My posting has been light of late, for a couple of reasons. The first is the obvious one: grad school is a lot of work. I just finished my third test in as many weeks, and putting regular homework and teaching duties on top of that doesn’t leave a lot of time for blogging.
The other reason is that the main things occupying my mind (other than school) of late are things outside the normal scope of this blog: the Presidential election and the start of the NBA basketball season. However, at the risk of being self-indulgent, I’m going to talk a little about each of these topics, beginning with the election in this post. And, to keep it somewhat in the lines of this blog, both topics will be related back to statistics and computer simulation.
I, like many people got addicted to election news in the few weeks leading up to the election. Hourly checks of fivethirtyeight.com became a near addiction:
This addiction had a variety of sources: the first candidate I could ever really get excited about in my lifetime, the fact that my home state of Indiana was actually a swing state, etc. But it was really driven by the content on fivethirtyeight.com: not just opinion polls, but computer simulations of the outcomes, and projections based upon the data.
The projections were interesting, even though my general distrust of statistics prevented me from fully believing the models. I think pollsters and the news outlets that report their polls put way too much stock in their numbers even though they usually have a 4 percentage point margin of error due to sample size. They also try to correct for systematic errors in their polls (a fact that I learned from reading fivethirtyeight), but there is no clear way to do this.
I did find the projections on fivethirtyeight.com to be worthwhile, because they at least recognized these limitations, and their weighting system and projection model tends to smooth out the inconsistencies between polls.
Another interesting thing that I learned from this website was the importance of the “ground game” of the campaigns: calling potential voters, getting multiple contacts in person, and a final get out the vote push the weekend before the election. This apparently makes a huge difference, and was credited, in large part, for Barack Obama’s success. Notably, Obama squeaked out a win in Indiana, where he had a much larger presence on the ground than John McCain did.
While I was interested to learn about this part of politics, I also found it somewhat unfortunate that this is so important. I would like our elections to be decided based upon the issues, and to have our elected officials to be the ones that most of the electorate thinks would do the best job. I would think that most people who chose based upon the issues made up their minds at least by the final debate, if not before.
This leaves a crucial part of the decision to people who are influenced by things like TV ads and campaign calls. I will not say that these people should not vote, because I believe in democracy far more than I believe in any liberal or conservative ideology. I simply wish that this were not the case.