Subject: | |
From: | |
Reply To: | |
Date: | Wed, 10 Jun 2020 12:59:16 +0000 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
" but with a sample size of 1037, we can have a
99% confidence level and a 4% margin of error"
The incidence of covid 19 in the US as a whole is about 3%. It varies greatly with age, region and economic status. But, lets just ignore all those variables for a minute. With a test population of 1037 and assume bee keepers are identical to the whole population for all variables besides susceptibility to covid 19. The 95% confidence levels would say you should see between 23 and 39 cases of covid 19 in your 1037 bee keepers. Anything outside that range simply says bee keeping has a statistical effect at the 95% confidence level. Medical stuff is generally reported as positive when you reach the 95% confidence level. If you want to push to the 99% confidence level the range would be 20 to 49. You need a number of cases outside that range to be 99% sure bee keeping has an effect on susceptibility to covid 19.
My 3% includes the 30% of infected people who never show symptoms so do not know they have had the disease. And, that 30% is not a well known number at all. We know for sure it is not zero. I picked 30% simply because that is what the Ohio prison data shows. That 30% may or may not be valid for the general population. After all, the high density crowding in a prison is far different from going to the grocery store in terms of concentration of active virus suspended in the air. Besides, prisoners in general are in better health than the general public, at least in the Ohio system. So, what we really need is not reported cases based on symptoms but cases based on antibody tests on each bee keeper. You have not controlled at all for risk factors. Not even blood type! No legitimate medical journal is going to publish data on a test population of only 1037 subjects for covid 19 unless the results are truly dramatic. Dramatic like 75 cases or 3 cases. And, without a lot of medical history data such a result would still be considered ancedotal.
I may remind you it took a meta study on over 100,000 to show that statin treatments of people with "high" cholesterol but no diagnosed heart disease (roughly defined as no angina upon exercise, no history of a heart attack, no stents, no coronary by pass procedures) to reach the 95% confidence level that such treatment had no statistical effect on heart attacks after the statin treatment was started. The rate of heart attacks in this group was around 3% during the studies. Not very different than the incidence of covid 19.
Dick
***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software. For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html
|
|
|