Unanticipated variables obscure the true cause and effect relationship.
uncritical analyses show relationship btwn con founder and disease.
Solution; stratify data analyses on levels of co-founder
i.e Diabetes rate in obese/non obese in regards to cholesterol level. Given RR. By stratifying we see Obesity likely cause of DM II, not cholesterol.
i.e Smokers, drinkers, and cirrhosis. Stratify and look at high/low alcohol use and determine Alcohol (not smoking) is cause of cirrhosis.