Don't Expect Many Nevada Polls

[silverscreenselect]

Even though Nevada is the next contested Democratic state (1/19 caucus), don't expect to see much polling done there. The major pollsters may be passing on the state to avoid a similar embarrassment as they faced in NH. Nevada has little history behind its caucuses, a large transient population and a lot of shift workers so gauging turnout could prove very difficult for pollsters. The last polls in early December showed Clinton with about a 20% lead but those obviously are way, way out of date.

One interesting detail for those who are trying to claim that the NH polling was a "historic" mistake along the lines of Dewey/Truman, in 2000, the last contested Republican NH primary, the polls were off about 9% as well (they were off about 10% here). They showed McCain beating Bush by 9% instead of the 18% he won by. No one claimed there was any sinister conspiracy then.

There is a real problem in conducting polls of two events on the same day which have a mutually exclusive results. People can't vote in both the Democratic and Republican primaries and independent vote is split as a result. Allocating independents to a poll that is looking for one party's outcome only is a very tricky proposition because it is very easy to overestimate them. "Likely" independents will probably vote, but they can't vote in both primaries and they wind up being counted in both polls some of the time.

[mrkelley23]

This is what I've been trying to say -- I understood SSS to be among those who were claiming that Hillary made a big rush right at the end to overcome Obama. I think it's more that independents who might've voted for Obama, voted for McCain instead. Maybe the big polls convinced them that Obama didn't need their vote. Maybe they're really more Republican-leaning, so at the last moment they stuck to their true colors.

But I don't think anything that happened over the last few days of the campaign would've change anything in New Hampshire any more than it would have in Iowa. The Hillary crying having a big effect is bogus, the racism thing is bogus. It just doesn't make any sense. And while people lie to pollsters all the time, the pollsters who know what they're doing can usually still come within the margin of error of the poll.

I'm now going to go check out how many people voted on each side this time, compared to the last several times.

[mrkelley23]

What little I found from the Union Leader is inconclusive. It is clear that New Hampshire shattered the record for their turnout in a presidential primary, but it's so different from 1992 and 2000 that it's difficult to draw solid conclusions.

[andrewjackson]

There is an op-ed piece in today's Houston Chronicle which originally ran in the New York Times claiming that the polling methods have difficulty with low income people's attitudes toward minority candidates.

The writer is Andrew Kohut of the Pew Research Center.

http://www.nytimes.com/2008/01/10/opini ... ref=slogin

I don't know if that link will work.


His theory is poor people tend not to respond to the pollsters and favor different candidates than do higher income people in what are otherwise the same demographic categories.

In exploring this factor, it is useful to look closely at the nature of the constituencies for the two candidates in New Hampshire, which were divided along socio-economic lines.

Mrs. Clinton beat Mr. Obama by 12 points (47 percent to 35 percent) among those with family incomes below $50,000. By contrast, Mr. Obama beat Mrs. Clinton by five points (40 percent to 35 percent) among those earning more than $50,000.

There was an education gap, too. College graduates voted for Mr. Obama 39 percent to 34 percent; Mrs. Clinton won among those who had never attended college, 43 percent to 35 percent.

Of course these are not the only patterns in Mrs. Clinton’s support in New Hampshire. Women rallied to her (something they did not do in Iowa), while men leaned to Mr. Obama. Mrs. Clinton also got stronger support from older voters, while Mr. Obama pulled in more support among younger voters. But gender and age patterns tend not to be as confounding to pollsters as race, which to my mind was a key reason the polls got New Hampshire so wrong.

Poorer, less well-educated white people refuse surveys more often than affluent, better-educated whites. Polls generally adjust their samples for this tendency. But here’s the problem: these whites who do not respond to surveys tend to have more unfavorable views of blacks than respondents who do the interviews.

In other words, pollsters know poor/less educated whites won't vote exactly the same as richer/more educated white people but the pollsters don't know how much differently they are going to vote since they don't answer the polls.


I offer no opinion on the accuracy of his theory.

[silverscreenselect]

The independent vote is one factor (actually, Hillary got more votes from registered independents than McCain did). Polls did show a bit of momentum back to her in the last couple of days, which may have intensified on Tuesday. Remember, the polls do not show voter preference on election day; they show voter preference on the 1-3 days prior to the election. With so many volatile events in this very short campaign, it's no wonder the pollsters couldn't keep up with it.

There was a record turnout, 55% of all registered voters and almost two-thirds of Democrats voted. Under these circumstances, all of the likely voter screens were badly underinclusive. If a significant percentage of those who were screened out as not likely to vote did in fact vote predominantly for Hillary and there is some anecdotal evidence to support this, then that would account for the discrepancy in her totals.

Another factor was a fairly large number of undecideds in the final polling. Again, if they broke predominantly for Hillary for whatever reason, that would explain things.

[silverscreenselect]

Poorer, less well-educated white people refuse surveys more often than affluent, better-educated whites. Polls generally adjust their samples for this tendency. But here’s the problem: these whites who do not respond to surveys tend to have more unfavorable views of blacks than respondents who do the interviews.

In other words, pollsters know poor/less educated whites won't vote exactly the same as richer/more educated white people but the pollsters don't know how much differently they are going to vote since they don't answer the polls.


I offer no opinion on the accuracy of his theory.

There is a big difference between your statement and his. Hillary did better among lower income voters than Obama did. That doesn't mean that the reason they voted for her was because they were racist.

Her appeal is greater to women, who usually do make less money than men, which is a part of the income difference between her votes and Obama's.

And unless there aren't any poor people in Iowa, that doesn't explain why those polls got the results right and the NH ones didn't. Pew is trying to cover his rear end for a poor performance and throw a little slime in the process.

[wintergreen48]

Maybe the big polls convinced them that Obama didn't need their vote.

I do wonder how much impact the polls themselves have on the results, as with instances like this-- if 'everyone knows' (based upon polling) that Candidate X will win, does that discourage Candidate X's voters from voting ('he/she doesn't need my vote, so why bother?'), or does it discourage people from voting for other candidates ('Candidate Y can't win, so why bother?'), or does it actually encourage people to vote for Candidate X (the bandwagon effect), or does it motivate people to vote for 'the other guy' (to 'make it an election, not a coronation,' or to prevent Candidate X from having a Nixonian landslide).

I think we should elect the President by going into a room with wonderful frescoes, and not leave until we have elected someone; we would announce the results by burning straw, with the color of the smoke letting the world know the outcome. This works pretty well for electing Popes, and of course Popes are elected through the help of the Holy Spirit, so maybe it would work for us in electing a President. Just need to find a big enough room with wonderful frescoes...