Thursday, November 15, 2007

The problem of polling in Nevada

In 2004, 5808 people voted in the Democratic Caucus in Nevada.

Doesn't that strike you as crazy? In a state of 2 million citizens and 516, 785 active Democratic voters, 5808 voters is about 1% of all voters. Yet, they are the ones that selected the Nevada delegates for the convention.

It is excusable since Nevada was simply a rubber stamp on John Kerry's march to the White House. However, in an effort to add diversity to the primary schedule, Nevada was moved up as one of the first 4. Thus increasing its importance.

So on the eve of the Democratic debate tonight in Las Vegas, CNN has released a poll and Hillary Clinton has a commanding lead.
The New York Democrat is the top pick of over half, 51 percent, of the likely caucus participants interviewed for the poll. Her closest rival, Sen. Barack Obama of Illinois, was the choice of 23 percent. Former Sen. John Edwards of North Carolina came in at 11 percent. All other candidates came in at single digits: New Mexico Gov. Bill Richardson, 5 percent;

This is, of course, extremely concerning since I've said that Richardson must win Nevada to win the White House.

Yet, there is more to the Nevada poll than meets the eye. In fact, CNN themselves are saying Nevada is a crap shoot.

Why? Well because polling is nearly impossible in Nevada. One could poll "Democratic voters" but that doesn't necessarily mean they will caucus. They may vote in the general, but not turn out for the caucus in January.

One could poll "likely caucus voters" as a better way to find actual voters. However, there are two ways of figuring that out. The first is asking people if they are planning on caucusing. Which is good, but we all know that what is planned is not always what is done. Folks plan on going out to dinner, making Bible Study, playing in their soccer game, etc, but life sometimes changes all that.

The other option is to ask if people had caucused previously. Well, that's a problem because those that caucused in 2004 are most likely all political insiders in the Democratic Party. Since the party establishment has backed Hillary, the sample will be overwhelmingly pro-Hillary.

Additionally, no pollster knows how many voters will turn out. Certainly more than the 5800 in 2004. Would a massive increase help Hillary? or Obama? or Edwards? or Richardson? What happens if that turnout is all located in Las Vegas? Would 8 votes be enough to carry Warren county again? or 3 votes to win Sherman county (which awarded 2 delegates in 2004)? What effect will performances in Iowa and New Hampshire have on voters in Nevada? Are they committed to their candidate or only have soft support for them?

Unlike Iowa, there are just too many unanswered questions. I'm not saying throw out the polls in Nevada. They have some relevance. However, do take them with a grain of salt. Hillary is far from the winner in Iowa and Richardson is far from out.