I began to think about various ways to cover his win and I became convinced that the color-coded map of who won which county would reveal some pertinent information about the nature of the primary.
It became apparent that in order to keep this story in the relevant news cycle, I would have to abandon my hopes of a comprehensive statistical analysis of various factors. The following is a very basic look at some key findings as well as speculative conclusions drawn from a visual analysis of data, maps, and tables. A deeper analysis, as it concerns statistical significance, may be forthcoming.
Correlation with votes for Trump in November 2016.
Both Pensler and James claimed to be the most “Trump-ian.” In the counties that they won, both seemingly enjoyed support from the same type of voters. The rate of Trump votes in a given county had a strong positive correlation with the rate of support for James and Pensler. This is expected. It’s a republican primary.
Were Cruz voters more predictive?
A visual inspection of the county maps provided by the New York Times for both yesterday’s primary as well as the 2016 presidential primary reveal that, with the exception of the thumb and a few of the UP counties, counties that James won were more supportive of Cruz than counties that Pensler won. Trump won most of Michigan, so a lot of that is based on how much Cruz lost by. For the areas that Cruz won outright, we do see the same phenomenon, support for James over Pensler.
When Pensler won a county, he won it with less votes.
97% of Pensler’s county wins were decided by less than 1000 votes. 71% of his wins were decided by less than 500 votes. That is compared to 52% at 1000 and 39% at 500 for James. On average, Pensler won his counties by 55% of the vote. James on the other hand won his counties with 58% of the vote. In some instances, James enjoyed up to 68% of republican support in a given county, while Pensler garnered at most 60%.
Participation is low, but not too bad for a midterm primary.
According to FairVote.org, about 40 percent of eligible voters vote in non-presidential election years. This number tends to be lower in primaries. The data I compiled compared participation rates with Trump’s 2016 win since we are dealing with a republican primary, not a general election. Participation, as it compared to votes for Trump, ranged from 13% to 54% with an average of 41% participation. This number translates to about 25% of the eligible voting population.
What factors might have influenced the election?
I looked into a variety of measurements, demographics, and other stats to see if something could explain why the state was divided into clean cut sections of support for either James or Pensler.
Here are some possible considerations: race, age, income, and education.
-Using data from the Census Bureau compiled by Randy Olson, The Atlantic ran an article on racial diversity by county back in 2014. These maps included information regarding diversity in general as well as specifics by race. Again, my analysis is limited to a visual inspection of the data until I have time to enter demographic information of 83 counties into Excel. John James won areas that are heavily Black, but he also lost areas that are heavily Black as well as other areas that are labeled “diverse.” James won areas that are heavily White or not diverse. Pensler won areas that are heavily Black or diverse and lost some areas that are heavily White. I suppose in order to analyze this information correctly, one would also need to have the rate of participation of minorities in primary elections. Alas, it's more information I don't have at the moment.
-This was all over the place. Nothing to report.
Income and Education
-I’m combining these too because they actually work well together and the information comes from the same source. An article by Julie Mack on MLive.com from earlier this year displayed an interactive map that listed this demographic information. I strongly recommend checking this article out. The maps are very interesting. It appears that in counties where education and income are higher, John James did better. There are obvious exceptions to this. The Grand Traverse and Petoskey areas have high levels for both income and education and voted for Pensler. There are other aberrations in the opposite direction, but the general appearance of these two maps reflects the result of the primary.
Does this mean anything? Probably not. I stated at the beginning that it’s just my own reckless speculation. I hope to explore these issues in depth with a more scientific approach. Until then, enjoy the links to the various maps.