STAT 378 - Quasi-experimental design

FC NSMA QFR

4 credits
Determining causal relationships (for example, the impact of a vaccine on preventing illness) is an important goal in statistical analysis, especially when building models. However, the conditions for causality are often prohibitively difficult to obtain, either due to ethical concerns, cost, or the infeasibility of constructing a controlled experimental environment. This course provides an overview of the diverse set of analytic approaches that fall under “quasi-experimental” designs, which allow for near-causal inferences using observational, rather than experimental data. Topics covered will include interrupted time-series, differences-in-differences, and spatial discontinuity models. Students will both review contemporary research that employ quasi-experimental designs, as well as conduct their own quasi-experimental analyses using publicly available data and open source computational tools.

Prerequisites: ECON 255 or PSYC 300 or any 200 level STAT course.

Undergraduate Research Intensive