Hacker News new | ask | show | jobs
by giantg2 1627 days ago
"Are you trying to defend the notion that COVID isn’t any worse than the flu?"

Nobody in this chain has suggested it. The person you responded to said that covid cases and death stats are garbage in garbage out. I would largely agree considering the underreporting of asymptomatic cases as well as the lack of distinction between causative and non-causitive hospitalizations and deaths.

The very argument you are making about possibly fewer deaths from other things supports my argument - the data doesn't exist to cover all the variables and some of the data is garbage. There is a huge difference between direct and n-order deaths, both for covid and other things. We could compare either one so long as we are all on the same page, but what's the point? The conversation was about data quality and not about lethality.
1 comments
cromwellian 1627 days ago
Do you think the data about excess deaths is less reliable than VAERS? On the one hand, we have actual factual data possibly with an unproven (but highly likely) correlation and on other hand, we have VAERS which can simply be spammed by anyone.
link
giantg2 1627 days ago
This feels like you are trolling at this point. Nobody said they are less reliable. The excess deaths weren't even a point in this discussion until you brought it up. The other commenter specific mentioned covid cases and deaths.

You mention "factual data". The excess death numbers are estimates, per the CDC. They are approximate and could be changed as new information comes to light, etc. No problem with that, but not exactly facts as far as attributing a hard death count.

The issue with the VAERS data isn't people "spamming" it. It's that legitimate potential events are not reported. The purpose of VAERS is to provide early warning. This means people are supposed to report events that don't have a different proven cause. Noise from coincidental events could then be removed using population baseline occurrence. I'd the data simply isn't there, then it's not very useful for that purpose. I believe they also perform some cursory validation with the provider before accepting/confirming it, but I could be wrong.

link