Tuesday, July 17, 2012

The Electoral College Map (7/17/12)

With just 16 weeks left for candidates to win -- 112 news cycles to spin -- before election day, here is what the electoral college map looks like using FHQ's graduated weighted average.

Now, FHQ will not focus too extensively on the electoral vote numbers. The intent at this juncture to create the baseline from which future updates can and will be compared. The electoral vote tally is secondary -- at least for now -- to the graphic below. The Electoral College Spectrum translates the weighted averages (more on those in a moment) into a ranking of states from the most strongly-aligned with Barack Obama to the state most firmly behind Mitt Romney. Of course, for our purposes, all the interest is not in those easy to peg states so much as the states that will decide the election within the electoral college. In 2008, Obama pushed toward the bottom of the middle column of states and was competitive into the next rightward column (states like Georgia and Arizona).

But in 2012 the terrain is different. The battle is still being waged overly closely competitive states in the middle column of the Electoral College Spectrum, but instead of a push to the right, the overall tilt is back toward the left -- bluer or traditionally bluer territory on the presidential level. Michigan, for instance, is much closer in the FHQ weighted average than it was in 2008. And that brings this discussion back to the utility of the spectrum: It gives us a glimpse into the relative arrangement of states in terms of the margins separating President Obama and Governor Romney in a weighted average of state-level polls. The common refrain from political scientists is that this process is like a pendulum swinging back and forth from cycle to cycle; that states will, overall, see a similar shift in the aggregate. But the underlying assumption is often that while the partisanship of the electorate may shift which state falls into which party's column in any given election, the alignment of states is largely similar. That is certainly not wrong. A glance back at 2008 shows that Rhode Island and Hawaii are every bit as much Obama states as Utah and Wyoming are Romney/Republican states. The same is true for the group of states that will see so many candidate visits and never-ending television ad buys over the next 16 weeks.

However, the relative ordering is important and so are some of the individual state shifts. Take aforementioned Michigan. The Great Lakes state is now essentially right where it was in the ordering in FHQ's final electoral college projection on election day 2008. But it is a lot closer in 2012. Iowa, on the other hand, is also closer in 2012 relative to 2008, but it has shifted over an entire column on the spectrum. That is significant in all of this.1 It also adds a layer to our understanding of the shifts or prompts at the very least to ask why one state has moved so much and another has not.

[For more on how to read the Electoral College Spectrum, please follow the link in the figure's footnotes.]

The Electoral College Spectrum1 
1 Follow the link for a detailed explanation on how to read the Electoral College Spectrum.

2 The numbers in the parentheses refer to the number of electoral votes a candidate would have if he won all the states ranked prior to that state. If, for example, Romney won all the states up to and including Ohio (all Obama's toss up states plus Ohio), he would have 272 electoral votes. Romney's numbers are only totaled through the states he would need in order to get to 270. In those cases, Obama's number is on the left and Romney's is on the right in italics.

3 Ohio
 is the state where Obama crosses the 270 electoral vote threshold to win the presidential election. That line is referred to as the victory line.

Finally, one additional figure helps to parse this out further: the Watch List. These are states that are a fraction of a percentage point in the weighted average from shifting categories. Florida, for example, is within one point of shifting into Mitt Romney's column. If one is looking for new polls that might have an effect on which states fall into which category (strong, lean or toss up), these are the states to be on the look out for.

The Watch List1
from Toss Up Obama
to Toss Up Romney
from Strong Romney
to Lean Romney
from Toss Up Obama
to Lean Obama
from Toss Up Romney
to Lean Romney
from Lean Obama
to Toss Up Obama
New Hampshire
from Toss Up Obama
to Lean Obama
North Carolina
from Toss Up Romney
to Toss Up Obama
West Virginia
from Strong Romney
to Lean Romney
1 Weighted Average within a fraction of a point of changing categories.

Methodology (quick hits)2:

  • Everything above is based on a graduated weighted average of polls in each state conducted in 2012. The weighting is based on how old a poll is. The older the poll is the more it is discounted. The most recent poll is given full weight. The main difference between what one sees here and what FHQ did toward the tail end of the 2008 cycle is that the collective weight of the polls other than the most recent one are not halved. The reason for the change is based on where we are in the process, or more accurately, how many polls have been conducted. Right now, there just are not enough polls in each state and the full weighted average is needed to ward off the influence of outliers that may arise. Later on, say, post-conventions most likely, we may have enough polls serving as a sufficient anchor on any outliers. There can be a fine line between the responsiveness of the averages and wild fluctuations. The way things are constructed, the average ends up being fairly conservative. If a change occurs, then, it is usually a lasting change.  
  • Strong states are those states with an average margin between Romney and Obama of over 10%. Lean states are those between 5% and 10% margins. Finally, toss ups are those states under 5%. Admittedly, these distinctions are somewhat arbitrary. There is enough space between New Hampshire and Michigan and the other Obama toss ups that those two states could be moved into the lean category. The same is true -- probably truer -- of Missouri on the Romney side of the ledger. There is a wide berth between Missouri and North Carolina. The 10% does appear to be a pretty good cutoff between lean and strong states though. One other addendum is that these cutoffs will be adjusted downward as we approach election day. If memory serves, the toss up threshold was moved to under 4% sometime between the conventions and the first debate and to under 3% after the final debate. Those moves are made under the assumption that greater distances are more difficult to make up for a candidate the closer it gets to election day.
  • It should be noted that there are still a host of states where no polls have been conducted. For those states and for those states with just one poll, the averages of the final margins from the 2000, 2004 and 2008 elections were used. Those three elections were averaged with the lone poll in the states with just one poll conducted thus far. FHQ is not totally comfortable with this (see particularly Arkansas and West Virginia), but it is the best that can be done in this simple framework. There is plenty of complex out there, but we like to keep it simple here if only to stack up a rather parsimonious model against some of the other more complex ones at the end of the process. 
  • Updates? If polling is trickling in at least somewhat regularly I will try to update every day even though that level of posting isn't really necessary at this point. Look, I've to make amends for starting this in July somehow. 

Comments, questions, things I missed? Please feel free to chime in on Twitter (@FHQ, hashtag #ec2012) or in the comments section below.

1 Of course, there are some issues with that comparison. First of all, the comparison points in the race and thus the amount of polling data available in July versus November is vastly different to say the least. Why, just this morning, FHQ was bemoaning the lack of polling in Iowa of all places. For a state as seemingly close as it is, it has been polled very sporadically in 2012. Secondly, one could also take issue with FHQ using its weighted average instead of the actual election results where Iowa was much closer to its current position. That is a fair point. Both datapoints are useful. Plus, that points out a potential shortcoming of the polls and the resulting weighted average (with more emphasis on the latter).

2 For a more extensive look at the methods behind the above figures please feel free to check out the FAQ from 2008. The underlying methodology is largely similar.

Recent Posts:

Are you following FHQ on TwitterGoogle+ and Facebook? Click on the links to join in.

No comments: