Legal Boxer Briefs :: Jewsade for Justice

According to RealClearPolitics, Barack Obama has an 8.7 percent lead over John McCain. However, this is only when Ralph Nader and Bob Barr are included in the polling data. When third party candidates are excluded, however, Obama’s lead shrinks to 6.8%. There are a variety of explanations for this.

One of these explanations is simply that Bob Barr simply siphons off voters who would otherwise be voting for McCain. However, this, at least by itself, is entirely inadequate. In fact, in a partial restoration of my faith in the American people, only 1% of Americans are projected to vote for him. Nader, meanwhile, is polling at 3%. If we accept political science orthodoxy that a third-party candidate will only draw votes away from the candidate closest in ideology to them, then Obama should be ahead by around 10.7%.

I got this number by adding the 3% of Nader, which would theoretically go to Obama, and subtracting the Barr’s 1%, which would theoretically go to McCain. Whille the Libertarian Party shares views with both major parties (it is socially liberal and [very] economically conservative), this year’s candidate is ultraconservative Bob Barr of Georgia, known for his both his economically and socially conservative douchebaggery during his time in the House of Representatives (for example, he both authored and sponsored the Defense of Marriage Act. He also proposed to ban the practice of Wicca in the military and wants to repeal the 16th amendment [which instituted the income tax]).

An alternative explanation is that there are simply fewer polls, which leads to less accurate overall statistics. The problem with this, however, is that a sheer number of polls will not ensure statistical integrity. In fact, the sample sizes of most of these polls have been around 1000 likely voters. With sample sizes like this, it is rather unlikely that the margin of error will be particularly large.

On the other hand, the accuracy provided by the sample sizes might be offset by the varying definitions used for “likely voter”. Gallup has experimented with using two different designs for its “likely voter” polls. The results of one experimental poll were as follows. Among “registered voters”, Obama led McCain by 11 points. Among the traditional definition of likely voters (that is, it takes into consideration “voting intention and self-reported past voting behavior”), Obama was in the lead by 7 points. Among the expanded definition of likely voters (people who claim they will vote on election day), Obama was in the lead by 10 points. There is a problem with the assertion that the “likely voters” model is less accurate. For the registered voters in this poll, the MoE is ±2%; for the traditional LV model the MoE is ±3%; for the expanded LV model, the MoE is ±2%. In other words, the margin of error is essentially the same, but the likely voter interview is more likely to predict the outcome of the election.

Another reason to believe that the inclusive polling (i.e. with Obama, McCain, Nader, and Barr) is accurate is that it has followed the same pattern as the McCain v. Obama polling, yet only slightly delayed (because there are simply fewer inclusive polls).

The explanation that I find most credible is that Nader voters are simply people who might not otherwise vote this year. Obama’s numbers are essentially the same across the two polling methods (49.0% average in the inclusive polling versus 50.1% in non-inclusive). In fact, if you add up Obama’s and McCain’s numbers in the non-inclusive polling, there’s a missing 6.7% (presumably, “undecideds”). Given that Obama’s numbers stay the same across the polling data, yet his lead only grows when Nader and Barr are included, my hypothesis is this: while some McCain voters are willing to bite and vote for Bob Barr (either because they don’t like McCain’s social conservativism are are, in fact, libertarians, or because they don’t think he’s socially conservative enough and are simply fans of Bob Barr), Obama voters are committed to their candidate. This argument is justified by polling data which indicates that Democrats are far more excited about Obama than Republicans are about McCain. At the same time, the 6.7% of unlikely voters will split between Nader and Obama for two reasons. First, because CBS polling indicates that undecideds and independents favor Obama by wide margins. Second, because many Nader supporters have said that they simply wouldn’t vote if he weren’t on the ticket.