So 9 days ago, I tweeted what I thought was a joke about states conflating antibody testing with their viral testing data.

It was so obviously bad to do this that, at first, I didn't believe it was happening. https://twitter.com/alexismadrigal/status/1260039815824736261?s=20
But actually, it turned out it was happening in many different states. Virginia, Maine, and Vermont have already repented. There's a big fight about it in Texas. And we report tonight that Pennsylvania is doing it, too.

The worst disclosure came today, though.
Worse, when the @CDCgov launched the new tracker, they explicitly said that they were only including viral tests.

On the left, there's the info they had up through May 18. And on the right, you can see the current version of the page.
Why is this failure so bad? No outside public health expert that we've talked to thinks it is a good idea to mix these tests.

Here's what Kathryn Turner, the deputy state epidemiologist for Idaho, told me about why *her* state isn't including antibody tests.
So, now the question becomes:

if scientists and data people *outside* the CDC immediately recoil at the thought of mixing viral and antibody tests, then ... why did the CDC do this?

For me, it is very hard to believe that the lifelong civil servants *wanted* to do this.
The most generous possible interpretation is that the CDC maybe didn't know that the states were mixing their test results. That's what Vermont claimed. https://twitter.com/EPetenko/status/1263138001879797762?s=20
But if you work for the CDC and you know the actual answer, like I said before: [email protected].
And also, some local reporters/pubs deserve huge shoutouts:

@MelLeonor_ blew open this story with her reporting in Virginia.
@texasobserver first reported Texas was doing it.
@asuozzo got Vermont.
And @toomuchme got the CDC on the record this morning.
So, one lesson you could draw from all this:

Subscribe to your local news(paper).

@COVID19Tracking would not work without so many local reporters pushing so hard for transparency in their cities and states.
Another lesson:

As they say in science and technology studies, Raw Data Is An Oxymoron. (Cf. @namleti)

That doesn't mean all data is bullshit. It just means the data production process needs to be rigorous, ethical, and transparent. https://mitpress.mit.edu/books/raw-data-oxymoron
Finally, these antibody tests probably account for *some* of the discrepancies that we were at a loss to explain in our @COVID19Tracking white paper comparing the state data with the new CDC data:

https://covidtracking.com/blog/tracking-cdc

So, I guess v2 of this paper has begun.
If you’ve been following this antibody test reporting saga, here’s a new tidbit.

Missouri’s health department says that the CDC directed them to report in this way. 😐😑🤨

https://www.kansascity.com/news/state/missouri/article242957701.html
You can follow @alexismadrigal.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: