4 ms·

TBH, I think the Lead Stories factcheck is slightly hand-wavy. Let's do some back-of-the-envelope calculations based on their numbers: - The breakdown of Covid

TBH, I think the Lead Stories factcheck is slightly hand-wavy. Let's do some back-of-the-envelope calculations based on their numbers: - The breakdown of Covid-19 deaths for the month of August in Scotland seems to be as follows: 75% fully vaccinated, 20% unvaccinated, and 5% partly vaccinated. - The Covid-19 vaccination progress in Scotland seems to be like this: 84% have had both doses, 9% are not vaccinated, and 7% have only had their first dose. Now, we can try to estimate the vaccine's efficacy as follows: the population of fully vaccinated people is ~9.33x the unvaccinated (0.84/0.9), but their number of deaths is only ~3.75x (0.75/0.2). That gives an estimated efficacy of around 60%. (https://www.wolframalpha.com/input/?i=1+-+%280.75%2F0.2%29%2F%280.84%2F0.09%29 https://www.wolframalpha.com/input/?i=1+-+%280.75%2F0.2%29%2...) Again from the numbers in the factcheck, the advertised efficacy is around 75 to 99% for the AstraZeneca vaccine.

eynsham 5y ago
Older people are more likely to die (by several orders of magnitude), and more likely to have been vaccinated.
- finite_jest 5y ago
  That's a fair point. My napkin math doesn't account for demographic differences between the vaccinated and unvaccinated. However, the factcheck does not even mention your point. I guess one could argue that's yet another reason that it is indeed hand-wavy, and doesn't live up to its promise of "telling the whole story" :-)
phonypc 5y ago
Worth noting that when you break down the data into age groups the efficacy is substantially higher than that overall number, since the vaccinated population skews older. An example of Simpson's paradox. Good explanation here: https://www.covid-datascience.com/post/israeli-data-how-can-efficacy-vs-severe-disease-be-strong-when-60-of-hospitalized-are-vaccinated https://www.covid-datascience.com/post/israeli-data-how-can-...
- finite_jest 5y ago
  Very interesting, and certainly much more nuanced than the factcheck. I had heard of the Simpson's paradox in another context (university admissions), but really didn't expect it to pop up here.

eynsham 5y ago

Older people are more likely to die (by several orders of magnitude), and more likely to have been vaccinated.

finite_jest 5y ago
That's a fair point. My napkin math doesn't account for demographic differences between the vaccinated and unvaccinated. However, the factcheck does not even mention your point. I guess one could argue that's yet another reason that it is indeed hand-wavy, and doesn't live up to its promise of "telling the whole story" :-)

phonypc 5y ago

Worth noting that when you break down the data into age groups the efficacy is substantially higher than that overall number, since the vaccinated population skews older. An example of Simpson's paradox. Good explanation here: https://www.covid-datascience.com/post/israeli-data-how-can-efficacy-vs-severe-disease-be-strong-when-60-of-hospitalized-are-vaccinated https://www.covid-datascience.com/post/israeli-data-how-can-...

finite_jest 5y ago
Very interesting, and certainly much more nuanced than the factcheck. I had heard of the Simpson's paradox in another context (university admissions), but really didn't expect it to pop up here.