I shelved a post on COVID-19, as not well written. Essentially I said that, apart from existing statistics, one needs an unbiased sample of a population. That is critical, but we don't have such sample anywhere. And we have only one complete sample, which is the Diamond Princess -- 1% mortality among an older population on a ship -- or rather, 1% of those who died had COVID-19 in their blood. But having said that, I was napping. A friend points out that the COVID-19 death rate is a subset of the total death rate. We don't have a clear idea of that relationship yet. OBSERVATION: It illustrates the importance of getting the academic questions right. We are in lockdown in South Africa, and the army has been deployed. Yet here and all over the world, the threat is still largely unknown. We do have death tolls, yes, but death tolls in themselves reveal comparatively little. Perhaps this is a sign of the times: we have become pragmatists so-called, rather than thinkers.
POSTSCRIPT: I discussed it with Lead Data Scientist Katharina Fenz, of the World Data Lab in Vienna. She wrote to me (30 March), "Unfortunately, you are right that the availability of reliable information regarding COVID-19 is still very limited ... [not] enough information to draw reasonable conclusions."
No comments:
Post a Comment