This reprint Worst statistical analysis I’ve seen this year Prompted by Mr. Bruce Hall’s tendency to give any ol’ link as support for a given post (actually, I think this is the worst I’ve seen).
(and I’ve seen a lot of horrible analytics) [Update 8/14/2020: the author has taken down the post, but here is an archived 8/13/2020 version of the webpage]
reader Bruce Hall recommend Link to this article It asserted that the 2020 death toll was not abnormal. In fact, it’s the 20th time in the past 21 years!
Now, Ms Colleen Huber of NMD*** concludes:
As of this writing, 2020 has passed 32 weeks. However, for the previous year, 52 weeks have passed. So how do we compare all-cause deaths in 2020 to previous years?
I divide the total number of deaths per year by the number of weeks. That’s 52 weeks for all years except 2020, which has 32 weeks as of this Saturday, August 8, 2020, the most recent week in the quoted CDC data. This gives us the average number of deaths per week for each of these years and allows meaningful comparisons between 2020 and previous years.
Then she generates the following table:
She concluded:
It doesn’t look like 2020 will see a pandemic of COVID-19 or any other disease, at least not in the United States.
Glad Ms. Huber told us there are 52 weeks in a year. She divides the 2020 data by the 32 weeks that the CDC has recorded (although reporting has been very patchy in recent weeks).
It would be a sensible way – to calculate the weekly death rate – If the data is not seasonal. However, death yes It’s seasonal in the US, which can be easily seen from the CDC data she’s analyzing.
figure 1: CDC data accessed August 7, 2020.
As we enter the second half of the year, the death toll usually rises (flu, etc.). Therefore, using 32 weeks in 2020 and all 52 weeks in previous years often produces meaningless comparisons. (There is a standard method, used in many economics publications – year-to-date counts. That is, Ms. Huber can compare the number of deaths in the first 32 weeks of each year for the past 20 years with the number of deaths in the first 32 weeks of the past 20 years Compare. This year.)
Another most embarrassingly stupid data analysis I’ve seen this year (maybe this decade, although the competition is fierce).
My survey uses the CDC’s estimate of the expected death toll, here.
*** “NMD” means “Natural Medicine Doctor”
Addendum, 2/19/2023:
Ms. Huber has recent postbut since I have to register The Epoch Times To read it, I’m just going to provide the link and let others risk their information being hijacked.
