Intervention

As explained in Causal Inference for Data Science

An intervention is an externally imposed change to a variable or system intended to reveal or alter causal effects.

An intervention deliberately changes a system to ask what would happen under a different treatment or setting. Unlike prediction, which describes likely outcomes from observed patterns, causal inference asks what can be changed and what effects would follow. section 1.2 section 2.2.4

For kidney-stone treatment, P(Treatment = A) describes patients who happened to receive A, whereas P|do(Treatment = A) describes a world in which everyone receives A. The latter supports comparing treatment effects. section 2.2.4

Formally, intervening on a node changes the data-generating process:

• Fix the treatment variable T to a chosen value, such as A. section 2.2.4
• Remove every arrow pointing into T, so prior causes no longer determine its value. section 2.2.4
• Use the resulting distribution to estimate what universal treatment or randomized assignment would produce. section 2.3

Kidney-stone treatment

Assigning treatment A to every simulated patient produces an approximately 83% recovery rate. Simulating an RCT, with treatment assigned randomly, produces similar estimates: about 83% for A and 75% for B. Randomization breaks the relationship between stone size and treatment, helping separate treatment effects from confounding. section 2.3 section 2.4.2

Figure 2.6 Data-generation process producing the intervened probability distribution P|do(T = A), which models what would happen if we gave treatment A to every patient

Use the intervened graph to compare the observed data-generating process with the scenario where treatment A is given to everyone. section 2.2.4

How Causal Inference for Data Science covers it

How deep it goes

Defines graph surgery formally, contrasts observed and intervened distributions, and demonstrates simulation without developing broader identification theory.

Examples

Uses kidney-stone treatment, universal treatment assignment, intervened graphs, and R or Python simulations of randomized trials.

What it assumes

Assumes basic causal questions and probability notation while building understanding of confounding, graphs, and randomized treatment assignment.

What you can do after

Apply via R or Python by fixing treatment assignments, simulating randomized assignment, and comparing resulting recovery rates.

How it connects

Connects interventions to causal inference, confounding, directed acyclic graphs, randomized controlled trials, adjustment formulas, and average treatment effects.

liveBookmanning.com

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

Also explained in

• 

  liveBookmanning.com
  Causal AIAn intervention is an externally imposed change to a variable or process intended to alter what happens in a causal system.

Compare how 2 books explain intervention →

Related concepts

• actioncontrasts with · An action is a choice or operation selected or performed by an actor that can affect a system or environment or produce an outcome.
• causal hierarchy theorempart of · The causal hierarchy theorem states that observational, interventional, and counterfactual information are generally distinct levels that cannot be inferred from lower-level information without assumptions corresponding to the higher level.
• counterfactual notationused by · Counterfactual notation represents the value or distribution of a variable under a hypothetical intervention or condition.
• action stepimplemented by · An action step is the stage of counterfactual inference that represents a hypothetical condition as an ideal intervention in a modified causal graph.

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.