Tuesday, May 4, 2010

And since I have already left the path: a bar graph

Once I have freed myself from the tyranny of confining myself to scatter plots, interesting questions have raised themselves. This one is demonstrated by a bar graph.
One of my conclusions about the "Core Electoral Votes" was that they showed that the election was indeed becoming more polarized. After thinking of ways to express this, I decided to make bar graphs showing the number of electoral votes that Obama scored a certain percentage of the popular vote in (This sentence does make sense). I then tried to find another electoral winner who had a similar-sized victory, and tried to break down their electoral votes down similarly. Although it is far from a good comparison, the best comparison to Obama's '08 victory was probably the first Bush, in '88. (2 candidates, similar numbers in popular and electoral votes).p some of the conventional wisdom:
This diagram confirms some conventional wisdom: elections are indeed becoming more polarized, at least as far as different regions of the country having different values. Bush in 1988 got more electoral votes than Obama did in 2008, but a great of them were in a narrow band between 50% and 55%. On either side of that, he fell off, getting below 45% in only a handful of states, and also getting over 60% in only a few states. The bar graphs form a triangle, of sorts. Obama, on the other hand, scored all over. His largest category for electoral votes was actually over 60%, and there are three separate peaks with two separate valleys. Even given his good popular vote and large electoral vote margins, there were areas that didn't move at all. The middle ground, between 45-55%, is much smaller than sense says it should be.

I think this does speak of polarization: in 1988, Massachusetts and Oklahoma might have different results, but the results would move together. Now, that is not the case.

Monday, May 3, 2010

Since I seem to have abandoned the "daily" thing, lets abandon the scatter plots as well.

So, I slowed down updating this, and haven't updated this in two weeks.

This is mostly because I feel I have exhausted a lot of what I wanted to talk about. Actually, I should at some point summarize what I have learned, but I don't have a lot more to say about most of what I was covering.

But, I did, on a lark, think of a question that can be used with another type of graph. And so, I present a non-scatterplot to you.

In the US election system, the president is chosen by electoral votes. The electoral system can greatly magnify a candidates success or failure. Also, electoral votes can be won by plurality, meaning a candidate can win the presidency in a landslide without actually winning any of the states by a majority. Clinton did this in 1992, only winning 9 electoral votes (Arkansas and DC) by majority, but getting 370 Electoral Votes.

So I decided to look at the history of how many Electoral Votes were won by over 60% of the vote. These show states that were won with what could be seen as a strong consensus. So from 1960 until 2008, here is the fate of both party's ability to truly capture states:
Much as with my scatterplots, the first lesson to be learned from this graph is unpredictability. The biggest lesson seems to be that holding on to Core Electoral Votes is very difficult. Even after the biggest landslides ('64, '72 and '84), the amount of core electoral votes drops. This also leads some credence to the belief that elections are more about candidates and circumstances than they are about the deep seated philosophical leanings of the electorate. Was Reagan's victory in 1984 a sign of a deep seated conservative bent in the US? According to this diagram, it would seem not, because just 4 years later, the amount of states showing a really strong commitment to support the Republican candidate shrunk down to just a few in strongholds. Of course, the same could be said of the movement from 1964 to 1968.

Another trend that actually shows up from popular political discourse in this is that some of the "Red State/Blue State" and "polarization" seems to have some evidence for it. In previous elections, one party might get a lot of Core Electoral Votes, or both parties might get none or close to none... but only since 2000 have both parties managed to have strongholds. So there is some truth to it: the current political situation is one where, regardless of candidate, Massachusetts will probably go over 60% for the Democrat, and Oklahoma will go over 60% for the Republican.

Of course, since the unexpected is expected, I bet 2012 will have some interesting changes to make to this chart.