Gov. Asa Hutchinson proudly displayed this graphic at today’s coronavirus press briefing to illustrate that Arkansas is seemingly a leader in the rate of testing. I added the scrawled question marks after some commenters on Twitter detected a problem with the numbers.
Hutchinson said the graphic was provided by the “White House.” He said it shows “Arkansas is doing very well” in comparison to the rest of the country. There was some question in his presentation about whether Arkansas fell in the 60-90/1,000 or the over 90 per 1,000 categories.
The state’s own data raises a question about either figure.
Little Rock lawyer David Couch weighed in first or was at least the first question I saw on Twitter.
I just don’t see how this can be accurate.
Mississippi’s population is 2.9 million and they have tested 51,434 and Arkansas’ population is 3 million and we’ve tested 26,483 but Mississippi’s per capita testing rate is lower? https://t.co/KFyUz5UNm1
— David Couch (@DavidCouchAR) April 20, 2020
Then came a response to Couch from an account devoted to COVID-19 news.
It’s not accurate. It’s off by a factor of 10. It’s Arkansas’ testing numbers for April, but with a population of 300,000 instead of 3,000,000. Here’s the math: https://t.co/ZHi4TghmnZ
— ArkansasCovid (@ArkansasCovid) April 20, 2020
That website commented:
Gov. Asa Hutchinson shared a graphic given to him by the White House that purported to show that Arkansas was one of the leading states in the country on testing. The problem? The White House apparently thinks Arkansas has 301,780 residents instead of 3,017,804. What’s worse than that? Either the governor’s office does too, or it has no idea how many people the state is testing for Covid-19.
Then it ran the numbers:
Arkansas conducted 5,794 tests by March 30. Assuming all of those tests were done in March, that would be .00191/per person or .191 per 100 people or 1.91 per 1,000 people. No where near the 50 or 60 per 1,000 shown on the map.
So far in April, Arkansas has tested 16,913 people. That’s 0.00560 per capita, or .56 per 100 or 5.6 per 1,000. I suppose, if the White House thought Arkansas had 301,780 people instead of 3,017,804 people, that would put us between 50 and 60 for the month. That fits – if they left off a 0.
Ok, so what if they were looking at mid-March through Mid-April? Well, let’s be generous and go ahead and include every single test Arkansas has done to date. That’s .00823 per person, which is .823 per 100, or 8.2 per 1,000. Even if we included every single test Arkansas has ever done, we would be at the bottom of the barrel among states on this map.
I’ve sought comment from the governor’s press secretary, who customarily doesn’t answer questions from Arkansas Times; the governor himself (who has responded once in a while), and the Health Department.
Given that the numbers Hutchinson highlighted reportedly came from a notoriously untrustworthy source of factual information, it’s probably best to just forget you ever saw the thing.
UPDATE: I heard first from the governor who chastised me for being unprofessional in putting the question to him without first waiting for his aide to respond. He also said I’d been unfair in my characterization on Twitter of his answer to a reporter who asked about the impact of the crisis on poorest citizens. (He referred them to community agencies.) He did not respond to my specific question about the erroneous White House data.
Later, I did get a rare response to a question from Katie Beck, his press aide:
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.”
This is not an answer to the question