I suppose Aristotle or Bertrand Russell might say that the statement "We don't know the denominator" is true, when referring to the actual Covid mortality rate. But in the real world, that claim is a lie.
It may have been true in the first few days of the U.S. epidemic, but any official who claims that no one can know the percentage of their populace that is infected because of too few tests is either incompetent or lying.
Anyone who has taken a statistics course knows that we can get a pretty good bell curve with as few as eight random samples.
Ideally, we might like to have several hundred people selected pretty much at random for testing. Depending on the size of the sample, the confidence in the estimate of the number of people with the Covid virus can range from fairly sure to virtually certain.
The last I looked, 2,000 random samples would give a near 100 percent confidence level for the entire U.S. population. It's curious that the press uses popularity polls all the time that use that scientific method. But most in the media seem clueless here. They should be asking what these figures are. How do they change from week to week, whether nationally, or by state, county or metro region?
Public officials, mostly Democrats, who hurl abuse at President Trump over the purportedly insufficient number of tests available are either ignorant (they slept through their college statistics course) or recklessly concealing information from the public. Those governors who say they will base decisions on restoring liberty to their citizens don't even release basic scientific statistical data so that the scientists in their states and elsewhere may evaluate it.
In any case, from what I can tell, a number of the computer models that project epidemic numbers are using the wrong input data. The most basic input data should come from the percentages of the populace infected at two points in time. Hospitalization rates and "confirmed" Covid case rates and other such data constitute useful secondary data. If health officials are not doing the type of random sampling outlined above, then the models are, pretty much, junk. As they say in computer land, "Junk in = junk out."
In addition, the claim by the FDA and others that the efficacy of hydroxychloroquine could only be established by clinical trials is not altogether accurate. The best information would come from random trials. But here is a method that would give a pretty fair indication:
Poll a random sample of ICU chiefs with two questions. Do you routinely administer hydroxychloroquine for five or more days upon admission? What is the average ICU stay of your Covid patients -- regardless of chloroquine use? [In addition, it would be helpful to ask, Have you observed adverse side effects that you suspect result from use of hydroxychloroquine?]
One can then do a scatter plot and obtain a regression line showing whether there is a correlation between hydroxychloroquine use and length of stay in the ICU. One must quiz enough ICU chiefs in order to filter out specific local policies (by the tendency of minor causes to cancel) on ICU length of stay.
Ron Paul on the mounting backlash
https://www.youtube.com/watch?v=H0OOqq29Ios
Details of econ restart plan
https://www.scribd.com/document/456756096/White-House-guidelines-for-reopening-economy#from_embed
Question: Will we need ID to prove we aren't vulnerable?
In other words, suppose, in Phase 1 of the White House plan, I am ambling about. I am in the senior citizen age bracket, as is apparent from my craggy, saggy face. Will police officers be checking my ID in order to verify that I am permitted to be out and about? This should be a silly question, but in the current atmosphere, it isn't, especially because different governors will be deciding on how the plan is to work, if at all.
Answer: the White House says the vulnerable "should" avoid people. An ambiguous word leaving plenty of room for abuse of power by state and local penny ante dictators.
Greenwald: Emergency powers rarely go away
https://www.youtube.com/watch?v=Nd7exbDzU1c
It may have been true in the first few days of the U.S. epidemic, but any official who claims that no one can know the percentage of their populace that is infected because of too few tests is either incompetent or lying.
Anyone who has taken a statistics course knows that we can get a pretty good bell curve with as few as eight random samples.
Ideally, we might like to have several hundred people selected pretty much at random for testing. Depending on the size of the sample, the confidence in the estimate of the number of people with the Covid virus can range from fairly sure to virtually certain.
The last I looked, 2,000 random samples would give a near 100 percent confidence level for the entire U.S. population. It's curious that the press uses popularity polls all the time that use that scientific method. But most in the media seem clueless here. They should be asking what these figures are. How do they change from week to week, whether nationally, or by state, county or metro region?
Public officials, mostly Democrats, who hurl abuse at President Trump over the purportedly insufficient number of tests available are either ignorant (they slept through their college statistics course) or recklessly concealing information from the public. Those governors who say they will base decisions on restoring liberty to their citizens don't even release basic scientific statistical data so that the scientists in their states and elsewhere may evaluate it.
In any case, from what I can tell, a number of the computer models that project epidemic numbers are using the wrong input data. The most basic input data should come from the percentages of the populace infected at two points in time. Hospitalization rates and "confirmed" Covid case rates and other such data constitute useful secondary data. If health officials are not doing the type of random sampling outlined above, then the models are, pretty much, junk. As they say in computer land, "Junk in = junk out."
In addition, the claim by the FDA and others that the efficacy of hydroxychloroquine could only be established by clinical trials is not altogether accurate. The best information would come from random trials. But here is a method that would give a pretty fair indication:
Poll a random sample of ICU chiefs with two questions. Do you routinely administer hydroxychloroquine for five or more days upon admission? What is the average ICU stay of your Covid patients -- regardless of chloroquine use? [In addition, it would be helpful to ask, Have you observed adverse side effects that you suspect result from use of hydroxychloroquine?]
One can then do a scatter plot and obtain a regression line showing whether there is a correlation between hydroxychloroquine use and length of stay in the ICU. One must quiz enough ICU chiefs in order to filter out specific local policies (by the tendency of minor causes to cancel) on ICU length of stay.
Ron Paul on the mounting backlash
https://www.youtube.com/watch?v=H0OOqq29Ios
Details of econ restart plan
https://www.scribd.com/document/456756096/White-House-guidelines-for-reopening-economy#from_embed
Question: Will we need ID to prove we aren't vulnerable?
In other words, suppose, in Phase 1 of the White House plan, I am ambling about. I am in the senior citizen age bracket, as is apparent from my craggy, saggy face. Will police officers be checking my ID in order to verify that I am permitted to be out and about? This should be a silly question, but in the current atmosphere, it isn't, especially because different governors will be deciding on how the plan is to work, if at all.
Answer: the White House says the vulnerable "should" avoid people. An ambiguous word leaving plenty of room for abuse of power by state and local penny ante dictators.
Greenwald: Emergency powers rarely go away
https://www.youtube.com/watch?v=Nd7exbDzU1c
No comments:
Post a Comment