Does the interpretation on the estimated coefficient on the treatment change if you use a proxy variable to control for a confounding factor?
A) No, the estimated coefficient is a consistent estimate of the causal effect.
B) No, but instead of getting causal estimate you're only estimating partial correlations.
C) Yes, the coefficient on the treatment is biased.
D) Yes, the coefficient on the treatment suffers from the simultaneity bias.

Question

Quizplus · Accepted Answer

Using a proxy variable to control for a confounding factor will not change the interpretation of the estimated coefficient on the treatment. The estimated coefficient remains a consistent estimate of the causal effect.

Does the Interpretation on the Estimated Coefficient on the Treatment