lukestein’s avatarlukestein’s Twitter Archive—№ 7,594

  1. This👇 is a good summary discussion of ten problems with inference from COVID-19 data: 1 Test rationing 2 Test inaccuracy 3 Retesting 4 Treatment lags 5 Changing standards 6 Reporting lags 7 Reporting errors 8 Attribution errors 9 Manipulation 10 Heterogeneity t.co/3ZO88PlKTX