Sep 25 2008

Is Operation Chaos Messing With The Polls?

Published by at 12:36 pm under 2008 Elections,All General Discussions

I was not a really big fan of Rush Limbaugh’s “Operation Chaos”, but the fact is he led a sizable number of Repubs and others to register as Democrats in order to vote in the primaries. Here is one example of the havoc he was able to achieve:

Joe Lenski, executive vice president of Edison Media Research – which did exit polling in Pennsylvania, where about 2 million registered Democrats turned out to vote – said research showed that about 100,000 voters (5 percent) said they had re-registered from Republican to Democrat. Another 4 percent who voted on the Democratic side were new voters who said they had no party registration before the primary, and an additional 4 percent said they were unaffiliated or with another party, he said.

Among all new voters, the preference was for Obama, 58 to 42 percent according to the exit polls, Lenski said, though the Republicans-turned-Democrats went 50-50 for Clinton and Obama, he said.

Limbaugh also cites exit polling that suggests some 100,000 Republicans-turned-Democrats cast ballots for Clinton in Ohio, along with nearly 120,000 in Texas and 38,000 in Mississippi.

While 100,000 out of 2 million is not a big number (5%), you need to add into those numbers Hillary Clinton supporters as well to see the potential pit fall here. The thing is, most polls rely on voter affiliation and registration percentages to determine their voter models and turnout models. So if the models are off by 5% on who they think is a Democrat verses a Republican, that is a big deal. For example, if a state like PA has voter registration as 50% Dem to 40% GOP, and those 5% of dems (2.5% of voters) are faux registrants, then the REAL number is 47.5% Dem and 42.5% GOP.

Addtionally, if 10% of Hillary supporters go to McCain, assuming Hillary supporters are half the Dem primary voters, these losses result in another 2.5% of the electorate shifting.  That brings the Obama-McCain race into a 45-45 tie. OK, this is silly speculation – but I do wonder how many die-hard Republicans are still registered on the books? How many Hillary supporters (or Dems in general) will Obama lose. I think the polls are actually a bit Pollyannish on Obama’s popularity.

Addendum: A few more thoughts on poll models.  One of my commentors, MerlinOS2, noticed a risky disconnect between one poll’s party affiliation and party registration in Colorado:

PPP just released a poll in Colorado which puts Obama up +7

Now what the issue is here is that the party split was

Dem 40
Rep 36
Ind 24

However August voter registration number per the spreadsheet available from the Secretary of State show the registration breakdown is

Dem 30.6
Rep 34.8
Ind 34.5

Clearly that poll is aggressively optimistic, and is allowing the sample to drive the model (i.e., more Dems participated in the poll than Reps). Obama better be very careful in relying on these kinds of polls, he could lose the women’s vote to McCain-Palin if Hillary supporters assume he is going to win. There is much more oppoertunity for a protest vote for Palin if Obama appears to be leading due to bad poll models.  And we all know how bad the polls were this year, and they seem to still have problems, given the wide variations in poll numbers in states from polling over the same time period.

5 responses so far

5 Responses to “Is Operation Chaos Messing With The Polls?”

  1. browngreengold says:

    The race is back to a tie in the Gallup Daily.

  2. MerlinOS2 says:

    Another factor that is throwing the exit polls off is early voting.

    AJ if I am correct you could start voting for the Presidential Election last Friday.

    The percentage of voters in both primaries and general elections is growing as to number of early voters.

    Growth is both in existing early voter states and new states adding the capacity.

    Va opened last Friday and a half dozen more states opened this week.

    All those voters (40 – 60% of registered voters now vote early) are not available for exit poll surveys.

    That is part of the reason the much talked about Kerry exit polls had little to do with the final vote tally.

  3. crosspatch says:

    Also one needs not to look at voter registration, but look at actual voter turnout averaged over, say , the last two elections. If you have about even numbers of registrations but a larger percentage of one party’s voters actually goes to the polls and casts a vote, that side is going to win no matter what the “polls” say.

    That is why “get out the vote” efforts are so important. It is actual votes cast that get a candidate elected. Being a member of a party and being in the majority of opinion polls means nothing if you don’t get to the polls and cast your vote. Republicans tend to vote in greater percentage than Democrats.

  4. WaPo/ABC polling Bias exposed: sample consists of 38% Democrats and 28% Republicans….

    poll internals show a party affiliation advantage to Democrats of +10. If leaners are forced to make a choice, the Democrats have a +16 advantage. That is quite simply hogwash. The average advantage of Democrats vs. Republicans in the last 6 or so nati…

  5. Wayne at Jeremiah Films says:

    I’ve linked to your post from WaPo/ABC polling Bias exposed: sample consists of 38% Democrats and 28% Republicans.

    Some of these sample rates look very, what should I say optimistic for Obama, … if any population should be discounted as unlikely history tells us that it will be the younger voters.