The diagram "US Senate Voting Similarity Networks, 1975-2012" visualizes relative differences and similiarities in U.S. Senate voting records over time. Use this tool to zoom in and navigate around the diagram.
"Sequestration" budget negotiations, gun policy debates and election results offer anecdotal evidence that Americans are increasingly divided into two political camps.
Now researchers at Duke University and the University of North Carolina-Chapel Hill have been able to visualize an aspect of that polarization by analyzing Senate voting records dating back to 1900.
Their conclusion: "We have not seen the current level of partisanship since the early 1900s," said professors James Moody of Duke and Peter Mucha of UNC in their paper, "Portrait of Political Party Polarization."
In their study, published in the inaugural issue of the journal Network Science, Moody and Mucha examined how each senator voted on each bill in a given two-year period. They then applied statistical methods to group politicians into Democratic and Republican voting blocs based on the similarity of their votes.
Olympia Snowe's changing political affiliations over time are charted in the study.
Senators who did not consistently vote with either party were designated as occupying a middle ground outside of either party’s camp. Each party voting bloc, as well as each individual senator falling outside of a bloc, was arranged based on the underlying polarization score. The higher the score, the more disparity there was in voting records.
The results of the analysis are presented in a diagram, "US Senate Voting Similarity Networks, 1975-2012" (download the diagram here). Voting data from before 1975 are not included in the main figure.
The trend is clear: The difference in voting behavior between the two parties increases over time, while the number of senators voting independently from their parties decreases.
"In the current era, middle positions seem fragile and even longtime middle residents follow party lines (Jeffords) or lose their seats (L. Chafee)," Mucha and Moody said, citing the examples of senators James Jeffords of Vermont and Lincoln Chafee of Rhode Island.
Because the study examines the voting patterns of each senator in each Congress, the careers of individual senators can be tracked over time. Take the example of Sen. Olympia Snowe, a Republican from Maine who cited excessive partisanship when she announced her retirement from the Senate last year. When Snowe first entered the Senate in 1995, the study shows she voted with her party; however, in the following session of Congress, she broke out as a moderate. Then, she returned to her party's flock, where she remained through 2004. In George W. Bush’s second term, Snowe moved back to the center for two years and then actually sided more with Democrats in 2007 and 2008. Under President Obama, she reclaimed her place in the middle.
"The goal here is to present a simple data-based puzzle for the scientific community to ponder," said the authors.
Moody is a professor of sociology and leads the Duke Network Analysis Center. Mucha is the Bowman and Gordon Gray Distinguished Term Professor in Mathematics, and also the chair of the Department of Applied Physical Sciences.
The journal Network Science is a new publication for academic research using networks to study natural, social, engineering and informational sciences.
CITATION -- "Portrait of Political Party Polarization," James Moody, Peter J. Mucha. Network Science, Vol. 1, Issue 1, published online April 15, 2013. DOI: http://dx.doi.org/10.1017/nws.2012.3.