Extended conditional independence and applications in causal inference (Q682301)

The authors combine the concepts of stochastic conditional independence and variation conditional independence (involving nonstochastic variables) and introduce the concept of extended conditional independence that satisfies their core properties. It is used to ``provide a rigorous basis for a wide range of statistical concepts including ancillarity, sufficiency'' and decision theoretic framework for causal inference. Supplementary material containing proofs of many of the results in the paper are given online.