Semi serious question, as I'm not a stats guy: We have a voluntary, opt-in testing program. Any voluntary testing that happens outside the U is asked, not required, to be reported.

So then how can this statement be true: "The main focus of the COVID-19 Community Sampling and Tracking Program is to systematically look for increases in COVID-19 that may be missed by other testing programs...

...We will do this through the testing of a representative sample of the U-M community each week."

(from here: https://campusblueprint.umich.edu/monitoring-testing/)

Is the claim of "representative" from this non-random, opt-in sample coming from some demographic weighting? I guess what I'm asking is: How is opt-in "representative" and "systematic", both technically and rhetorically?

Ok, so I'm poking around the FAQs and seeing responses and I think I get it: Students can opt-in voluntarily to the program. The program participants are randomly sampled, but can decline to be tested and can remove themselves from the program at any time.

Or, as @asociologist put it: You’re getting a representative sample of an unrepresentative population.