DataTerminology.comNEW
The probability of receiving treatment given observed covariates, used to balance groups in observational studies.
DESCRIPTION
The propensity score is the conditional probability of treatment assignment given covariates. Matching, weighting, or subclassification on the score aims to balance observed covariates, approximating randomized experiments. It supports estimation of treatment effects like ATE or ATT when ignorability holds. Diagnostics for overlap and balance are critical, and sensitivity analyses assess robustness to unmeasured confounding.
ALTERNATIVE DEFINITIONS
Wikipedia
In statistics, the propensity score is the probability of a unit being assigned to a particular treatment given a set of observed covariates.
Provides the canonical probability-based definition.
NIH PMC (Austin, 2011)
The propensity score is the probability of treatment assignment conditional on observed baseline covariates.
Highlights conditioning on observed baseline covariates and its role in observational studies.
RELATED TERMS
Matching Inverse Probability Weighting Doubly Robust Estimation Confounding Treatment Effect Average Treatment Effect Causal Inference Covariate Balance Overlap Sensitivity Analysis
COMPETING TERMS
Instrumental Variable Randomized Controlled Trial
RELATIONS
- Broader: Score
KNOWLEDGE AREAS
TARGET AUDIENCE
Data Scientist Information Analyst AI Engineer
EXTERNAL SOURCES
- Wikipedia - Propensity score
- NIH PMC - Introductory review of propensity methods
- Penn State STAT 508 - Course lesson on propensity scores
- Scott Cunningham - Matching and subclassification chapter
TOP QUESTIONS
What is a Propensity Score?
It’s the probability of treatment given covariates, used to balance observed differences between treated and control groups.
Category: Definition
How are Propensity Scores estimated?
Commonly via logistic regression or ML classifiers using pre-treatment covariates only.
Category: Requirements
What are ways to use Propensity Scores?
Matching, inverse probability weighting, stratification, or covariate adjustment using the score.
Category: Application
What is a common mistake with Propensity Scores?
Including post-treatment variables or instruments as predictors, which biases estimates.
Category: Mistakes
How to diagnose balance with Propensity Scores?
Check standardized mean differences and overlap; visualize propensity distributions for treated and control groups.
Category: Practicality
What assumptions are required?
Conditional ignorability and overlap (positivity) are key for unbiased effect estimation.
Category: Requirements
When is Propensity Score not advisable?
When important confounders are unobserved or overlap is poor; alternative designs may be needed.
Category: Risks
Is ML better than logistic regression for Propensity Scores?
ML can improve balance but risks overfitting; prioritize covariate balance over predictive accuracy.
Category: Opinion
How does Propensity Score relate to Randomized Controlled Trials?
It attempts to emulate randomization by balancing observed covariates in observational data.
Category: Example