How can latent variable models be used in causal inference to address unobservable confounders?
Instruction: Describe the application of latent variable models in controlling for unobserved confounding in causal analysis.
Context: Candidates should explain how latent variable models can infer unobserved variables and their role in addressing omitted variable bias.
Updated
Official answer available
Read the opening below, then unlock the full answer and practical guidance.
Latent variable models can represent an unobserved construct using measured indicators, but a fitted hidden factor does not automatically remove confounding. Different causal structures can produce the same observed correlations while implying different treatment effects...