Showing posts with label descriptive statistics. Show all posts
Showing posts with label descriptive statistics. Show all posts

Tuesday, March 17, 2009

More on 2008 Candidate Visits

Let's look a bit deeper at the 2008 presidential candidates data I posted the other day. Instead of looking at it in terms of raw percentage shares for each state, I grouped the states according to the size of those shares. These cut points are rather arbitrary, but it does drive home the point. Recall that if all states were created equally, each would receive two percent of the total number of candidate visits in each presidential cycle. And just by eyeballing it, that 2% mark is a good cut off for the uppermost grouping. That group, then, is comprised of those states that garnered at least a two percent share of total candidate visits. I wanted to allow for another couple of categories, but no natural breaking point really emerged. What I did was to set the upper bound of the bottommost grouping at .5%, but as you'll see below, an additional category could have been created to provide for further differentiation. [I'll discuss this a bit more below.] The resultant three groups, then were states that received a:
  1. 0-.5% share of 2008candidate visits
  2. .5%-2% share of 2008 candidate visits
  3. greater than 2% share of 2008 candidate visits
If we look at the picture overall, here's how the states fall into those groupings:

[Click Chart to Enlarge]

In other words, only nine states got anything more than a two percent share of candidate visits (including both parties' candidates). Over three times that many states had a less that half a percent share of overall visits during the 2008 primary campaign. Now, this lowest category could further be broken down into states that got less than .25% of visits and between .25% and .5%. That would basically split that group in half with the former group containing 14 states and the latter, 16.

Looking at those below that .25% line, half (7 states) had concurrent Democratic and Republican contests on Super Tuesday. Of the other seven states, four had their Democratic contests on Super Tuesday while the Republican Party's contests came later (after the point at which McCain had wrapped up the nomination). There are several factors at work here. First, size is a common theme among these seldom-visited states. I'll use electoral votes as a proxy here. Of those 14 (<.25%) states, all fell at or below the 10 electoral vote line and nine have five or less electoral votes. The other obvious points here are that competition for candidate visits matters, and so too does the fact that a contest may fall after the point at which the nomination has been decided. Small states already fighting for attention are even more up against it when there are, say, 25 other states going on the same date. The saving grace for those Super Tuesday small states is that their voters at least had the opportunity to weigh in on both nominations. And while those states with split GOP contests (the ones with Democratic contests on Super Tuesday) were able to avoid the competition for attention, they missed out on the attention altogether by being so late in the process. The voters in those states were in a lose-lose situation. But shifting back to those top nine attention-grabbing states, we see that they accumulated 84% (57% in Iowa and New Hampshire) of the total amount of attention. Now granted, the advance build up of visits in Iowa and New Hampshire in the year(s) prior to the presidential election year skews these figures to some extent. However, when the Iowa and New Hampshire visits are dropped altogether, there are still only 11 states overall with visits shares over 2%. That's a net gain of two states in that category, but the cumulative share of visits to that group of states now drops to just under 73% of the total. Whether Iowa and New Hampshire are withheld does not change the fact that this group of states had one or more of three basic properties. These states were early, big and/or the only event on a given date. Ah, but what happens when these figures are separated by party?
[Click Chart to Enlarge]

The distribution of visits across Republican primaries and caucuses didn't stray too far from the overall distribution above. But once the same procedure as in the above example is employed there are some subtle differences under the surface.

If the large collection of seldom-visited states is split along the .25% line, 18 of the 31 states fall below that line and 13 above it. The thing about the Republican nomination race was that it conformed for the most part to previous nomination races, and that leaves us with two main sets of contests: the compressed states held prior to the nomination being decided and those that are more spread out yet fall after the contest is over. Those are the two categories represented by an overwhelming number of those 18 states below the line. Just three of those states weren't either on Super Tuesday or after March 4 when John McCain became the presumptive nominee of the Republican Party.

On the flip side of the chart, those 7 states in that (>2%) category made up 84% (52% in Iowa and New Hampshire) of the Republican candidate visits. That matches the percentage in the overall case. Yet, if Iowa and New Hampshire are dropped from consideration, there are only six states that exceed that 2% level and they only comprise just under 70% of the total Republican visits.

[Click Chart to Enlarge]

In the Democratic contests, the distribution looks pretty much the same, but there is a trade-off between the lower two categories, with the the two being closer in frequency than they were in the previous two examples. Splitting the lowest category along the .25% line doesn't have the same effect as it did in the previous two instances. 20 states fall below that point and just six above it. Three-quarters of those 20 very seldom-visited states were on Super Tuesday. And that is telling. Since the Democratic race extended to the final contest, many more states had an opportunity to have attention that otherwise would not have. The states that paid the price, then, were those in the most compressed environment, Super Tuesday.

And the attention-grabbing states? Well, those seven states received just under 80% (62% in Iowa and New Hampshire) of the Democratic candidate visits. And that number hardly changes when Iowa and New Hampshire are dropped. However, double the number of states fall into that (>2%) category when the two lead-off contests are withheld. Those fourteen states make up 76% of the Democratic contests. Again, that speaks to the longevity of the competition on the Democratic side. The tie that binds those contests (with or without Iowa and New Hampshire) is the fact that most were stand-alone contests or on a date where there was far less competition for attention.

NOTE: I have to confess that I've put most of these last two posts together for an exercise on descriptive statistics that I'm doing in one of my classes. But I thought I'd share and provide a bit of background information in the process. That will help us down the road if I get around to doing a projection model for 2008 based on candidate visits.


Recent Posts:
2008 Presidential Candidate Visits by State and Party

Should Indiana Frontload in 2012? (Part Two)

Michael Steele by the Numbers