[robertlagrant]

There were upper and lower bound stats that accounted for many discrepancies. The high numbers were "death with covid", which means the covid might not've done it, and the low numbers were "death by covid" where we were (fairly) sure Covid did it due to lack of other factors, but missed things where Covid tipped someone over the edge.

[tightbookkeeper]

That sounds right. Based on my reading the CDC numbers are basically death with Covid test or symptoms, with the exception of an obvious distinct cause.

Which honestly I can’t think of a better way for mass data collection.