An effect changes sign under a small plausible unobserved confounder. What conclusion is warranted?
StatisticsLevel: AdvancedCausal stress testing
An effect changes sign under a small plausible unobserved confounder. What conclusion is warranted?
Hints
- Fragility is conditional.
- Do not infer existence from a simulation.
Answer
This answer is part of mentoring
Advanced questions come with a full breakdown and a mock interview with a working quant researcher. The first call is free.
Source and licence: DoWhy — sensitivity to unobserved confounding · MIT · answer: Quantoros editorial