ECO331H1 Lecture Notes - Lecture 2: Internal Validity, Nash Equilibrium, Field Experiment
Document Summary
Pay for performance, which sometimes means paying for preferences circumstances fixed or random. Allow, in principle, variation of putative causal factor while keeping all other relevant. Allows researcher to control all observable characteristics. However, there are still many characteristics that are unobservable to researchers: hope to resolve unobservable characteristics through randomization so average participant is same across all treatments, replicability. In theory: publication bias can only replicate experiments that have been published. Is hard to publish a study someone else already did so few academics will replicate in social sciences. Replication crises in psychology: many well-known studies in psychology were recently replicated and researchers failed to replicate results. Testing theory for robustness: searching for facts, start with theory and see if experiment results match theory, economic engineering. Will not look at this in this course. Effects of variables about which theory is either silent or cacophonous. Search for most likely nash equilibrium if more than one.