PSCI 2701 Lecture Notes - External Validity, Experiment, Internal Validity
Document Summary
Psci2701b class 7 oct 22 2013 (from lecture 6 slides) Ideally you do experiments several times to correct for errors or abnormalities in just one and use the average. Laboratory experiment: in a controlled setting, looks like a clinical trial. Field experiment: instead of having people come to you in a controlled setting, you go into a natural environment and conduct it there, still you are forcing the treatment. Natural experiment: you go into a natural setting, but you are just observing something that will naturally occur, you are not adding anything to the natural environment for the participants (maybe just doing a survey before) Everyone has an equal chance of being exposed to the treatment. Only way you can prove causation with your measurement/design. *professor accidentally used the senate for the graphic example and the house of representatives for the written example on the previous slide*