kollel?man> They are counting anyone vaccinated within 14 days as unvaccinated.
in this paper, I believe they would include into “unvaccinated” – indeed those who were just vaccinated and also those who were already infected. Both would affect the numbers. Let’s see how we can clean it up:
1) there is separate data for immediate 2 weeks after vaccine. First, the Phase 3s answered that question clearly for the original virus. It was also reviewed for Delta later on, for example in Israel. It is the easiest to check what happens exactly after vaccine is given.
2) You can also separate effects of vaccination from recently vaccinated: in July-August, the rate of vaccination was very low (0.2% per day) and most were for younger people, as older ones were already vaccinated some time ago and not yet getting boosters: so those numbers should be reliable.
3) including those who were previously infected affects numbers much more. But presuming that previous infections decrease new infection (as many reasonably think), this would only make numbers for not infected even higher. Effect of previously infected on vaxed is harder to bound, but as the overall effect is on the order of 10, and proportion of previously infected is (way) less than 50%, the effect would still exist under very stringent assumptions.
4) similar results come from Israel and England, where vaccinations were performed much faster, so confusing effects are much smaller.