It feels like this is something that should be easily amenable to statistical analysis.
Large numbers of people get diagnosed with lung cancer; smoking rates vary widely and smoking is typically recorded in patient histories; typical radon levels vary widely with geography and are recorded / known. This looks like a near perfect natural experiment, no?
Large numbers of people get diagnosed with lung cancer; smoking rates vary widely and smoking is typically recorded in patient histories; typical radon levels vary widely with geography and are recorded / known. This looks like a near perfect natural experiment, no?