About Those Polls

Michael Barone provides an excellent short course on the perils of polling, in the context of this year’s election. The key point is that in 2004, people who actually voted split exactly equally between Republicans and Democrats. On the other hand, current polling is showing party identification favoring the Democrats by 5 to 12 points. As Barone notes, party identification has historically changed only slowly. It is highly unlikely that the turnout on November 7 will favor the Dems by anything like 5 to 12 points. Which throws into question most if not all of the poll resuts we’ve all been seeing.
In addition, Barone points out that far fewer people generally vote in off-year than in Presidential elctions: 75 million in 2002, 122 million in 2004. Which means that this year, the parties’ turnout operations are critical. Nearly everyone agrees that the Republicans have an edge in this department.
Barone doesn’t draw this conclusion, but the implication of all of this is that Democrats shouldn’t start popping the champage corks just yet.
PAUL adds: Last Friday, I had the pleasure of hearing Barone speak to the Washington D.C. chapter of the Federalist Society. Michael was full of insights such as those mentioned by John in this post and in Scott’s post below. He was also tremendously funny — who knew that there was so much to laugh at in this election cycle.
Towards the end of the lunch Michael, polled the 100 plus conservative lawyers present as to our preference for the 2008 Republican nominee. Mitt Romney was the clear winner.

Responses