[joshuahedlund]

Simpson’s paradox is fascinating. If the result above does not seem intuitive, think of an extremely simplified version: imagine that half of everyone in their 50’s is vaxxed and no one on their 20’s is. Next imagine that of 1000 people in their 50’s, 100 of 500 vaxxed people die from non-covid sources, 100 of 500 unvaxxed people die from non-covid sources, plus 50 unvaxxed people in their 50’s die from covid and no vaxxed people do. And imagine that 0 out of 2000 people in their 20’s die from anything.

150 out of 2500 unvaxxed deaths in 18-59

100 out of 500 vaxxed deaths in 18-59 (higher rate)

But: 150 of 500 unvaxxed deaths in 50-59

100 of 500 vaxxed deaths in 50-59 (lower rate)

It would be true that more vaxxed people died than unvaxxed in the 18-59 age group without the vaccine causing any death, but only because the larger number of surviving young people swamps the higher relative death rates among the higher relative vaxxed elderly.