POLI 30 Lecture Notes - Lecture 9: Confounding, Observational Study, Randomized Experiment
Document Summary
Lecture nine: observational studies and controlled comparisons 02/11/19. We"d like to assess whether there is a causal relationship between x and y. This can be hard: temporal ordering, correlation, causal mechanism, and confounding variables. Without time machines, we can"t prove that x caused y but we can still do our best. Studying a single case or event in great detail. Some of the research designs we will study: Random sampling- everyone in the population has an equal chance of being in the sample. Random assignment- assignment of treatment is randomized. There is usually a trade-off between internal validity and external validity. Field experiments: a study conducted in the real world, has both internal and external validity if conducted properly. Suppose you are studying how a democratic campaign ad affects vote intention. When conducting experiments we check for balance are treatment and control groups similar on obvious confounds. The original idea for experiments was that may different scientists would repeat.