Saturday, October 27, 2012

Sloppy Polling Memo By Mason Dixon Says Akin's Getting Closer to McCaskill-Updated

Update: McCaskill campaign poll blows the Mason Dixon poll out of the water! (see below) A new poll paid for by the St. Louis Post-Dispatch, the Kansas City Star, and KMOV and conducted by Mason Dixon is out claiming that Senator Claire McCaskill is only leading Todd Akin by 2 percentage points. However, the polling numbers are strange and the memo, if not the poll itself, is very sloppy. As Brandon H notes, they spelled Jason Kander's name "Kandor:"
Brandon also pointed out that the memo used "Conservative Party" instead of the correct name "Constitution Party" on the question about the Presidential race. But what's really strange, and possibly a larger problem if the sloppiness goes beyond just the memo, is that the poll questions as written on the document asked about third party candidates by name in all of the races *except* the Senate race (h/t Alexandra). For example, here's the wording for the question about the Treasurer's race, which includes the Libertarian Party candidate:
On the other hand, for the Missouri Senate race, the question on the memo only lists Claire McCaskill and Todd Akin, despite the fact that Libertarian candidate Johnathan Dine has been polling at anywhere between 6 % and 9 % in recent polls:
Note that Dine is listed as getting 4% of the vote, so perhaps he was included on the original question (despite what the memo says), but 4 % is lower than his averages in other recent polls. I'll also point out that in the two most recent polls Todd Akin's unfavorables have been waaay over 50%. In a poll taken on Oct 17, Rasumussen found 62% of respondents had an unfavorable view of Akin. And on a poll conducted from Oct. 19 through 21, PPP had his unfavorables at 57%. But on the new poll, conducted only a few days later from Oct 23 to 25, only 42% of respondents viewed Akin negatively. So yeah, after Todd Akin said that women's bodies can magically shut down pregnancies from "legitimate" rapes, after suggesting that McCaskill was "unladylike" because she was aggressive at a debate, after comparing McCaskill to a "dog," not to mention his crazy policy positions, only 42% of Missouri voters have a negative impression of him? I have a hard time believing this poll ended up with a representative sample. Anyway, what's always important to note is that individual polls should never be taken as perfectly describing the elections. Polling averages are generally much more accurate, and every once in a while even a well-conducted poll will end up with an average that's off the mark. And if the race was really genuinely close, would the RNC be acting like this:
An NRSC spokesman said it is a keeping "a close eye on this race" and hopes Akin can win.... But two Republicans with knowledge of the situation said the committee has looked at the race several times in the aftermath of Akin's remarks and decided he is too far adrift of McCaskill for them to get involved with the contest. The committee has instead poured money into other states. The Republicans spoke on the condition of anonymity because they were not authorized to discuss the committee's decision-making.
Update:Scott Charton is tweeting out the results from an internal poll from the McCaskill campaign showing her with a lead of 53-39. And (cough cough) it shows Akin's unfavorable rating at 58%. So, if you're trying to decide which poll is more plausible, consider that the Mason Dixon poll requires us to believe that within a period of 3 days, 1 out of every 5 voters (20%) previously had an unfavorable opinion of Akin but then changed their mind. On the McCaskill poll, Akin's numbers stayed exactly where they have been in two other polls taken earlier this week.

No comments:

Post a Comment