One tip: if a model shows a symmetrical curve then it's too simplistic and wrong - a realistic model would show a sharper growth and slower decline after it's leveled off.
Simon ...
Agreed. But as I see it, the entire sampling process is at best questionable, or more realistically, unreliable.
Since statistically taking an entire population is not possible, a sample is the only realistic method to do it. However, we are not taking a random sample to arrive at a mean since there are restrictions on who is getting tested and who is not. IMO, you would likely need to take the samples on a case by case basis and then somehow find some correlation between the various elements.
If you lower the testing rates and less people are positive, how do we interpret that?
If you increase the testing rates and more people are positive, how do we interpret that?
It feels like the more I think I know the less I actually know.
Rub a Dub