Rejoinder to the discussion on ‘Causal inference with misspecified network interference structure’

Key Points

• The aim is to clarify the importance of network specification in causal inference under different probabilistic regimes.
• Analyzed feedback from discussants on network specifications and interference.
• Revisited the coupled design-based approach with exposure mapping.
• Addressed proposed modifications and extensions suggested by the discussants.
• Emphasized the critical role of accurate network specification in causal inference.
• Identified the advantages and risks of using multiple networks in analysis.