Flames@Ugent seminar: Causal inference from observational data: Instrumental variables

Causal inference from observational data: Instrumental variables

This seminar is about 'Causal Inference from Observational Data: Instrumental Variables':
We have previously presented methods to mitigate the impact of measured confounding on the estimation of a causal treatment effects such as Matching and Weighting using propensity scores. A main critic of these methods is that they only control for measured confounders and do not control for unmeasured confounders. Instrumental variables analysis is a method for controlling for unmeasured confounding. This type of analysis requires the measurement of a valid instrumental variable, which is a variable that (i) is independent of the unmeasured confounding; (ii) affects the treatment; and (iii) affects the outcome only indirectly through its effect on the treatment. In summary, the IV method seeks to find a randomized experiment embedded in an observational study and uses this to estimate the causal treatment effect.
This tutorial discusses the types of causal effects that can be estimated by instrumental variables analysis; the assumptions; methods of estimation and sources of instrumental variables.
We will wrap up with an example that uses the ivpack package in the R software.