Interactive GRAPH: Vaccine Hesitancy for COVID-19: State, County, and Local Estimates

[progree]

11. Probably not. This Minn Public Radio Covid commentator discussed that ...

http://view.connect.mpr.org/?qs=acb5804de0956fa3dc4e555bf1bab66a2f29c8529e91df0604af6fd7f6747108fe7723df1c356f27f09b0c8a0439e3577f0d426ac24f7750ee1fdd5e61eb5db7d819630251024871da321a7a9c11d684

However, the dataset gets more granular than that. Government statisticians used the demographic details of people responding to the vaccine hesitancy survey to build a model extrapolating it to other Census data — and thus letting them get estimates down to the local level.

These should be taken with a pretty big grain of salt, given the layers of abstraction between these models and the actual views of the population. (You've got two different surveys, a model to connect the two, and then further statistical tricks to get county-level approximations.) One warning sign: the map of county-level hesitancy estimates shows pretty big differences on state lines. The idea that Minnesota as a whole might have different levels of vaccine hesitancy than North Dakota as a whole is plausible. But eastern North Dakota having hesitancy rates two to three times higher than similarly rural counties just over the Red River? That suggests something's up with the data.

...

(Though this data is by county, you might notice lots of big blocks that are the same color. That's because this data really exists at the level of the Public Use Microdata Area, or PUMA, an under-used piece of Census geography with no fewer than 100,000 people; in rural areas PUMAs are made up of blocks of counties, while in urban areas they follow city or neighborhood boundaries. The county-level data is extrapolated from the PUMAs.)

blah blah blah snore