Make sure you’ve read the previous post on this topic. Otherwise, you’re at risk of being swallowed in a puddle of confusion.

To recap, we looked at the implausibility of Arkansas’ testing numbers being correct in this graphic provided by the White House to Gov. Asa Hutchinson. Arkansas has, to date, tested 8.7 out of every 1,000 residents, according to data provided on testing numbers from the Arkansas Department of Health and population numbers from the U.S. Census. That varies dramatically from the 60-90 per 1,000 showed here for the state.

What I didn’t do yesterday was look at the numbers for the other states. I had previously created a ranking map for testing per 1,000 residents by state. If you peruse here, you’ll see that that Arkansas ranks 35th in testing numbers provided by the Covid Tracking Project.

It only takes a minute to see that all the states are off – by a lot. There doesn’t seem to be a pattern to it. My initial theory was that they just left a zero off the Arkansas population. But that’s not the case for all the states.

For example, let’s take Louisiana. This chart shows that they’ve done 50-60 tests per month per resident, fewer than Arkansas. If you look at the data being put forward by the State of Louisiana, they’ve done 30.6 tests per 1,000 residents. That’s much, much better than Arkansas’ testing rate. In fact, Louisiana is ranked third in the country for testing per capita. But yet, the White House chart shows them doing worse than Arkansas, which is ranked 35th per capita based on our own numbers.

How about Mississippi? The White House chart shows Mississippi at 40-50 tests per 1,000 people. State data has them at 17.2 tests per 1,000 people, ranked 13th in the nation. Yet, like Louisiana, Mississippi is shown as having done less tests than Arkansas.

Arkansas Blog reached out to the governor’s office for clarity on the graphic. The response, from the governor’s press aide Katie Beck below:

“The sources cited by the graphic is ‘Cepheid; Hologic; Abbott; Roche.’ The last two are commercial labs.  I presume the first two are commercial labs.  The graph was sent to me by the White House Task Force on Covid-19.

This makes sense and is probably the most authoritative source since we often only get positive test results from the commercial labs and all of them do not send us the negative results.  This means our internal numbers are lower and this is the first indication of the actual testing being done.”

The sources do matter. They are indeed all commercial labs – which means the White House graphic wouldn’t include any of the tests from the Arkansas public labs, and would bring the testing numbers down significantly.

I’ve reached out to the Arkansas Department of Health to confirm whether the state receives (or doesn’t receive) negative test results. If they don’t receive negative test results, why not? Have they requested it? That would seem to be a major factor in determining an accurate percent positive rate – one that’s cited every single day in the governor’s press conference.

I’ve also asked the Department of Health what consistitues a test for counting purposes. For example, the commercial labs could be reporting every time they’ve run a sample through a machine, even if it’s the same sample from the same person. Or maybe there are multiple samples from the same submission. The assumption is that each positive or negative result is based on “submission.” The thinking is a doctor submitted your slime for assessment, and whatever the lab returns (positive or negative) would count as 1. But that may not be the case. If the commercial labs are counting negatives multiple times, does that me we count positives multiple times? And are the public labs counting the same way the commercial labs are? It’s a slippery slope.

It seems like it would make more sense for the governor to say that his office is unsure about the background of the data presented, but that they will find out. Instead, what they said is “this is the first indication of the actual testing being done.” By saying that, they are implying that the data they have presented this far has been inaccurate and that the data underlying the chart – which neither they nor we have access to – is correct. The idea of the governor sacrificing the integrity of the state’s own data at the altar of a flawed White House chart is a bizarre one, but that may be where we are.