01:220:322 Lecture Notes - Lecture 1: Randomized Experiment, Econometrics, Randomized Controlled Trial
Document Summary
Suppose we wish to determine the effect of a new drug on the incidence of heart disease in human patients. This difference is known as the treatment effect: any difference in the means of the two groups has to be because the two groups received different treatments. This difference is known as the treatment effect. (cid:1) (cid:1) (cid:1) If the assignment of individuals is not random then we have some problems: Suppose, for example that group i contained all individuals who were genetically predisposed( ) to have heart problems and group ii had people who were not predisposed. Then, without treatment we would expect the incidence of heart problems to be higher in group i than in group ii , all other things being equal. Suppose without knowing that the groups were not chosen randomly we run the experiment as before.