How is the national unemployment rate determined? How about the percentage of people living in poverty, the quantity of crops produced in agricultural counties, or the educational achievement levels of the nation's fourth graders? The answer to these questions is: through surveys. In this course, we investigate the mathematical and statistical underpinnings of survey design and analysis. We examine probability sampling designs that can yield valid estimates of unknown population quantities, including stratified, clustered, and multi-stage designs. For each of these designs, we derive estimators of population quantities and measures of uncertainty of these estimators. We also discuss optimal sample allocation, methods for handling nonresponse, resampling techniques for variance estimation, and other advanced topics. Emphasis is on understanding surveys through real data sets in economics, education, health care policy, and the sciences. Performance is evaluated primarily through homeworks and projects. Prerequisite: Mathematics 243 or permission of the instructor.

REITER