MATH 148 Chapter Notes - Chapter 2: Confounding, Observational Study
Document Summary
Control subject who did not get the treatment. Controlled experiment study where investigators decide who will be in treatment group and who will not. Observational study subjects assign themselves to groups, investigators just watch. Ie: comparing smokers (treatment group) to non-smokers (control group) Heart attacks, cancer, other diseases more common among smokers- strong association between smoking and disease. If cigarettes cause disease, it explains association that death rates are higher for smokers because cigarettes kill. Could be confounding factor that makes people smoke and makes them get sick. If there is confounding, something has to be done, but we can"t expect perfection. Statisticians talk about controlling for confounding factors. Make comparisons separately for smaller, more homogenous groups (ie: crude comparison of death rates among smokers and non-smokers could be misleading because smokers are mainly men and men are more likely to have heart disease anyway)