Yeah, small population size creates fake outliers. "Rural states have both the highest and the lowest rate of kidney cancer" - what they in fact have is small populations, so they just have much larger variance.

It's not "just". Cancer isn't purely random on its own authority. It has causes.

Variance could be due variation in environment, pollution, personal habits, quality of record keeping, and many other meaningful or spurious causes. Outliers and 95%iles are worth studying, not merely writing off.

Iceland's models might not work for larger countries due to size difference, and might not work for other cities due to differences in sovereignty structure, but it's worth investigating.

While the data might not be ‘just’ caused by random variance, it provides little evidence that there is something more going on than just randomness.

You have no particular reason to study those cases more than anything else, since the data we have does not suggest that there is anything actually going on.

Things are not as simple as one or the other.

The uneven spread of heavy industry creates pockets where cancer rates are higher becausw of environmental factors.

Random variance is responsible at a micro-level but at a macro level the trend still holds true.

I wasn't making any claim about the actual causes... I was simply stating that the statistical effect you see can be completely explained by random variance, so you can't use the result as an argument for a cause. You have to search elsewhere for that evidence.

Since people frequently miss this point, the population of Iceland is approximately the same as the Kalamazoo, MI metro area.