I blogged the other day about James Conca's latest masterpiece in Forbes. I tried to correct him on a few things in the comment section, but he seems insistent on his pseudo-epidemiology which goes something like this (if I understand him correctly):
"There are 6 billion people on the planet, cancer rates aren't much different across the planet, natural radiation levels vary across the planet, therefore low level radiation doesn't cause cancer".
You just can't make this stuff up!
Cancer rates do vary as does natural radiation levels.
Other possible confounders also vary across the planet like smoking, diet, diesel exhaust, etc.
And different sub-populations (of the global one) have different genetics.
In order to determine radiation's cancer risk, one has to define a specific population (global sub-population) of interest. Then one has to determine the actual doses received by individuals, one can't just use background radiation levels. Why? Because they are not homogenous, and people travel to different areas, sometimes they are indoors, sometimes outdoors, different construction buildings will affect the dose, etc.
And after doing that, the people need to be followed until they die.
Then compare the cancer mortality to the absorbed doses and see what the relationship is. Any identified confounders for that population should be identified and measured (if possible).
Conca is committing the ecological fallacy on a global scale!