A lot of the differences in the number of reported coronavirus cases from state to state are the result of testing capacity rather than necessarily reflecting the actual spread of the disease. Thread.
For instance per @COVID19Tracking, let's take two groups of states.
1) High tests-per-capita states; these states 1) reliably report negative test results and ii) have conducted at least 0.4 test per 1,000 residents. They are WA, ME, NM, VT, MT, SD, AK, RI, MN, NH and ND.
1) High tests-per-capita states; these states 1) reliably report negative test results and ii) have conducted at least 0.4 test per 1,000 residents. They are WA, ME, NM, VT, MT, SD, AK, RI, MN, NH and ND.
2) Low tests-per-capita states. These states reliably report negative results but have conducted less than 0.2 tests per 1,000 residents. They are AZ, MO, ID, FL, WV, OK, PA, AR, KY, SC, NE, GA, VA, KS, IL.
Which states have more spread? Well, the high-testing states report an average of 0.03 cases per 1,000 residents. The low-testing states, only 0.01 cases per 1,000. So the high-testing states have 3x more spread, right?
Well, maybe not. *Of the people who get tested*, 9% of cases come up positive in the low-testing states, on average. By comparison, only 3% of cases in high-testing states come up positive. So that makes it seem like the *low-testing* states have 3x more spread, right?
What's the answer? Which states are safer / have less community spread? The answer is—it's hard to say. You'd have to do a *LOT* more work on this problem. BUT THE NUMBER OF POSITIVE TESTS IS OFTEN NOT A GOOD INDICATOR BECAUSE TESTING VARIES SO WIDELY FROM STATE TO STATE.
There are *some* things you can say. WA has done more tests per capita than anyone and still has a fairly high positive rate (7%). And NYC is quite scary, though NY is not reliably reporting negatives. Still, reported differences in + cases may mostly reflect testing capacity.
