Note: If the data here is old, try to refresh the page using ctrl-F5 on Windows and cmd-alt-r on macOS, and/or clear your browser’s cache.
There are different ways to calculate the mortality and here I present what I call the “incremental” mortality. This is the deaths for a given day divided by the new infections for that day. The problem with this is obviously the lag between when someone gets sick and possibly dies. Over time as infections and deaths reach a steady rate this measure should approach the true value. This way of measuring mortality is not skewed by the initial undercounting of those infected, but there is still likely undercounting of infections occurring.
Antibody testing now being prevalent raises another question, does the number of infections consider people that were never tested while they were infected? The problem is determining how to count those people, i.e. do you add them as new infections or try to modify the old data based on when they believe they had the virus. While this might give us more accurate historical data, it would make keeping track of things very difficult. Even though it is important to know how many people have antibodies and, thus, not likely to be reinfected, it is more important to know how many new infections there are, thus the data here tracks new infections, i.e. those tests while they had the virus.
States sorted by population
Click to enlarge.
