Assume each member of the population under study, labeled i, has two possible outcomes, Y0i and Y1i which are the outcomes that i would display if it didn't get the project, Ti =0, and if it did get the project, Ti =1. We would like to know something about the distribution over i of the effects of the project. What we can observe in the data is the difference between the average outcome in the treatment and the average outcome in the controls: E(Yi1|Ti=1)-E(Yi0|Ti=0)

(1) Explain why this difference could be a biased estimate of the effects of the project. (2) Explain how randomization could solve the problem in (1).