Update: Nate Silver now calls Brown a 3-1 favorite to win. I respect his analytical skills but will remain cautious about any GOP prospects in Massachusetts.
Like many armchair analysts I’ve been following every new poll in the Coakley/Brown race and I’ve found the results hard to believe. The idea that in the hard blue state of Massachusetts a virtually unknown Republican could defeat a long time state Attorney General seems as bizzare as the idea that I could play center for the NBA.
But as noted by a number of analysts, the trendlines have been steadily in favor of Brown. The last poll to come out showing Coakley ahead was conducted on the 12th and 13th, every other survey since has put Brown ahead or in the case of the new Daily Kos poll shows the race tied (the mere fact that a Kos poll has the race tied is pretty telling).
One of the tools that many people use is to compile an average of the various polls on the theory that this works out the various biases. The web site Real Clear Politics did this during the 2008 campaign and in most cases the averaged poll came very close to the actual result.
So I decided to do the same and took the information from RCP. I decided to use the last ten polls, all of which were conducted on January 11th or later. Taking polls prior to that seemed to be going a bit far back in terms of reflecting the current trend.
The combined average of those polls gave me a Brown lead of 4.4 points (Brown: 49.4, Coakley: 45.0)
I then decided to see what happened if I eliminated the all three polls showing Brown up by double digits, reducing it to 7 polls.
The combined average of these surveys left me with a Brown lead of 1.28 points (Brown 48.14, Coakley: 46.86)
So in both cases we have a Brown win by at least a full percentage point and thus outside the realm of a recount.
However we also have to factor in the heavily Democratic nature of the state. As Chuck Todd put it, in almost any other state this race would be over but this is Massachusetts. For this I turned to the analysis of Nate Silver on the web site fivethirtyeight.com. Although Nate is open about his strongly liberal politics, he does offer a lot of interesting and fact based analysis.
His review of races in heavily Red or Blue states did show a tilt towards the majority party when compared to the polls on those contests. In the case of strongly Democratic states the variance was 3.4 points. I’m not sure that this really applies to the current race but just to be fair I’ll apply the numbers.
If you adjust the ten averaged polls above by 3.4 points then the Brown lead is reduced to 1 point.
If you adjust the seven averaged polls (eliminating the double digit Brown polls) then you have a Coakley edge of about 2.12 points (the only averaging that gives her an edge).
So looking at these polls we find three averages which give Brown an edge and one that gives it to Coakley.
In theory that should cause me to predict Brown as the winner, but the one major intangable we still do not (and indeed cannot) resolve is turnout. If the turnout tomorrow is similar to that projected in most of the polls then Brown will not only win but will do so comfortably. But if the turnout is closer to the 2008 model then Coakley will pull out a victory.
Of course I think a strong argument can be made that the mere fact we are even discussing the outcome of this race is in itself either a win for the GOP or a loss for the Democrats. It’s a race that Coakley should be running away with and yet she is struggling. There seems to be to be little room for the Democrats to claim a victory here, either they lose a safe seat or barely hold on to it. In either case a bad day.
Whatever happens tomorrow night is going to be very interesting to see.