Qm
Statistics

Over-adjustment: controlling for a proxy of the treatment

over-adjustmentbad-controlscausal-inference

You estimate the effect of a value signal on returns. To "be safe," you also control for book-to-price ratio, which is essentially the same construct your value signal is built from: the two are nearly the same information.

Explain why adding this near-duplicate control drives the value coefficient toward zero, and why that is a bias, not a discovery.

Your answer

Solving needs a free account

Answers, streaks and solutions unlock when you are signed in. Reading the question and the hint stays free.

Create a free accountSign in

Discussion

Sign in to join the discussion · reading is open to everyone

💡 Discussion rules

  1. No full solutions here. Hints and approaches only.
  2. Complexity, edge cases and intuition are the point.
  3. Interview experiences are welcome. Respect your NDAs.

Loading discussion…

Learn the concepts

The theory behind this question.

ATE, ATT and Other Treatment Effect EstimandsBad Controls and Post-Treatment BiasBerkson's ParadoxEstimating Conditional Average Treatment Effects With MLCausal Discovery from Observational Data

Related questions

The other kind of bias: controlling for a mediatorConditioning on a collider opens a fake pathWhich of three candidate controls belongs in the regression?Total effect versus direct effect when you control a mediatorEstablishing causation: the index-inclusion effect
All questions →