Romney chips away at Obama’s lead, but electoral math still favors president (Originally posted on Yahoo!'s "The Signal" Blog)
DavidMRothschild on October 10, 2012 @ 6:05PM
Both campaigns declared victory in last night's debate between Vice President Joe Biden and Rep. Paul Ryan, with Democrats focusing on Biden's passion and Republicans focusing on Biden's aggression. Neither acknowledged that it is a futile point.
Immediate polls from CBS, NBC, and Xbox Live all reported that a majority of undecided voters believed Biden won, and the prediction markets ticked up a few points in President Barack Obama's favor in the hours after the confrontation. Yet, everyone was missing the point. The question pollsters should have asked was this: Is Obama still bleeding?
DavidMRothschild on October 04, 2012 @ 1:23AM
Debates have a reach beyond the immediate bump or slide in the polls as they seep into the narrative and offer up ammunition for campaign commercials. With nearly two full weeks until the next presidential debate, the results of this one have a long time to hang around. Romney's solid performance can lead to new donations that, in turn, lead to better poll numbers in the following weeks.
In this way, debates are the opposite of conventions, in which we advise you to ignore the bump in the polls since they inevitably fade. After debates, we advise you to ignore the non-bump in the polls, because it may grow.
Our prediction moved in Romney's favor because, with the wind in his sails, he is slightly more likely to be able to close the 4.5 percentage point gap in the polls over the next few weeks. That remains, so to speak, an uphill sail..
Academics love models, but their window of opportunity has passed (Originally posted on Yahoo!'s "The Signal" Blog)
DavidMRothschild on October 03, 2012 @ 10:24AM
In case the new issue of PS: Political Science and Politics is still on your junk mail table, here's a primer on the journal's recent publication of 13 distinct predictions of the 2012 election: Five academics predict an Obama victory, five predict a Romney victory, and three say it's too close to call.
And here's a prediction I feel good about: Five of them will be correct.
All 13 of the predictions in this peer-reviewed journal are the product of fundamental models, which examine broad historical trends that influence elections rather than simply aggregating polls and prediction markets. Some of the models use polls as a guidance, but the focus is on information like economic indicators, incumbency, past election results, the state of war, and other lofty data points divorced from public opinion surveys.
I wholly endorse the idea of academics working alongside journalists in the popular election prediction industry—obviously—but PS looks silly publishing these forecasts at the end of September. Models are useful in painting a broad electoral picture six months ahead of time, before public opinion has coalesced. They typically cannot account for the narrow margins of victory that shake out weeks or days before polls open. Relying on fundamental models in October is like relying on pre-season baseball predictions in October. I would look stupid—or at least delusional in my fandom—if I forecasted the Philadelphia Phillies winning the National League East today, when they are eliminated from the running, even though they were pre-season favorites.
Even before debates, electoral map appears largely written in stone (Originally posted on Yahoo!'s "The Signal" Blog)
DavidMRothschild on October 02, 2012 @ 3:06PM
Many see potential for Wednesday's presidential debate to be a deciding moment in the 2012 election. From the our perch here on Forecasting Mountain, we don't see a whole lot left to be decided.
Since we posted our first forecast of the state-by-state presidential election on Feb. 16, 2012, six months before the Republican Party even had an official nominee, only three states have flipped camps at any point in time. Virginia pointed toward the Republican nominee for several months during the summer, while both Florida and North Carolina have recently shifted to President Barack Obama's column. Almost all of the other 47 states have moved further in whichever direction they were leaning in February as the game clock has ticked down from more than 250 days to fewer than 40 until the election.
DavidMRothschild on September 25, 2012 @ 3:12PM
There are five states that have a double digit number of electoral votes, where Obama and Romney each still have a decent 15 percent shot at winning: North Carolina (where Obama has a 31.8 percent chance), Florida (57.9), Virginia (73.6), Ohio (77.3), and Wisconsin (83.9). There is something striking when you look at a chart of the likelihood of any of them voting for Obama since August 1; there is a lot of movement, but very little crossing.
States respond to local shifts. Wisconsin's dip started just after their native son Paul Ryan joined the Republican ticket. We also know that the candidates' fortunes in states correlate with local economic trends, and that the huge differential across swing states in spending on advertisements and get-out-the-vote initiatives impacts the support and turnout in those states as well.
Yet trends still tend to be more national than regional or local. Year after year, granular predictions of the state-by-state outcome over the months preceding an election look similar to the above chart with few states cutting dramatically across the other states.
The likelihood of Romney winning North Carolina and Ohio is not much greater than the likelihood that he will win Ohio. Because the most likely way for Romney to win Ohio is a national trend that moves all of the states between North Carolina and Ohio into his column, picking off Ohio before states that currently favor Romney more is very unlikely. Conversely, the likelihood of Romney winning Ohio and losing North Carolina is almost negligible. (I am further examining the nature of these state-by-state relationships in a prediction game on my website, which readers are encouraged to play.)
This method of determining joint probability is called the ranking method and it has proven surprising difficult to beat this simple/transparent and reliable method. Drawing this method out to all states in the Electoral College, if you list all states from most likely for Romney to most likely for Obama, it is unlikely that any state moves more than a few points in rank. Thus, the quick and easy estimate for the likelihood of any candidate reaching 270 electoral votes is to figure out the state that flips the election if every state stays in their order. Below is the key section of the states 23-32 in that list:
If Romney takes Missouri he has 191 electoral votes, 206 electoral votes with North Carolina, onward until he crosses 270 with a victory in Ohio. Similarly, Obama has 243 electoral votes if he takes Nevada, 253 with Wisconsin, and he crosses 270 with Ohio.
The flip state is Ohio at 77.3 percent for Obama and 22.7 percent for Romney. This is a great, simple approximation for the likelihood of the election. It has tracked my more complex model of electoral victory, within 2 or 3 percentage points, for the entire election cycle.
This article is syndicated on the Huffington Post.