Always Check the Dataset
By A while ago I used a story about the relative score of North Carolina on “Electoral Integrity” to frame a piece about democratic backsliding. I spend a lot of time looking at datasets, so I have no excuse for not looking at the one used in that piece. It turns out to be… problematic. And by ‘problematic’, I mean ‘not a basis for relative comparison.’ Indeed, it’s based on surveys of ‘domestic’ and ‘international’ experts—which, I suspect, means the scores generated for each country are often not comparable across states and regions.
That doesn’t mean that North Carolina has not been seeing significant democratic backsliding, especially since Shelby County. But let this be another lesson in truthiness. The fact that North Carolina scores lower than many non-democratic countries is probably an artifact of the way that the data was compiled.
UPDATE: First, I’ve made a few edits. Second, academics on social media are reporting that one of the researchers and his family are receiving death threats. That’s… insane and frightening.
