vesicles
Colonel
Yes, a power analysis would allow you to figure out a minimal sample size. However, how to choose a "truly random" sample set is an art. When we read a clinical paper, we typically spend most of our time evaluating their sampling table (typically Table 1 detailing the demographics of their samples). Have the authors picked a truly random sample? With a sudden viral outbreak like the COVID-19, it would become so confusing how you should choose your sample. Let's assume that we have 3 groups that try to conduct their own statistical studies. I can almost guarantee you that they will give you 3 completely different results because they have chosen different samples.
To be honest, my own feeling is that figuring out an overall infection rate would be a little pointless. Infection rate is dependent on so many different external factors, most of which unrelated to the contagiousness of the virus itself. A good example would be a comparison between Italians who love to hug and kiss and Chinese who wear face masks all the time. Also, China has administered strict quarantine and lockdown. In comparison, Americans downplay the virus and would not do anything major. We will have vastly different infection rates in these countries. Even when you compare Wuhan and the rest of China, you will get vastly different infection rates. What does that tell you about the virus itself? Nothing. Your chance of getting infected will differ so much when you go to these different countries.
A more effective way of looking at this thing is to figure out the main routes of transmission. Air? Fluid? Touch? Aerosol? Once we are clear of how the virus is transmitted, we can do things to protect ourselves. No matter how weak a virus (not very contagious) is, if you start kissing an infected person on the mouth, you will be get infected with almost 100% certainty. Even if you have something highly contagious like Ebola (air-transmitted), you will not get infected if you stay at home and isolate yourself.
