All research designs have their strengths and limitations. A comparison of lung cancer in the US and India with average tobacco intake in the two countries could say something about whether cigarette smoke causes cancer. So could following two groups, of smokers and of non-smokers, over ten years, to see how many get sick, and how many die, and of what. A look at a death register, tallying causes of death to the dead people's occupations could yield information on occupational risks. Comparing school children's exam score with their blood lead levels indicate the effect of lead.