3 Comments

Very old hat. Statistically-minded Substack denizens such as El Gato Malo and Prof Norman Fenton have been exposing this statistical chicanery (which goes back to Pasteur) for the past two years.

In fact Professor Fenton has shown that, in any normal viral epidemic situation, inclusion of the 0-14 day vaccinated in the unvaccinated category will inevitably lead to the initial appearance of a high degree of vaccine efficacy, even for a placebo. This arithmetical effect diminishes over time (as it must) - exactly as the supposed benefits of vaccines were seen to diminish over time.

Expand full comment

The definitional bias you mention is hugely important as well, and is part of the journal article I'm working on.

Expand full comment

The efficacy is negative in the first weeks.

The mass injections are kicking off waves.

Older vaccinations and previous infections protect to some degree.

Since vaccination rates and daily injections correlate highly across regions, you need to look for times when new daily injections p.c. and total adminstered injections p.c. diverge.

In the USA this happened during Delta, Q3 2021.

That's where you can see the real dynamics. The rest of the time these dynamics are masked by collinearity.

https://q3deathwave.pervaers.com/

Expand full comment