Thread : this is going to be my penultimate post on Swedish mortality & covid. After SCB's next update, Jan 18th, we can consider the prel. numbers for all cause mortality 2020 fully stable, and I will then do a final update. What follows below, is still subject to minor updates.

2/n: To get ahead of media, who no doubt any hour now will report "Sweden has highest number of deaths since 1918!", here's what the absolute deaths deaths look like:

3/n: So expect media headlines painting Sweden as "The World's Cautionary Tale". However, with a little bit of thought and analysis, we might reach different conclusions... Let's first take population size into account:

4/n: when adjusting for population, 2020 as we speak (remember, the numbers are not fully finalized yet) has mortality on par with 2015. Can you spot the two outliers...? (hint: 2020 is not one of them).

5/n: As I've mentioned earlier, it's not enough to adjust for population - different years have quite different demographics in terms of age. So we need to take that into account, by adjusting for age:

6/n: Another way we can illustrate the importance of age is by looking at the proportion of deaths - actually: the proportion survivors - in a number of age intervals (I'm using only 4 age intervals, since that's how the SCB prel data are presented): Can you spot the outlier...?

7/n: Finally, let's look at one of those very popular Linear Regression models for Swedish deaths: first, non-age-adj. mortality 2010-2020 : can you spot any outliers...?

8/8: And lastly, a Linear Regression for age-adj mortality: again, can you spot any outliers...?

As mentioned, in a week, I'll do one more posting. After that, I'm off Twitter & most other social media.

@threadreaderapp unroll, pls.