COVID-19 may have been impartial in that it has affected just about every facet of our lives but its decade-long impact is likely to manifest itself in a larger Census undercount in some areas than in others. In this post I wanted to take a quick look at how our fair state is faring (no pun intended). Let us start with a view from the eagles’ crag – looking at our Census Tracts. We start by downloading the 2010 and 2010 response rate data, respectively.
Let us build a simple map that allows us to see if Southeast Ohio looks any different than the rest of the state.
Note the lower response rates for the Appalachian counties. This story is even more conspicuous if you focus on the Census 2020 internet response rates (shown below), although this is not a surprise given our patchy internet connectivity.
What about our cities? We could certainly map those data but I would rather use a table and perhaps something other than a map to see the 2010 versus 2020 difference. Time to pull the city data and dig in.
The first look is at the final response rate in 2010 (x-axis) versus the current overall rate (y-axis). The percent of the population living below the poverty line is layered in, with brighter colors highlighting places with a larger percentage of their population living in poverty as per the 2015-2019 American Community Survey. In general and as expected, poorer communities tend to have lower response rates.
The second plot measures the gap between the final 2010 response rate and the current overall response rate (y-axis) against the current internet response rate (x-axis). Again, the percent of the population living below the poverty line is layered in, with brighter colors highlighting places with a larger percentage of their population living in poverty as per the 2014-2018 American Community Survey. As expected, internet response rates are lower in poorer communities. What is more striking is the tendency for these poorer, low internet response communities to be farther below their 2010 response rate today. The tight clustering of low poverty places with very high internet response rates and very small 2010 - 2020 response rate gaps only underscores the punch line.
Another way of looking at the same pattern might be to swap the poverty measure with the current internet response rate; see below.
Finally, here are the data for individual places, arranged alphabetically by place name.