.
smitty
May 13, 2009
Amos - double check the data again (you have it flipped)… 45% in the Dartmouth poll OPPOSE gay marriage, 41% support it.
steve vaillancourt
May 14, 2009
I have no problem with scientific polls. As we learned in presidential season, when you have numerous polls on the same subject, the best result is probably to take the average. So if we have two reliable polls on NH views on gay marriage, the best result is probably the average. The significant word here is “reliable”. I’m not sure if the Dartmouth poll is scientific; I know Cornerstone one is definitely not-you don’t poll 50,000 to get a scientific result; you poll 50,000 to try to influence them, thus subverting the purpose of polling. Some pollsters have a bias; some slant their results the way their clients want. (ARG has been accused of that). Andy Smith has no bias. In fact, he’s been criticized by Dems as having too large a Republican sample in his polls. Until other scientific data proves me wrong, I’ll stand by the Smith poll that by a 55-39 percent margin, NH supports gay marriage.
My point, however, is a less debatable one. You can’t take a few calls from people incited to call by groups like Cornerstone and say that represents the will of the people. Any poll that is as far out of the margin of error as the Smith poll is much more likely to represent the will of the people.
Steve V (no slave to numbers but wise enough to read them)