ECON 104 Lecture Notes - Lecture 47: Scientific Control, Clinical Trial, Regression Analysis
Document Summary
Analysis at country level, need to ensure that some countries follow a policy randomly - which didn"t happen. Ethical problems when you randomise clinical trials of drug - who gets into the treatment plan and how gets treated within the randomised trial. Experiments are expensive - need a lot of money to have extra capacities. Experiment was designed and implemented consciously by human researchers, and experiments randomly assigns subjects to treatment and control groups. Quasi experiment - main regressors is almost randomly assigned to objects in regression analysis, but no one intended for the regressors to be analysed in an experiment. Sometimes it"s just nature, some countries are/ are not affected - since this is random, it"s as if nature is running an experiment. There"s no random assignment of treatment - not done with intention of being random, but it turns out to almost be as if it were randomly assigned.