Electoral College: Florida Flips Again, Still Really Tied

So yesterday I said:

As I’ve cautioned before with Florida (and a couple other states), the state is close, the five poll average has bounced back and forth across the line repeatedly. The average today moves slightly to the Obama side of the line. But it could easily go back to the Romney side of the line tomorrow. There have been no moves that indicate Florida is moving definitively toward one candidate or the other. Absent such a move in the next few days, we’ll basically just need to wait for the actual vote count.

Florida did indeed flip back to just barely Romney in today’s update, and I’ll just let my quote from yesterday on that stand for today as well, with the names reversed.

For reference, the chart of polling in Florida shows the situation there quite clearly:

Updated map and summary:

Goodbye to 332-206 for now. Perhaps we will see you again soon.

And now to the “Tipping Point Margin”:

This looks like Obama continuing to increase his lead in the tipping point, while the tipping point moves to Virginia. But… I urge caution… this is due almost entirely to what looks like a clear outlier in the Virginia polls that showed Obama up by 17% in Virginia! 17%! That outlier pushed Obama to a 3.5% lead in Virginia. But that new poll is way out of line with every other poll in the state. A 17% Obama lead in Virginia is simply not believable. (This was a registered voter result rather than a likely voter result, but still…)

I don’t manually remove outliers, but if you did, then Obama’s lead in Virginia drops back to 0.5%, which is probably closer to reality. If you did this, the tipping point state would still be Nevada rather than Virginia, and the margin in Nevada is still 3.2%. So the chart above is showing a 3.5% lead in the tipping point state (actually 3.46% before rounding) rather than 3.2%. That last little gain is an illusion due to the outlier. The outlier will wash away soon enough, in the mean time, this isn’t really that big a difference given the noise in this metric.

Either way, the picture has stayed consistent for the last few weeks. Obama has a slight lead in the electoral college as predicted by current polling, and that lead seems to generally be getting more solid. But that lead is still by no means secure. A 3.5% (or 3.2%) lead in the tipping point state is much more solid than the 1.0% lead Obama had on October 10th.

But 3.5% is also not all that secure. A 3.5% lead could disappear in a day or two with a bad news cycle for Obama. That seems less likely as we get closer, but it is still quite clearly not impossible. And as Romney supporters have been particularly fond of pointing out lately, there is also the possibility of systematic error in the polling, even when you average across many pollsters. It has happened before (most recently in the 1996 Clinton/Dole race), and it will happen again. Maybe it will be this year.

If there are no major changes in the few remaining days before the election, we go into the election with Obama a heavy favorite, but with a Romney upset a long shot, but still within the realm of reasonable possibility… which is where we have been most of the last year.

Note: Eight hours or so between my daily poll sweep and the blog post today. Polls released during that time period will of course be included in tomorrow’s update.

Note: Chart and map from the Abulsme.com 2012 Electoral College Prediction page. Both assume Obama vs Romney with no strong third party candidate and show polling as it currently exists. Things will change before election day. On the map red is Romney, blue is Obama, gold states are too close to call. Lines on the chart represent how many more electoral votes a candidate would have than is needed to tie under several different scenarios. Up is good for Obama, Down is good for Romney.