Image courtesy of Adobe Stock.
We have been locked down long enough for the volume of statistics to build up to the point of allowing deeper analysis. As is always the case, the statistics can also be manipulated to give whatever answer you want. Deeper, honest analysis is starting to show surprising results, for example, the lack of correlation between infection rates and lockdown policies.
Surprising results on the low correlation will be mentioned after some game playing is described.
Wall Street Journal – 9/9/20 – The Sturgis Statistical Misfire – For this story remember the old saying
- Figures don’t lie, but liars figure
The annual motorcycle rally in Sturgis, South Dakota routinely draws huge numbers of people to the town, whose population is only 7,000 people. Attendance at the 10-day event this year was lower than usual with an estimated 460,000 motorcycle enthusiasts hanging around.
Frightening news reports at the time said this would cause massive numbers of Covid infections leading to massive numbers of deaths.
Unless you turned off the news lately you have surely read about a report from the San Diego State University’s Center for Health Economics & Policy Studies which validated fears the Sturgis rally would cause massive infection.
The report calculated the infection rate in South Dakota increased jump from 3.6 all the way up to 3.9 per 100,000 people which means they were somewhere between 3,185 and 3,441 new cases statewide. That is a significant portion of the total cases in the state.
Let’s put that in perspective.
The increase inferred by the researchers took place over three weeks. It also reflects increases during that time across the whole state.
More perspective – that is equal to about one day’s worth of new infections in California this week. For the week of 8/16/20 new infections in the state of California were three times that tally each day of the week. That tally of infections over three weeks in South Dakota is equal to about four days’ worth of new infections currently in the County of Los Angeles.
While the entire increasing caseloads in South Dakota may or may not be caused by one event in Sturgis, it is rather unlikely because after the event the attendees scattered across country.
So seems to me the assumption for South Dakota is uncertain to start with.
The researchers then went off the rails quite spectacularly.
The reports you’ve heard, because they were probably covered in every media source that exists, say the number of new infections across the entire country is 266,796. Rounding, let’s call that a quarter of 1 million.
Pull a number out of the hat for the health cost of each and every one of those cases being $46,000 (the fallacies in that assumption are delightful, but we won’t pursue that false argument) means the total cost was $12.2 billion.
Consequences of the Sturgis rally include 267K new infections at a cost of $12.2B.
That is, until you look at the details…
The researchers tracked cell phone data of the attendees on an anonymous basis back to the counties where they live. Researchers then looked at the infection rates in those counties and pulled an assumption out of thin air that the increase infections (in only those counties which saw increases) was due to people who attended the rally.
Extend out that massive string of incorrect assumptions and you get to the 267K new infection count.
According to the reporting in the Wall Street Journal the researchers did not adjust for any of the long list of factors which should have been taken into consideration. The biggest issue is many counties in the study were already seeing accelerating infection rates before the rally started. Amongst the powers of the Covid virus is to make people in a county sick before the motorcycle fans climbed on their ‘hogs and headed towards South Dakota. I guess the ‘rona is so poweful it can travel back in time.
Such as the math behind many published reports on the coronavirus.
On to some research that passes the smell test…
Donald Luskin – Wall Street Journal – 9/1/20: The Failed Experiment of Covid Lockdowns – At the opposite end of the spectrum from starting a project knowing where you are going is research which produces counterintuitive results which cannot yet be explained.
Author of the article, who is the chief investment officer of an analytics firm, pulled the state-by-state infection rate data from public sources and analyzed each state from the beginning of the pandemic to the point of maximum lockdown in the state, which was typically sometime in mid-April.
On state-by-state basis they correlated infection rate increases with the strength of the lockdown.
They then took infection rate data from the peak of each state lockdown to the end of July. They correlated the timing of when restrictions started to be released with the subsequent infection rate.
What they found is completely contrary to what everyone would have expected and is totally contrary to what we’ve been told every day about happened. Neither these researchers nor anyone else has been able to explain why they are seeing the results they are seeing.
They did not find that strong lockdown requirements reduced infection rates. On the contrary, they found a very weak statistical correlation between lower infection rates and less harsh lockdown requirements. In other words, a strong lockdown does not seem to have had a favorable impact on reducing infection rate. Instead a milder lockdown seems to have produced a reduction in infection rate.
In terms of loosening up restrictions they did not find a correlation between when restrictions started to be reduced and subsequent infection rate. In other words, the timing when each state started to open the economy does not seem to have had an impact on the infection rate.
Author of the article says other studies are starting to find the same relationships, specifically an inverse relationship between the harshness of lockdowns and infection rate and also minimal correlation of when the restrictions started to be relaxed and subsequent infection rate.
Author says this is completely counterintuitive yet the data which is beginning to emerge supports that assessment. Author says no one is currently able to explain why those relationships exist.
Far more research is necessary. The preliminary indications are, in the author’s description:
“But now evidence proves that lockdowns were an expensive treatment with serious side effects and no benefit to society.”
So the author’s tentative conclusion indicates we incurred serious economic harm, a year of lost education for tens of millions of children, declining health, increasing mortality, and still-unquantified psychological damage without any benefit.
But the lockdown continues. And the cost rises.
0 Comments