Unexpected Event during Survey Design Framework
The Unexpected Event during Survey Design (UESD) framework is used for causal analysis of coincidental events.
Description
Unexpected events such as natural disasters and violence can be difficult to study. Between the difficulty in predicting where and when an event will occur (so as to collect pre-treatment data) and the impossibility of random treatment assignment, there are substantial challenges to experimental design.
Over time, there are coincidences of an active survey that collects relevant measurements and an unexpected event. Experience of the event can be used as an experimental treatment. This can be exploited for a causal framework.
- The event must be unexpected; anticipation of an event is expected to lead to smoother behavioral adjustments.
- Excludable: the estimated effect can be confounded by collateral or simultaneous events.
- Ignorability: the control and treatment cohorts must be comparable.
- Noncompliance: the treatment cohort must be uniformly treated.
History
This framework was formalized by Muñoz, Falcó-Gimeno, and Hernández.