Sunday, January 17, 2010

Massachusetts Preliminary Projection

It's two days before the election, therefore I no longer afford myself the luxury of calling the Massachusetts Senate special election a "toss-up". Let's analyze the data we have.

Polling and Statistical Data
We have three polls available to us that were:
(1) Taken in the past week
(2) Non-partisan in nature

These polls break down as follows:
Polling Firm Sample Size Coakley Brown
ARG 600 45% 48%
Suffolk 500 46% 50%
Rasmussen 1,000 49% 47%

So, here are the averages:
Sample-Weighted Average: Brown +0.9%
Pure Average: Brown +1.7%
Median: Brown +3.0%
Average of Averages: 1.9%

Statistical Projection: Lean GOP Pick-Up, Brown +1.9%

It is worth noting that Coakley's range is 45% to 49%, with an average of averages of 46.6%, while Browns range is from 47% to 50% with an average of averages of 48.5%. Since neither candidate appears to be over the 50% threshold, this election could very much be decided in the last two days as undecideds (estimated in this average at 4.9% of the vote) break one way or another. If these statistical projections are correct, Coakley would need 70% of the remaining undecideds.

It is also worth noting that there is a third-party Libertarian candidate in the race. It seems highly unlikely, despite the fact that he has an unrelated "Kennedy" last name, that he will garner a significant number of votes in such a close race, but different polls have treated his candidacy differently (some listing his name among the possible candidates in the poll, some not.) As we frequently see (most recently in New Jersey), third party candidates tend to underperform their polling on election day. And one would have to consider that Kennedy voters would largely swing to Brown. On the flip side, purely undecided voters, one would suspect, might break more than 50% for Coakley, given the state demographics.

So, all in all, I believe that this polling indicates, based on the data available today, that Scott Brown has a 70% chance of winning on Tuesday.

Of course, lots of variables will be important, from the impact of final-weekend spending, to the impact of President Obama (who is still popular in Massachusetts) stumping for Coakley, to, most importantly, the turnout on Tuesday (less is good for Brown, more is good for Coakley), especially relative to the polling assumptions.

I will update this statistical projection tomorrow, if, as I suspect, at least some new polling data are available.

The Betting Public
Intrade betting odds peg the race (as of this moment) at even odds. This reflects the close polls and also some disbelief that Massachusetts will actually elect a Republican. We'll see if that skepticism is warranted.

The Buzz in Washington
The inside buzz in Washington is that Brown is going to win. The GOP is already prepping a celebration dance, the DEMs already pitching talking points about how this is a unique race, not a vote on the Democratic agenda. When you start hearing these points, you know who they think is going to win. There are also several inside reports that internal polling by the Democrats shows Coakley in big trouble.

What to Make of It All
My overall conclusion: this is an extremely close race that is extremely hard to project, not only because of the closeness of the polling, but also because of the dynamics of a special election and the difficulty in estimating turnout.

Brown appears, at least at this point, to be more likely ahead than not. But he is not far ahead, if he is ahead.

So, if you live in Massachusetts, whether you support Coakley or Brown, you should make sure to vote. This may be the closest statewide race in the state for some time and voting in this type of election is far more critical than in a Presidential race (where the state is typically a slam-dunk for the DEMs.) So, get out and vote. And we'll all stay tuned Tuesday to see how you did.

I will publish an update on Monday if there is any new information to report on.

If you like this site, tell your friends.

No comments: