>>In fact, in Western Europe, there is a striking correlation between population weighted density – which may influence both the speed of virus transmission and the amplitude of infection waves – and covid mortality per country (see map below). This correlation does not apply, however, to Denmark, which has a low covid mortality but an average population weighted density (3434).
To make clear, I was incorrect in generalizing all the other Nordic countries as having a lower population-weighted density. But it's an outlier and I think the support for the broader point is still strong.
I'm sorry, but when your sample size is 4, you can't throw out one of them and call it an outlier because it doesn't fit the narrative. This sounds like classic cherry picking to me, especially when the narrative is one the author chose specifically to make a comparison favorable to their main argument! For all the submissions on HN decrying bad statistics, you'd think this kind of argument wouldn't fly here.
"We can't compare against the Nordic countries because they are not like Sweden, and we can't use population density because that's the wrong way to look it. Instead we look at population weighted density, and we throw out the one that causes our argument to fail. Then when we compare to Austria, look what a winner Sweden is!"
I understand that there's some merit to the line of reasoning you are presenting, especially population weighted density, but as far as the article goes, to just throw out Denmark as an outlier because to include it would refute the central argument of the article really puts it in a bad light.
> in my opinion
Right, let's just be clear that's what this is.
"There's no accompanying analysis to support its rejection from the comparison, let alone an attempt to quantify exactly how much of an outlier it actually is."
There's no accompanying analysis to support its inclusion in the comparison, let alone an attempt to quantify exactly how comparable it actually is.
Claiming Sweden underperformed Denmark is also a far cry from claiming that Sweden underpeformed other Nordic countries as a whole, as the original article does.
Is it though? Because neither you nor the linked piece have provided any framework by which to call it an outlier. Calling something a statistical outlier has no meaning without statistics.
> There's no accompanying analysis to support its inclusion in the comparison
The case for its inclusion is the fact that it's a country in the region, and we're comparing countries in the region. You don't need a good reason to include it, you need a good reason to exclude it, especially when it's the only one excluded, and again, especially when the title of the piece contains the word "vindicated". Calling something an outlier and causally tossing it aside, especially when it contradicts your point is classic cherry picking.
That's entirely false. The study provided the framework, which is a map and graph showing the countries of Europe and both their population weighted density, and their COVID mortality rates, respectively:
https://i0.wp.com/swprs.org/wp-content/uploads/2021/12/popul...
https://i0.wp.com/swprs.org/wp-content/uploads/2021/12/swede...
It's plainly a statistical outlier.
>>The case for its inclusion is the fact that it's a country in the region, and we're comparing countries in the region.
There's no accompanying analysis to support grouping all countries of a region together for a comparison.
You need a good reason to include it, given you're using it a reference point to criticize a partricular country's response.