And Nate Silver’s Early Weekend Analysis Was Right On
Looking at the Primary results in Arizona it’s too bad Mr. Santorum did not beat 99 to 1 odds and win the state, but then nobody really beats 99 to 1 odds. Of course it didn’t help that Mr. Santorum spent the last week of the Michigan/Arizona campaign building his lack of electability. One wonders what the results would have been in these two states had Mr. Santorum been prepared for the lone debate, and had decided not to make a bunch of idiotic statements. But then, he would not have been Rick Santorum, would he.
In Michigan it probably would not have made any difference. Mitt Romney did what he had to do, which was to generate turnout in his strong areas. Actually, it looks like Mr. Santorum may have generated the turnout for Mr. Romney by disgusting voters with his recent comments. No matter, Republican voters turned out for Mr. Romney.
The Detroit area came out heavily for Mr. Romney, and his vote total in Michigan was more in 2012 then it was in 2008. This dashed the media’s hope for a really close election, and while Mr. Romney did not win by as great a percentage as he did in 2008, he will claim momentum going into next Tuesday’s Super Tuesday primaries.
This Forum has followed the forecasting models of the New York Times’s Nate Silver, whose Five Thirty Eight is the place to go for objective quantitative analysis. Before the weekend Mr. Silver had Mr. Romney up about 4 points, pretty close to the actual result. Over the weekend the polls showed a closing of the race, but turnout for Mr. Romney and the stupid remarks by Mr. Santorum made those polls incorrect. As this Forum stated, when a race is projected to be very close it is very easy for the polling and the projections to be wrong. Here are the comments that were made here three days ago.
The Romney win was 3.2% points, 41.1% to 37.9% pretty close to what Mr. Silver's model was saying. Both candidates picked up votes from the collapse of the Paul and Gingrich vote (although Mr. Paul, like Mr. Santorum is claiming victory because they having a different defintion of winning then the rest of us.) Below the surface are some very important results that the major commenators will miss. Foremost of these is that Mr. Romney won amongst Catholic voters, and he won the anti-abortion vote. The point, Mr. Santorum's appeal on social issues is a loser in a state like Michigan. The contests in March will now determine if that is a permanent state, or if social issues can propel Mr. Santorum in other, more conservative states.
This Forum has often lauded the analytical abilities of Five Thirty Eight’s premier political analyst Nate Silver, for the simple reason that he is the most accurate forecaster of election results. Mr. Silver uses quantitative models of polling and other data to forecast the percentage a candidate will get and the probability that the candidate will win.
Because the models used by Mr. Silver rely on data as it becomes available, they are not good predictors until just before an election. So now we have the best forecast available for the Republican primaries in Michigan and Arizona , a day before the election. The expectation is that Mr. Romney will win Michigan by about 4 points.
Michigan | VOTE PROJECTION | CHANCE OF WIN |
---|---|---|
Mitt Romney | 38.9% | 77% |
Rick Santorum | 34.6 | 23 |
Ron Paul | 15.6 | 0 |
The Romney win was 3.2% points, 41.1% to 37.9% pretty close to what Mr. Silver's model was saying. Both candidates picked up votes from the collapse of the Paul and Gingrich vote (although Mr. Paul, like Mr. Santorum is claiming victory because they having a different defintion of winning then the rest of us.) Below the surface are some very important results that the major commenators will miss. Foremost of these is that Mr. Romney won amongst Catholic voters, and he won the anti-abortion vote. The point, Mr. Santorum's appeal on social issues is a loser in a state like Michigan. The contests in March will now determine if that is a permanent state, or if social issues can propel Mr. Santorum in other, more conservative states.
As far as projecting Super Tuesday, it will no be until the weekend for polls to be reliable. The voters will have to digest the results of the Michigan and Arizona primaries, and this will take a few days. So don’t pay any attention to what anybody says until Saturday at the earliest. Then look at Mr. Silver’s modeling results. That should tell you what to expect.