Thursday, February 11, 2010

A Look at Silver’s Analysis of Palin-2012

Summary:’s Nate Silver explores Sarah Palin’s 2012 presidential primary chances, but leaves out an important variable.

For those who haven’t seen it yet, Nate Silver’s analysis of Palin’s 2012 chances is quite interesting. While long and in-depth, he makes some good points.

From the post:

So where is Palin likely to run strongest? Obviously, it would depend on the candidates she's up against and the type of campaign she might want to run, but I think we can make some basic inferences. What I've done is to create an index of how favorable each state is to Palin based on six variables: fundraising totals to date for SarahPAC and five demographic and attitudinal variables taken from 2008 exit polls.

Fundraising: What I looked at is the ratio of contributions that Palin has received in each state so far through SarahPAC to the amount of contributions received by all Republican candidates in the 2008 cycle. The idea is to see how Palin compares vis-a-vis a typical Republican candidate -- indeed, I've found fundraising data to have quite a bit of predictive power in the past, even if the data is a little rough at this stage. Relative to other Republicans, Palin's best fundraising state is, of course, Alaska. Her fundraising has also been quite strong in the Pacific Northwest, and many of the prairie/frontier states. It has been weakest on the East Coast, as well as several other urban and industrial states located throughout the country. The fundraising data receives double weight in our index.

Variables from 2008 Exit Polls: What I looked at is not what a state's electorate looked like overall, but rather, the characteristics of the McCain (i.e. Republican) voters in each state. This is an important distinction -- for instance, although Oregon is a fairly progressive state overall, the conservatives there are quite conservative and rural, and this is what matters in the context of a Republican primary. Note that, although it would probably have been better to look at exit polling data from the 2008 Republican primaries, a lot of states either didn't have a competitive primary in 2008 or didn't have exit polling data available; thus, we look at McCain general election voters as our best proxy.

Specifically, the exit polling variables that I evaluated were as follows:

Rural and small town voters. That is, the percentage of McCain voters in each state that come from communities of less than 50,000 people. Palin spent a great deal of time campaigning in exurban and fairly rural areas in 2008, and I suspect that it's here -- not necessarily among soccer moms in the collar suburbs -- where her most enthusiastic voters lie. And Palin herself, of course, comes from a very rural area and is appealingly outdoorsy and self-reliant. This variable receives a double weight.

No college voters. Early polls of the 2012 Republican field, such as from Marist and Rasmussen, show Palin overperforming among this group (or, if you prefer, underperforming among college graduates), which certainly squares with my intuition about where her appeal lies. This variable also receives a double weight.

Conservatives. We also look at the number of McCain general election voters who described themselves as conservative in each state, although it receives only a single weight. Although clearly Palin wears the conservative label very comfortably and is liable to be harmed in states where there are a relatively large number of moderates and independents in the primary electorate, there are likely to be at least a couple other capital-C conservatives in the Republican primary field, which means we need to temper this somewhat.

White Evangelicals. Although Palin also polls well among this group, a lot of this may be because a lot of white evangelicals are also rural and lack a college degree. That is, although Palin runs well among the sorts of voters who happen to be evangelicals, it may not be because they're evangelicals. Nor, although Palin has increasingly invoked religious rhetoric in her speeches, does she have the scholarly religious credibility of someone like a Mike Huckabee or a Pat Robertson. It's conceivable that Palin could get outflanked by a Huckabee or lose votes to a Santorum among voters who are evangelicals first and working-class whites second. Thus, although we include this variable, we only give it a single weight.

Energy and Terrorism voters. Lastly, although this is a bit speculative, we look at the percentage of McCain voters in each state who said their votes were determined because of energy or terrorism policy, which appear as though they'll be Palin's core issues. These issues -- particularly terrorism -- lend themselves relatively well to the meta-narratives that Palin prefers and require less policy nuance than something like the economy or health care. This variable receives a single weight.

He then runs the data and comes up with chart showing how well Palin should theoretically do on a state-by-state basis. In the map below, the most red states is where she is strongest and the most blue is where she is weakest.

Next - taking other candidates’ strenghts and weaknesses into account - he gives his outlook for how Palin could secure the nomination. (For clarification on the colors he mentions, see the map at the bottom of this post).*

Palin's path to victory, then, would seem to consist of one of the following scenarios:

Palin Plan A. Win Iowa. Win South Carolina. Clean up in orange states. You probably have enough momentum to survive the consolidation of the GOP field which is liable to occur at this point.

Palin Plan B. Lose Iowa narrowly, especially to a Midwestern candidate. Hope that a Southerner isn't running strongly and win South Carolina. Clean up in orange states. Then you anchor in the South, winning Texas (green group), Florida/Georgia (gold group) or Indiana/North Carolina (purple group). At some point, you need to break through and win a big Midwestern battleground like Ohio or Wisconsin.

Palin Plan C. Win Iowa. Lose South Carolina narrowly to a Southern candidate. Regain momentum in orange states. Hope that green states vote next and aim in particular for a big win in Texas. If it's the gold states instead, go all-in in Ohio and Pennsylvania. If it's the purple states, you'll need some help.

Among the other possible candidates he mentions:

- Former Governor Mitt Romney (R-MA)
- Former Governor Mike Huckabee (R-AR)
- Former House Speaker Newt Gingrich (R-GA)
- Senator Scott Brown (R-MA)
- Governor Haley Barbour (R-MS)
- Senator John Thune (R-SD)
- Congressman Mike Pence (R-IN)
- Governor Rick Perry (R-TX)

This set off a huge red-flag for me. Last year, I wrote about Governor Tim Pawlenty (R-MN) being a particularly strong candidate to win the GOP nomination in 2012.

Looking at the map, who is better positioned to win both Iowa and New Hampshire than Pawlenty? Nobody. Romney can win New Hampshire, but probably not Iowa. Palin and Huckabee could both win Iowa but probably not New Hampshire. (This is still according to Silver’s analysis).

Pawlenty, on the other hand, has contacts within the Iowa GOP and is better positioned to use the party infrastructure early on. Because of his work in Minnesota, he is a very strong candidate in the agricultural Midwest. In fact, I know the Wisconsin GOP already gave his Freedom First PAC access to their donor list.

Additionally, Pawlenty doesn’t have the same “uneducated” style of politics about him (for lack of a better term, I swear) that candidates like Palin or Huckabee do. This will give him strength in a state like New Hampshire, which otherwise seems to be a given for Romney.

Sure, he probably can’t win Nevada or South Carolina as easily, but he’d be in great position for the next few states (a few of which boarder his own) and the momentum he builds could easily deliver him a victory.

Obviously, we’ll have to wait and see, but I think it was a mistake on Silver’s part to ignore Pawlenty’s presence come 2012.


The first states to vote are the traditional early states of Iowa, New Hampshire, Nevada and South Carolina. These states are shown in light blue. Note that this list does not include Florida and Michigan, which jumped in the queue to try to vote early in 2008 -- although who knows whether they'll be in a more cooperative mood this time around when push comes to shove.

Next to vote are the orange states, which are grouped together by virtue of their small populations. This includes 14 states and several territories, the largest grouping of which is on the prairies and the Western frontier, although there are also several New England states. Notably, no Southern states vote in this group -- the Republican calender definitely de-emphasizes the South.

Finally, there are gold, purple and green groupings of some of the larger states. These groups have some geographical integrity -- for instance, most of the traditional Midwestern Rust Belt states are in the gold group, whereas the purple states tend to be more coastal and the green more in flyover country -- although there are some exceptions. The order in which the gold, purple and green states vote will rotate every cycle and, to my knowledge, has not yet been determined for 2012.

No comments: