Adding on to and clarifying the University of Illinois testing program: Since the advent of their testing efforts, the University has made several changes to how it calculates % pos and defines a "case". Let's look... 1/* https://twitter.com/EWoodhouse7/status/1307438758824992768

Here's the "case" graph directly from UofI's testing dashboard, annotated with their own definitions featured here: https://go.illinois.edu/COVIDTestingData Note the change in definition occurred the day they reported their highest number of pos tests. 2/*

Here's the "case" graph annotated with UofI's own explanations of changes to calculating % positive. First change came on the date of highest "case" number. Second change came on the day of an "uptick" too. 3/*

Interestingly, with it's new explanation of how positivity rate is calculated, UofI says "in accord with standard practice" and links to @JohnsHopkins COVID tracking site https://coronavirus.jhu.edu/testing/individual-states Why was the school using "standard practice" previously? Did the standard change? 4/*

So, w/all that in mind, I revised the graph I tweeted earlier 2day, which had shown the total no. of positive tests without noting the definitional change of "case". See revision below. The 876 pre-Aug 31 includes >1 pos test/person. The 1,272 from Aug 31 to current does not. 5/5

P.S. Here's a summary of the University of Illinois reactions to their "rising" pos rate & unique new cases. Corresponding emails linked below:

9/2/20 https://massmail.illinois.edu/massmail/61640956.html

9/7/20
https://massmail.illinois.edu/massmail/1094914864.html

9/16/20
https://massmail.illinois.edu/massmail/1329736540.html