Monica Cuddy points us to a news article by Margot Sanger-Katz, “Your Surgeon Is Probably a Republican, Your Psychiatrist Probably a Democrat,” which reports a data analysis by political scientist Eitan Hersh and psychiatrist Matthew Goldenberg performed by mashing up “two large public data sets, one listing every doctor in the United States and another containing the party registration of every voter in 29 states.”
The above graphs shows the partisan leanings for several medical specialties, and they also supply this scatterplot of partisanship by income:
This set of graphs from Red State Blue State is also relevant. But what really struck me about the that graph of political leanings vs. salary is that lower bound of the salary scale is $200,000. That’s a high-paying field where the average salary is so high for even the lowest-paying specialties!
P.S. In a welcome and unusual step for a news article, Sanger-Katz summarizes the methods in some detail:
Hersh and Goldenberg constructed the data set by assembling a large sample of doctors from the federal government’s National Provider Index, a file of every physician in the United States who either bills insurance or shares data digitally. There are very few doctors who are not included in this file.
They then matched each physician to data from state voter files maintained by Catalist LLC, a political data vendor. The researchers searched for doctors with matching names, living within a small geographic radius from their practice address. Not every doctor matched. Some had moved; some were not registered to vote; some had changed their names; some had common names that made it hard to make a definitive match; some lived nearby in states where the voter file does not include political information; and there may have been some mistakes in each file. But over all, the researchers were able to collect complete data for more than 55,000 physicians living in the 29 states where voter files include party registration. (Those states contain about 60 percent of the population, and are roughly, but not perfectly, representative of the country.)
For many of the measures in this article, we looked only at the percentage of “partisan” doctors, that is, doctors who recorded a political party. There was a substantial fraction of the physicians with no political affiliation, and there was a small fraction who were registered with smaller political parties. Altogether, the analysis looks at just over 36,000 doctors.