For most other things you would do a placebo controlled trial, but it might be hard to get people to sign up for a trial where you're giving them enough radiation to expect to cause a significant number of them to get cancer.
So controlling for confounders is really the only thing left if you want an answer.
And it works pretty well -- at least better than nothing -- if you actually want the answer. The problem with it is that it unfortunately can also produce whatever answer you want from a political perspective by reverse engineering which confounders to control for based on which ones shift toward the desired outcome, so you can't trust somebody else to do it for you without checking their work at a level of detail that basically amounts to redoing the analysis yourself.
> It turns out skin cancer is a solid proxy
Wait, what are you using it as a proxy for?
It’s not great, but skin cancer is more closely tied to radiation exposure rather than everything that can cause lung cancer etc.
Plus, you still have the same problem with confounders. Expect more skin cancer in a population that spends more time outside, but also potentially a very different lifestyle in that case.
And problems with radiation type as well, since many types of radiation don't penetrate the skin, so they could cause skin cancer but not other cancers. Which has implications for the fatality rate because skin cancer has a much higher survival rate than most other types.
The problem with this assumption is you can’t keep everything else identical. Population X and Y don’t grow up live and die in exactly identical areas expect for different radiation levels.
What you can do is control for as many factors as possible or use a proxy. It turns out skin cancer is a solid proxy, * and populations like Australia and skin cancer support the linear model down to the limits of our ability to detect it. But, then people want to suggest the non linear model kicks in as soon as the signal falls below the noise floor without evidence, even though that’s the same behavior in both models.
* As in you can detect population differences and work out whole body radiation dose equivalents.
PS: By this logic we should be operating under the assumption that the Greek Gods actually exist because the theory is very old and it is difficult to measure. People are promoting a new idea, not just trying to disprove an old one. Aka, Greek gods and the flying spaghetti monster have equal support. That’s absolutely fine, it’s saying some other untested theory is supported without evidence that’s at issue.