Potential outcomes framework
The potential outcomes framework defines causal effects by comparing outcomes that would result for a unit under alternative treatments or interventions.
Causal Inference for Data Science
A potential outcome is the outcome that would result for a unit under a particular treatment or intervention level.
- How it is used
- Frames treatment-effect questions by comparing outcomes under alternative intervention levels rather than observed group differences.
- How deep it goes
- Defines individual and population notation, observability, DAG equivalence, and adjustment assumptions, while deferring advanced DAG counterfactuals.
- Examples
- Uses patient recovery, chatbot access, movie-release time series, and movie-saga groups with synthetic data.
- What you can do after
- Evaluate/understand which outcomes are factual or counterfactual and express intervention effects with potential-outcome notation.
Read how Causal Inference for Data Science explains potential outcomes framework →
Causal AI
A potential outcome is the outcome that would result for a unit under a specified treatment or condition, whether or not that outcome is observed.
- How it is used
- Frames causal effect inference and counterfactual decision questions by comparing observed outcomes with outcomes under alternative actions.
- How deep it goes
- Defines possible worlds and potential outcomes, then connects them to notation, probability, ignorability, and backdoor identification.
- Examples
- Uses partner choice, aspirin and headache relief, classifier feature changes, and intervention graphs to make alternative outcomes concrete.
- What you can do after
- Evaluate counterfactual claims, identify actionable changes to predictions, and reason about causal effects under explicit assumptions.
Related concepts
- directed acyclic graphequivalent to · A directed acyclic graph is a graph with directed edges and no directed cycles.
- do operatorequivalent to · The do operator denotes an intervention that sets a variable to a specified value in a causal graph while replacing its usual assignment mechanism.
- average treatment effect estimationproduces · Average treatment effect estimation quantifies the population-average difference in outcomes that would result from two alternative treatments or interventions.
- continuous functionconfigures · A function whose output changes arbitrarily little when its input changes arbitrarily little.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.