Revisiting Coronavirus Data Analysis

kirstenhacker.wordpress.com

1 pointnixtaken2 comments

Not data analysis, just weakly argued opinion. Author even accuses “Germany’s virologists” of “lying”.

While this seems like pessimistic bias I think much of the confusion about mortality rates stems from the multiple definitions. Is the denominator infected, symptomatic or hospitalized? These three estimates - still very uncertain today - will vary drastically.

Here’s some actual data analysis on this subject: https://www.nature.com/articles/s41591-020-0822-7

If the death rate turns out to be 0.5% as claimed by the virologists on the German nightly news, I'll stand corrected. They say that the dark number of unreported cases is a factor of ten of what is reported, but if it is a factor of ten in Germany and South Korea, then it has to be a factor of 100 in places like the US, UK, Spain, and Italy. I like clear data presentation and many people mistake complex data presentation for quality. Just compare the CDC site and the worldometer site. Anybody can do simple calculations to interpret worldometer data, but the CDC site presentation is harder to parse. I think this is a deliberate form of gatekeeping. Incidentally, from where I am in Germany, I can only view the first page of the Nature article you linked. As with all articles of its sort, the devil is in the details.

Reply on news.ycombinator.com