Friday, September 05, 2008

SUSA Poll on VP Choices

I prefer using the SUSA polls to those done by the big media organizations because of the consistent methodology and the consistently illuminating cross tabs. (Under no circumstances trust any poll from Zogby.) The latest poll was brief and direct. It asked 1061 registered voters, out of 1200 individuals called, only eight questions, which I'm reproducing here to show how a well constructed poll is done:
  1. Did you hear Sarah Palin's speech at the Republican National Convention last night?
  2. What grade do you give her on the speech? An A, B, C, D, or an F?
  3. On the whole, is Sarah Palin an asset to John McCain? A liability to McCain? Or, do you not know enough to say?
  4. Does McCain's selection of Sarah Palin as a running mate reflect well on MCCain? Reflect poorly on McCain? Or do you not know enough to say?
  5. On the whole, is Joe Biden an asset to Barack Obama? A liability to Obama? Or, do you not know enough to say?
  6. Does Obama's selection of Joe Biden as a running mate reflect well on Obama? Reflect poorly on Obama? Or do you not know enough to say?
  7. If you were placing a bet today, would you bet that Barack Obama will be elected president? Or, John McCain will be elected president?
  8. Is the media rooting for Barack Obama? Rooting for John McCain? Or trying its best to be fair to both?

I won't go diving into the crosstabs of the poll very deeply, so click on the link to see for yourself (they have a very convenient downloadable PDF of the poll, btw), but will try to touch on the overall picture it presents.

The first question is vital because it defines who it is that they are polling. Everyone who was asked these questions is a registered voter but not a likely voter. 75% of them actually listened to Palin's speech, and there is no indication what counts as "listened" (All on the TV? Highlights from teh talking heads shows? YouTube clips?).

The sample was 52% female and 47% male, was weighted towards voters under 50 (56% of those sampled), was predominantly White (75%), and higher income than the average population (70% over $40K annual income). There were more Republicans in the poll (36%) than Democrats (33%) or Independents (26%) which doesn't reflect party registration but probably does echo voter turnout. Idoelogically, 34% identified as Conservative, 38% as Moderate and 14% as Liberal. The sample was divided into 4 geographic regions: South (36%), Midwest (23%), West (22%) and Northeast (19%).

This gave the pollsters a sample that is relatively young, relatively affluent, slightly more female than male, and significantly under represented Hispanics as their portion of the population (8% polled). This strikes me as a sample that calls upon demographic groups that are aligned with both candidates, though the significantly higher levels of white, Republican, conservative and Southern demographics I think provides a slight bias for McCain. What we don't know is the location the people live - urban, suburban, or rural.

The response to Palin herself was strongly positive, noting that saying she gave an "A" speech doesn't mean the listener agreed with or approved of the message delivered; it is an opinion of the effectiveness of the speech. The take away from me on this is people who liked the speech really liked it and there was a greater number who strongly liked it than who strongly disliked it. Independents in particular were not turned off by it, which is the political demographic both parts need to capture for a victory. Political moderates were not turned off by the speech. Higher income earners had a better response to it than lower income, which may seem odd given the "Aw shucks, we're just folks," packaging, but aligns precisely with Krugman's writings that the Republicans' electoral gains since the 70s occurred among higher income whites more than any other group. Thus, both McCain and Obama owe much electoral success to white affluent issue voters. The data points that must trouble the Republicans most are the low scores in the West, a full 10% less than those in the South. It matters, however, where those Western samples came from. If from California, not electorally important given it is a blue state. If from Colorado, a different kettle of fish. The data points that must rattle the Democrats is the lack of strong negatives in the Midwest, where the election will be decided.

When comparing the VPs and their effect on the races, Palin was considered a solid though not overwhelmingly strong asset to the ticket, while Biden barely broke 50%. There were a lot of people who had not made up their minds about Biden being an asset, which is not good considering that he is a known quantity. When comparing what the VP pick says about the presidential nominee, Obama is hurting compared to McCain. He does not even get 50% and is 7% behind McCain. This selection does not reflect well on him, particularly among female voters, the 35-49 age bracket (his own age cohort), wealthier voters (which is a big change from his primary voter patterns), and with Midwestern and Southern voters. In every region, McCain's choice was seen as reflecting well on him, all above 50%. This affects key swing states like Ohio, Michigan, Missouri and Virginia.

The part of the poll people point to most, question 7 about placing bets on the outcome, is interesting mostly because there was little indecision. Only a few categories showed indecision 10% or above, which means people feel confident about their take on the election. They favor McCain by 3%, though the undecided pool is large enough to move either candidate to over 50%.

Overall, Palin is viewed as a bigger asset to McCain than Biden is to Obama, people think more highly of McCain's selection of Palin than of Obama's selection of Biden, and a slim majority now place their bets on McCain to win. Conservatives and Republicans are more strongly behind their candidates than Liberals and Democrats are behind theirs. Hispanics are significantly undecided and so are people in the Midwest. Palin did not provide as large a gain in the West as the Republican strategists probably had hoped for, but a lot of that will depend on where in the West the respondents live.

At this point, national polls are almost useless except to measure broad trends. What matters is how opinion breaks down in swing states.


1 comment:

sister of ye said...

Under no circumstances trust any poll from Zogby.

Oddly enough, I just answered an online Zoby poll yesterday. I'm not sure how I got on their list, but it's my policy to answer pretty much any opinion poll. I mean, there's many a time I said, Where the f*ck did they find the odd people who gave those answers? So I figure one of those odd people may as well be me.