Thomas Lumley • 7/2/2013

Problems with faithfulness and the causal Markov property (I)

This technical article analyzes foundational problems in causal inference, specifically the causal Markov property and the faithfulness assumption. It details how measurement error—broadly defined to include imperfect proxies and single-measurement averages—can break the conditional independence relationships these properties require, complicating the accurate representation of causal structures.

0 comments

#statistics #Causal Inference #Measurement Error