There is a good rule of thumb to use when looking at associative studies - if the effect is not 300% or more then it is unlikely to hold up when tested in a dou

There is a good rule of thumb to use when looking at associative studies - if the effect is not 300% or more then it is unlikely to hold up when tested in a double blind study. Anything under 3 fold is mostly noise or bad experimental design. As a comparison the risk of lung cancer is 1500% higher in smokers than non-smokers.