Who Slowed Down the Vaccination Rollout?

Background

As vaccinations slow down many have begun to point fingers at certain demographics for the cause of this. Some have stated rural Republicans are responsible for the slowdown while others have blamed inequity in the health systems for the slowdown. We take a look.

The Data

The CDC has been reporting on the demographics of the vaccinations and made the data publicly available. The Census Bureau of the Unites States also reports on the county level demographics of each county allowing us to dig into variances and correlations between these two data sets. Finally, we know which counties voted for which president in the 2020 election allowing us to determine the political leanings of each county. Together, with this data we were able to identify the variance caused by each of these categories and the factors (values) within each; Race, Age, Sex and Political Party.

We should note that ‘Hispanic’ is generally not considered race however the CDC reporting reports ‘Hispanic’ as a race, similar to White, Black or Asian.

Methodology

We sat down and reviewed the variance of each category had on the overall vaccination rates across the country for each of the 144 permutations between all factors of each of the four categories. We then went a step further and identified the demographics that could be targeted for two goals; herd immunity and equity. For herd immunity we focused on targeting the largest unvaccinated demographics and the counties they were most concentrated in. For equity we focused on the largest counties that contained the demographics with the lowest vaccination rates.

The Results

We found that race was the largest driver of the four, accounting for 40% of the variance in vaccinations. Party and Age were the next drivers at 24% and 22% respectively with Sex accounting for 15% of the variations in vaccinations.

Below you can find the vaccination rates for each category and factor. Based on the CDC’s reporting we found that the lowest vaccinations rates resided with Race (African American – 21%) and Race (Hispanic – 28%). The highest vaccinations rates were, expectedly, in the higher age groups, who were prioritized early during the vaccination rollout.

Taking our analysis a step further we sought to identify ways the rollout could be improved. First, we took the goal of vaccination equity to mind. We would target the demographic factor intersection that had the absolute lowest vaccination rates. This would be the African American Male 18-24 Republican demographic. We chose the top 10 countries that had the largest population of this demographic to be targeted: Mobile, Alabama; Suffolk, New York; Oklahoma, Oklahoma; Polk, Florida; Bell, Texas; Macomb, Michigan; Madison, Alabama; Jefferson Parish, Louisiana; Collin, Texas; Denton, Texas. These are just 10 counties and in total account for just 30,000 members of this demographic. The rollout would need to be much more widespread to have a meaningful impact, likely to the next 300 counties with the largest concentration of this demographic.

Second, we identified the demographic and locations that could be targeted if the goal were simple herd immunity. This would focus on the counties with the largest number of those that are unvaccinated. The single largest group of unvaccinated individuals are White Male 25-39 Democrats with others following close behind:

The two above findings may seem at odds with each other at first glance. However, after looking more deeply we found that the large relative proportion of the population that was Male 25-39 and Democratic was very large. When combined with a slightly below average vaccination rate (38%) it produced the largest unvaccinated percentage of the total national population. For this same reason, we do not find populations with low vaccination rates in this list; the overall proportion of the population is relatively small. For instance, African American Male 18-24 Republicans account for just 0.11% of the population, whereas was Male 25-39 Democrats account for 3.3% of the population, or 30 times more of the population.

The illustration below helps to illustrate this with a scatterplot.