12 Comments

I am still digesting your work...

Are you on Twitter? Take a look at this: Excess Mortality in EU countries plotted against Vaccine Doses over time https://twitter.com/_shloka/status/1620888062812364806?s=20&t=6Ubc4UyU795RJV075ChQrQ

The inflection point was April 2022!

Expand full comment

Thanks!

No, I am not on Twitter. If you ever spot me there, feel free to shoot me... :)

Expand full comment

You're a wise man!

Expand full comment

Thank you! :)

They opened schools in late 2021.

When a student got a positive test, this was forwarded to the federal state's public health department. They forward the results to the RKI (once per day I think).

Now, this result has to be confirmed with a PCR test, but here comes the kicker:

Meldepflicht only applies to positive results, so if the PCR test came back negative, this did not need to be reported. :D

But it gets better. If the test came back negative and the school decided to still report it, there was a deadline. I think it was Friday of the same week, 5 pm, but I am not sure.

You can imagine what this means for false positive antigen test results on Thursdays and Fridays.

EDIT: I haven't quite understood what you did with the DIVI data, but the result is intriguing.

I think the higher frequency of waves is due to the approval of Monupiravir. I am going to dedicate an article to that, but I have a couple of more that need attention before that.

Expand full comment

Here are the data behind the DIVI diagrams (a little untidy, particularly the attic).

https://www.file-upload.net/download-15087396/baselines.xlsx.html

Expand full comment

Mein Problem ist eher, dass ich mit Bootstrapping nicht vertraut bin. Muss ich mir gründlich anschauen wie das funktioniert. Ich bin ja eigentlich kein waschechter Analyst, nur Hobbyist.

Expand full comment

An dieser Stelle ist es eigentlich nur ein Rückrechnen einer Größe aus Daten, die auf den ersten Blick gar nicht dazu geeignet scheinen. Bootstrapping als statistische Technik gibt es natürlich auch (Erzeugung künstlicher Daten durch Ziehen mit Zurücklegen aus den Originaldaten), das wird aber eher zur Fehlerabschätzung genutzt.

Expand full comment

Not sure why I started replying in German. Good thinking on the age asymmetries you noticed!

Expand full comment

Answering in German is like using strange terms like bootstrap: it mystifies the uninitiated...

Expand full comment

Achsoooooo. Ich hatte nachgeschlagen was das bedeutet und war dann endgültig verwirrt. :D

Expand full comment

I'm reading through the Kuhbandner paper after reading Josh Guetznow's new post. Looking forward to hearing your thoughts and how it compares with your own approach to excess death estimates.

Also, I tried to send you a mail to see if you might be interested in chatting together to compare notes and share ideas also with Fabian.

Expand full comment

Mail was delivered successfully, but went into my spam folder. Check yours if you do not see my answer.

Expand full comment