Social scientists often rely on nonrandom samples, whether due to nonresponse, convenience sampling, or practical design considerations, when studying descriptive or causal relationships. This class focuses on design and analysis techniques that allow researchers to draw generalizable inferences from non representative data. In particular, we will study sampling and randomization design; weighting and modeling methods; and sensitivity analyses with applications to survey analysis and external validity of experiments.
Discussion sections held Wednesdays 2-4pm in SOCS791.
Political Science 231A and 231B or equivalent. Experience with R is assumed.