October 17, 2010

3rd CD Polls All Over The Place

So, what to make of the most recent polls in Washington's 3rd Congressional District?  A recent Survey USA poll has the GOP candidate up 11 points while a recent poll by Washington, DC-based The Hill has the race a statistical tie.  Broadly, polls have been all over the map in the Washington State Senate race as well, so what gives?

Regardless of what pollsters might say, polling is an inexact science.  There are so many variables that can effect the outcome of a poll from phrasing of the question(s) to sample size and universe.  Looking at the internals of the two most recent WA-03 polls, it's pretty easy to see why there is such a big discrepancy - the samples are different.

This may be a bit simplistic, but it speaks to the discrepancy.  The Survey USA poll interviewed a sample of likely voters that identified 34% of the voters as Republican.  The Hill poll interviewed a sample of likely voters that identified 28% of the voters as Republican - a significant 4 point swing.  Most pundits agree that voter intensity among Republicans is much higher than it is among Democrats this year, so The Survey USA sample could be spot on.  However, commentators have also concluded that Washington appears to be somewhat of anomaly when it comes to Democrat intensity.  That is, Washington Democrats appear to be more motivated to vote than their counterparts in other states.  If that is true, than The Hill poll could very well be an accurate sample.  So the answer?  Who the heck knows which poll is right?

I think we can all agree that turnout in 2010 will nowhere near mirror turnout in 2008 where there was a highly motivated group of Democrats voting in their party across the country..  A more accurate turnout model would probably be the 2002 elections in which Republicans picked up a significant number of seats.  Interestingly though, the partisan breakdown of 2002 results in WA-03 more closely mirror the sample in The Hill poll than Survey USA.  Again, all very strange.

One last point.  Survey USA has been exceptionally accurate over the past few election cycles.  Even they concede, though, that this has been a difficult year to poll in Washington, so their hedging their bets.  So be it.  ballots are in the mail, people are voting, and, in the end, the only poll that matters is the one on election day.  Onward.

No comments:

Post a Comment