The State Similarity Index attempts to quantify how similar American states are to each other relative to other states. The index is a statistically-based way to measure this. It weighs equally five major aspects of states: their demographics, culture, politics, infrastructure, and geography. The methodology is exactly the same for each state.
The data from the State Similarity Index was used to cluster American states into different regions. Some unique states, like Alaska and Hawaii, are very difficult to group with other states. Therefore, it was necessary to make the largest regions still have a great deal of variation within them. This resulted in 6 distinct macro-regions:
Please see this article, which compares these groupings to other regional maps created by geographers. These macro-regions are further broken up into 12 sub-regions, since there are still some significant differences within them. States as different as California and Utah are still grouped together. Three changes were made from the initial hierarchical clustering. The data showed that two states have attributes of two different regions. Therefore, Pennsylvania is grouped with both the Northeast and the Midwest, while Missouri is grouped with both the Southeast and the Midwest. Furthermore, although Oklahoma clustered with the states in the Southeast region, final analysis of the data showed it has more in common with Texas than the average state in the Southeast.
The states in each region and their individual pages are as follows:
— Texas, Oklahoma
— Alabama, Arkansas, Kentucky, Louisiana, South Missouri, Mississippi, Tennessee
— Georgia, North Carolina, South Carolina, Virginia
— West Virginia
It is hoped that this study has created a more rigorous and logical regional map of the United States than any previously created. More information on the similarities of each region to follow in later articles.
Here is the hierarchical clustering dendrogram, the primary source of the map’s regions:
A larger, high resolution version of the United States regions map:
Finally it may best to look at states as a connected web, not hierarchical clusters.
Here is a version of the map as a connected web: