10 Interference

10.1 Interference intro

Nov 18 Slides.

Today we will introduce the idea of network interference. After class you should be able to:

• Explain what interference is and why it causes problems for standard estimators
• Use exposure mappings to model structure in interference

10.2 Lab

Nov 19

Lab today will be a chance for you to get started on the course project and ask questions with the TAs.

10.3 Discussion on interference

Nov 20 Slides.

Today we will discuss two approaches for estimating the global average treatment effect under interference given data from a randomized control trial. After class you should be able to

• Use inverse probability weighting (IPW) to estimate global average treatment effect
• Explain the implications of the choice of randomized design on the variance of the estimator
• Use model based regression to estimate global average treatment effect