Guest “Excel-lent!” by David Middleton
What happens if you crossplot the “lockdown” rating of the Lower 48 states and DC with the COVID-19 infection rate?
To the extent there is a correlation, the states with the tightest lockdowns have the highest infection rates. Alaska and Hawaii were the only states ranking in the top 10 most aggressive lockdowns that didn’t have high infection rates. They are also isolated relative to the contiguous United States.
But, but, but… Correlation is not causation! States could have locked down more tightly in response to the infections! The lockdown rating was as of April 6, 2020 and the infection data are as of May 11, 2020.
Data Sources
Worldometer. WORLD / COUNTRIES / UNITED STATES. Last updated: May 11, 2020
| State | Lockdown Rating | Total Cases/1M | Deaths/1M | Deaths % of pop. |
| New York | 1 | 17,755 | 1,378 | 0.14% |
| District of Columbia | 2 | 8,887 | 458 | 0.05% |
| Alaska | 3 | 518 | 14 | 0.00% |
| Hawaii | 4 | 446 | 12 | 0.00% |
| New Jersey | 5 | 15,763 | 1,043 | 0.10% |
| Rhode Island | 6 | 10,642 | 398 | 0.04% |
| Washington | 7 | 2,313 | 122 | 0.01% |
| Massachusetts | 8 | 11,287 | 722 | 0.07% |
| New Hampshire | 9 | 2,259 | 98 | 0.01% |
| West Virginia | 10 | 760 | 30 | 0.00% |
| Minnesota | 11 | 1,999 | 102 | 0.01% |
| Vermont | 12 | 1,486 | 85 | 0.01% |
| Maryland | 13 | 5,390 | 272 | 0.03% |
| Connecticut | 14 | 9,411 | 832 | 0.08% |
| Delaware | 15 | 6,621 | 230 | 0.02% |
| Louisiana | 16 | 6,797 | 492 | 0.05% |
| Maine | 17 | 1,068 | 48 | 0.00% |
| California | 18 | 1,719 | 69 | 0.01% |
| Pennsylvania | 19 | 4,691 | 299 | 0.03% |
| Ohio | 20 | 2,060 | 115 | 0.01% |
| Indiana | 21 | 3,584 | 224 | 0.02% |
| Montana | 22 | 429 | 15 | 0.00% |
| Illinois | 23 | 6,135 | 269 | 0.03% |
| Idaho | 24 | 1,248 | 37 | 0.00% |
| Oregon | 25 | 765 | 30 | 0.00% |
| Wisconsin | 26 | 1,755 | 69 | 0.01% |
| Tennessee | 27 | 2,194 | 36 | 0.00% |
| South Carolina | 28 | 1,486 | 64 | 0.01% |
| Georgia | 29 | 3,187 | 132 | 0.01% |
| Kansas | 30 | 2,387 | 60 | 0.01% |
| Colorado | 31 | 3,421 | 169 | 0.02% |
| Missouri | 32 | 1,631 | 81 | 0.01% |
| New Mexico | 33 | 2,319 | 95 | 0.01% |
| Kentucky | 34 | 1,441 | 68 | 0.01% |
| Virginia | 35 | 2,937 | 100 | 0.01% |
| Iowa | 36 | 3,790 | 84 | 0.01% |
| North Carolina | 37 | 1,424 | 54 | 0.01% |
| North Dakota | 38 | 1,957 | 46 | 0.00% |
| Arizona | 39 | 1,528 | 74 | 0.01% |
| Michigan | 40 | 4,720 | 456 | 0.05% |
| Nevada | 41 | 1,980 | 99 | 0.01% |
| Texas | 42 | 1,376 | 39 | 0.00% |
| Utah | 43 | 1,950 | 21 | 0.00% |
| Florida | 44 | 1,890 | 80 | 0.01% |
| Mississippi | 45 | 3,192 | 144 | 0.01% |
| Arkansas | 46 | 1,329 | 30 | 0.00% |
| Wyoming | 47 | 1,144 | 12 | 0.00% |
| Alabama | 48 | 2,020 | 80 | 0.01% |
| Nebraska | 49 | 4,298 | 51 | 0.01% |
| South Dakota | 50 | 3,835 | 38 | 0.00% |
| Oklahoma | 51 | 1,160 | 69 | 0.01% |
Addendum 5/12/2020 11:04 AM CST
The population density of the 50 US states correlates fairly well with the infection rate.
Lockdowns are like mosquito nets. There is no doubt that they are effective. But you mustn’t have even just ONE tiny hole in them or all the effort is in vain.
But even if you successfully establish that there are no holes in it before you go to bed, The sad reality is that when dawn breaks you are faced with a choice. Either you stay protected and meander about like a ghostly apparition or you bravely discard the protection in full knowledge that somewhere, someday, somehow those little critters are going to get you!
Tricky dilemma!
David
After trying to explain these two graphs to a third party, I’ve come to a different conclusion to what they tell us.
The strongest lockdowns were in more populated areas, so these graphs don’t really tell us much about how success the lockdowns were. The second graph demonstrates quite clearly that population density causes more cases per million. Which makes sense.
I can’t see any evidence one way or another that the lockdown causes the higher cases per million. This aspect seems to be simply the higher population density, and is also the location where the stronger lockdowns happened.
As I see it,to appreciate the effect of lockdown,we need a no lockdown state to compare to.
Now I have been told Brazil is a good candidate for no government reaction and limited medical response.
However ,where do we find good numbers in any country at the moment?
Even the USA has biased the reporting of infection by the very perverse rewards of declaring an emergency,suddenly all deaths become Covid 19.
How long do we need to wait to see real numbers?
How long should it take for this virus to peak in Brazil(for example) and which country has had the most effective lockdown process?
For Brazil,if what is said about their action is true,is the best test case for worse case models..
Is South Korea or Taiwan the best lockdown states?
Forget comparing the craziness of the US approach with the dithering of the UK, the failure of Belgium or the alleged cool Swedish approach.
Out in front is Vietnam….not a small country, population 96 million and a 1400km border with China.
288 [not thousands or tens of thousands], 288 cases to date and no deaths.
People [other people] might dislike their methods but they were bloody effective.
Best on the planet
https://www.orfonline.org/expert-speak/vietnam-emerges-victorious-in-fight-against-covid19-65666/
Or maybe the virus was never as bad as people thought and the numbers have been massaged upwards. Strange that 11, mainly 1st world, countries have been responsible for 92% of the deaths.
8% of deaths spread amongst the other 184 countries with Corona.
The severity has certainly been over hyped by an excitable media…anything for a good story. Now there are “fears” of a COVID-19 “second wave” if lockdowns are lifted.
I’m glad you provided the addendum; now you need to work a measure of (lockdown rating “x” population density) vs Covid-19 cases per million, where “x” is some useful function (not necessarily straight multiplication; I’m not a statistician, but suspect there may be some useful function available in statisticians’ toolboxes).
Also, for good measure, what are the criteria used to determine lockdown rating? (I’ve just tried to follow the link to wallethub, but got blocked from having too many requests from my IP address, despite never having visited that site before in my life). Is lockdown rating purely a measure of how tightly restrictions on activity are defined, or does it take into account a general change in activity level as a result of the virus?
To be honest, the first graph is utterly pointless unless you’re suggesting that those at the top end of the lockdown rating scale did not lockdown enough, and therefore allowed a lot of infections to happen.
It’s not hard to understand why the tightest lock downs cause the highest infection rates. There are still essential tasks that must be done in contravention of any lock down. Failure to carry out many of those tasks will raise infection rates or result in deaths by other causes. Carrying out those tasks will result in convictions for violating the lock down. The very act of enforcing a lock down will also increase infection rates for both the enforcers and the violators. It’s a self destructive catch 22 situation. It’s also not hard to understand why the wealthiest and most politically powerful people are the most infected. Anyone who commands them to self isolate, wash their hands or wear ppe while is subject to punishment for insubordination.
Civil disobedience has begun. Commissioners of Madison County, IL have declared their county “open for business.” This is just across the river from St. Louis. I may be visiting there soon…
Confused reasoning. The argument seems to be
1) We have a correlation between no lockdown and less epidemic
2) The lockdowns preceded the uptick in infections
3) Therefore the lockdowns caused, or had no effect on, the uptick in infections.
It doesn’t follow. Another plausible explanations is that the locked down areas always were more at risk, they locked down and thus minimized the epidemic. So the upticks occurred in spite of the lockdown and would have been much worse without them.
If you really wanted to show lockdown caused, or was compatible with, or had no effect on, the epidemic, you would have to find two comparable areas one that did lock down and one that did not, and show that the epidemic was similar in both, or worse in the locked down one.
How about Sweden and Denmark?
The evidence seems to be not, that lockdown has no effect, but that avoiding behavior starts before the official lockdown and is reinforced by it, not originated with it. And that it does have an effect on reducing infections, albeit at appalling cost, and inefficiently, because risk is not uniform across the population whereas the lockdown applies to everyone whether they are at risk or not.
The evidence suggests that a better plan than blanket lockdowns may be protection of the groups disproportionately at risk, the old and the infirm and the obese, rather than the blanket lockdown.
But that is a different argument from the argument that lockdowns don’t work. This argument is that they are not being implemented sensibly and are too costly in their current form for what they deliver.