Friday, November 19, 2010

Political Divisions of the U.S. House of Representatives

I became curious about the history of the political division of the U.S. House after some of the media discussion about what a historic change had occurred in the last election, and how unstable elections had been recently. As it turns out, neither assertion is true.

The chart below shows the election results from 1867 through 2010. These data were obtained from the web site of the Clerk of the House.

I’ve included on the chart the approximate times of depressions in the U.S. obtained from Wikipedia . I’ve shown the recent major “recession” in gray rather than black because it officially hasn’t been labeled a depression. I’ve also shown the major wars and a sampling of presidents to help orient the viewer. The sizes and location of the graphics on the time line are only approximate.

As you can see, the last 50 years is not nearly as unstable as the first 80 years. And, the recent swing from democrats to republicans in the House is not an extraordinary change at all. Eight other times in this history the change has been larger, and fifteen times it’s been within five percent of this year’s change.

The pattern of these changes indicated to me that it might be representative of a complex system at work. Analysis of the change pattern resulted in the development of a power curve. A power curve is one indication that the system might be a system exhibiting self organized criticality, like earthquakes and markets. The results of that analysis is show in the graph below. This analysis indicates that the probability of occurrence is inversely proportional to the magnitude of the change. However, the correlation coefficient is not as high as in other examples I have worked with.

The calculated probability of a change being greater that the most recent change is 21%.

This analysis is not conclusive evidence that the election system for members of the U.S. House of Representatives is a self organizing critical state system, but it does hint that this might be true.

U.S. Wars


  1. Very interesting. Just happened that Flowing Data today posted a link to a different political analysis that you might be interested in : by David B. Sparks. Enjoy!

  2. Erika,
    Thank you. That's a very interesting analysis. Watching the video reminds me how a massively parallel models of agents evolve. Models of infectious diseases and fires look like his models. It makes me wonder if the voting public isn't operating somewhat like flocking birds as well. Each agent operating with simple rules and influenced by what their neighbors do.