Statistically, Santorum is the anti-Paul

An interesting way to look at the results of the Iowa Caucuses is to view each county as an individual data point. This gives us a convenient set of 99 data points and yields some insight into the Iowa voter, if you believe that there is a correlation between a voter’s physical location and his or her political propensities.

Cluster Analysis of 2012 Iowa GOP Primary Results

Each dot represents one county in Iowa.

k-means Cluster Analysis was used (in R) to group each county into one of three clusters. For you non-quants out there, this simply means that we asked the computer to group each Iowa county into one of three groups — based on how similar that county’s voting percentage results (for all 6 top candidates) were to other counties.

The green cluster appears to be counties of strength for Ron Paul. The red cluster appears to be counties of strength for Rick Santorum, and while Mitt Romney did well in the black cluster — it’s not quite as distinct of a set as the green (Paul) and red (Santorum) clusters.

The middle row, rightmost column’s chart clearly shows the interplay between Paul and Santorum. In counties where Santorum did well (towards the top of the chart), Paul typically fared poorly. In counties where Paul did well (towards the right of the chart), Santorum typically fared poorly. Such results lend evidence to the hypothesis that supporters of Paul are not likely to be supporters of Santorum, and vice-versa.

Mitt Romney’s results present an interesting case. On the one hand, Romney (like Paul) does worse-than-average in the red (Santorum) cluster. However, Romney performs reasonably well in the black and green clusters. Such a distribution could speak to the theory that the Romney areas and the Paul areas have more in common, or that Romney has a broader appeal than either Paul or Santorum. Another theory holds that Romney is a “default” candidate for those who lack the enthusiasm about Santorum or Paul. Either way, all of the top three candidates will soon have a test in New Hampshire that will help shed more light on their long-term prognosis. Plus, we have the wildcards of a potentially re-energized Gingrich, Bachmann’s graceful exit, and the perhaps under-rated Jon Huntsman.

Global Decision will leave it to the various SuperPACs to sling the mud.

This entry was posted in analytical methods, campaign analytics, politics and tagged , , , , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published.