Tuesday, September 2, 2008

And What About the Green Party?

This past week we've looked at how both the Libertarians -- behind Bob Barr's candidacy -- and Ralph Nader would fare in the upcoming November election given the information we have at this point in the race. The obvious goal is to see if, in a close race, either would pull enough of the vote away from McCain or Obama to affect the outcome of the race. The conclusion we've drawn is that in the most competitive states there are a few instances where either third party candidate could affect the race, but that across the full list of toss up states, it isn't likely to make all that much difference.

But are we getting the true Nader effect? Often we talk about the potential for a third party candidate to affect the fortunes of one or both of the major party candidates. What we don't discuss is how third party candidates can affect each other. Nader ran as the Green Party nominee in 2000 but as an independent in 2004. In 2004, however, there was something of a redistribution of 2000 Nader voters. Some returned to the Democratic Party having been spurned by the Nader and the perception that he cost Gore the presidency, others followed Nader and some stuck with the Greens and their nominee, Daivid Cobb. [Yes, I'm sure some stayed home as well, but we'll gloss over those folks.] The Nader effect can, then, be thought of as a Nader/Green effect to some extent. So what we really have here are two questions:
1) Does the separation/combination of the two add anything to our understanding of the effect they are predicted to have?

2) Does Cynthia McKinney's presence in polls -- and later on the ballot -- augment the effect and/or detract from Nader's vote share?
Reader (...and political scientist/blogger), MSS, asked about the latter in the comments to the second Barr/Nader post from last week, and as I said in response, there just isn't enough information out there on McKinney yet to draw any firm conclusions. The handful of national polls that include her give her an average of about .67% (not exactly lighting the world on fire). And the only state polls that have included her are the four released last week by CNN in Colorado, Nevada, New Mexico and Pennsylvania, which give the former Georgia congresswoman between 0 and 3%. [It should also be noted that all of the McKinney polling comes from just two polling firms, CNN and Zogby Interactive. We don't, therefore, have the benefit of the view of this from several different angles.] There just isn't at this point enough evidence to state that her presence in the race is having any effect on Nader or anyone else. Part of McKinney's problem in November will be ballot access. It plagued the Green Party in 2004 without Nader as the standard bearer and that trend looks to be repeating itself in 2008. A quick glance at the vote results on Dave Leip's US Elections Atlas shows that the David Cobb was not on the ballot (...or simply written in) in 23 states in 2004. You can't have an effect if you aren't on the ballot. Ballot Access News (and click on "View State-by-State Chart" for more information) confirms as much for 2008 as well. McKinney will be on the ballot in 27 states in November as of September 1. She is in court or petitioning for access in 6 other states. Whether she's on the ballot in those six states or not, there is a sizable enough number of states where not being on the ballot will have an effect on how much influence she could have in the race.

Now, I'd like to report that I ran the same sort of model I ran on the other third parties -- and I did -- but the lack of data really screwed up the resulting model. For starters, the relationship between the prediction and the 2004 Green Party vote share was negative; the more support in 2004, the less McKinney would get in the fall election. And that makes sense, right? But the lack of polling data for McKinney during this cycle is simply too sparse at this point. That may change, but at this point getting an accurate prediction just wasn't going to happen. [And yes, that same charge could be levied against the models run for the other third parties. As I argued last week though, the goal was to develop a rank ordering of where each is most likely to have an effect. But I digress.]

What we can do is address the first question posed above: predict a Nader/Green effect for November.
[Click to Enlarge]

If you glance back at the Nader scatterplot from the original post, you can see that the basic ordering of the toss up states is not fundamentally different when the Green data is added into the mix. All that we're really doing is raising the bar a bit. Instead of a series of predictions for Nader just under two points, you have a series of predictions for the Nader/Green vote that ranges from three and a quarter points to three and a half points. Those three states at the upper right of the plot are likely to be states where the Palin selection will play well -- especially in Alaska. That is certainly a point that is up for debate since Palin is "not a hit with undecideds" and is back on In-Trade, but the trading is over whether she'll be dumped as the VP choice. That aside, McCain still has the advantage in that trio of states. Is the Green/Nader presence hurting there? Well, Nader maybe, but McKinney isn't not on the ballot in two and is short over 6000+ signatures in North Dakota. So, perhaps it isn't a factor.

Recent Posts:
It's Never as Easy as Taking Away Half the Delegates

The Electoral College Map (8/31/08)

From Wyoming: An Answer to the "Will the GOP Sanctions Have Teeth" Question

No comments: