How deeply does Causal Inference for Data Science cover FWL strategy?

Demonstrates residual construction and numerical equivalence in a simulated linear setting, without extending beyond that example.

FWL strategy

As explained in Causal Inference for Data Science

The FWL strategy estimates a regression coefficient by removing confounder-predicted components from both the treatment and outcome and regressing the resulting residuals on each other.

The FWL strategy estimates a treatment effect by removing the confounder-predicted components from both treatment and outcome, then regressing the two residuals on each other. section 8.1.1

A practical workflow is:

Predict the outcome from the confounder and retain its residuals. section 8.1.1
Predict the treatment from the confounder and retain its residuals. section 8.1.1
Regress outcome residuals on treatment residuals; the coefficient matches linear regression controlling for the confounder, apart from rounding. section 8.1.1

Synthetic treatment-effect data

With a simulated treatment effect of 5, the adjusted regression estimates 4.9473069, while the FWL residual regression estimates 4.947307. section 8.1.1

How Causal Inference for Data Science covers it

Cover of Causal Inference for Data Science

liveBook manning.com

Read Causal Inference for Data Science on liveBook Get the book at manning.com

Related concepts

residual erroruses · Residual error is the difference between an observed value and its corresponding value predicted or fitted by a model.
linear regressionalternative to · Linear regression is a statistical modeling method that estimates a continuous response as a linear combination of predictor variables and fitted coefficients.
strategykind of · A strategy is a planned approach for achieving objectives through coordinated choices, priorities, and actions.
confounderaddresses · A confounder is a variable associated with both an explanatory or treatment variable and an outcome, whose influence can distort their apparent relationship or estimated causal effect.

This explainer was generated by Manning's knowledge graph from the text of Causal Inference for Data Science. The section links point to the parts of the book it draws on.