Member-only story
On The Sharing of COVID-19 Statistics
I’ve worked in statistics for 25 years. It’s my avocation and my occupation. With the advent of COVID-19 and daily CDC updates, my sistren and brethren of the craft seem to be welcoming a burgeoning number to our ranks. People on both sides of the re-open/don’t reopen debate (yes, both sides) are using statistics like a drunkard uses a lamppost: for support, not illumination.
When it comes to the primary charts and tables released daily by the CDC, I have scrutinized every base, every legend, and every label. And I can tell you that exactly *one* of these things is true:
1. The statistics are confusing.
2. The statistics are conflicting.
3. You don’t understand statistics.
Watching people cherry-pick a stat to support the belief they already held in the absence of data is like watching someone grab a platypus in the dark and argue that it’s a duck. It’s not a duck. In the comments of nearly every Facebook post I’ve seen about ending the lockdown, somebody makes this comment: “Well, statistics can be made to show anything.” My dog-eared copy of Strunk & White tells me to Eschew Passive Language, so let’s just put the subject of that sentence back in, shall we? “YOU make statistics show anything YOU want to show.” Statistics are not misleading. Statistics simply are. People…
