SURV722: Causal Inference from Randomized and Observational Data


Survey methodology research sets out to answer questions regarding the effects of particular design decisions : do self-administered modes increase the reports of socially undesirable behavior? Does the use of incentives increase response rates? Does dependent interviewing decrease seam-effects? Do the employment rate estimates change with adding additional response categories like “maternity leave”? This course teaches the fundamental concepts behind the estimation of causal effects, including potential obstacles to causal inference, faulty measurement, spuriousness, specification errors, and other problems that can lead to inappropriate causal inferences. We will discuss the benefits and the difficulties of randomization in survey research in the first half of the class. The focus of the second half is on the design of observational studies and inferences from prediction. Real-world examples will be discussed, with an emphasis on examples from survey methodology. Students will come away with an understanding of how to estimate causal effects in both randomized and observational settings, with a particular focus on careful design of both types of studies.

Course objectives: 

By the end of the course, students will be able to…

  1. understand the fundamental concepts behind the estimation of causal effects;
  2. understand and evaluate studies that produce point estimates and assert causality;
  3. select an appropriate design to estimate causal effects;
  4. design randomized experiments;
  5. understand how to estimate causal effects in observational settings;
  6. understand application of matching methods for causal inference.

Grading will be based on the folowing elements:

  • 7 Quizzes (10% of grade)
  • Class participation (10% of grade)
  • 7 Homework assignments (60% of grade)
  • Final exam (20% of grade)
  • Participation in the course Surv400 “Fundamentals of Survey and Data Science”
  • Knowledge of basic statistics including regression analysis
  • The students should be familiar with the statistical software R or STATA
Course syllabus: 

Course Dates


Summer Term (June – August)