Discussion of "Causal inference using invariant prediction: identification and confidence intervals" by Peters, B\"uhlmann and Meinshausen

TL;DR

This paper discusses the methods of invariant prediction for causal inference, emphasizing the importance of identifying causal structures and constructing confidence intervals, and provides insights into their theoretical and practical implications.

Contribution

It offers a critical discussion on the invariant prediction approach for causal inference, highlighting its strengths, limitations, and potential directions for future research.

Findings

Invariant prediction effectively identifies causal relationships.

Confidence intervals can be reliably constructed using this method.

The approach has promising applications in various statistical fields.

Abstract

Contribution to the discussion of the paper "Causal inference using invariant prediction: identification and confidence intervals" by Peters, B\"uhlmann and Meinshausen, to appear in the Journal of the Royal Statistical Society, Series B.