The question observational studies cannot answer

I suppose that and observational study found that red meat consumption above 300 grams a week increases the risk of colorectal cancer by 10%.
Is it a random distribution of the relative risk or is it a selective risk in a sub population with both genomic and environmental factors which are linked with the carcinogenic effect of red meat?
The difference between a random effect and a selective effect on redmeat conception in some people only cannot be solved by observational studies.