- Atlas
- Mathematics
- Causal Inference
Causal Inference
Causal inference is the part of statistics that tries to answer not just what is associated with what but what would happen if we intervened.
13 frameworks, from 1923 to today · Statistics · Mathematics
- Rubin causal model 1923-Present
- Randomized Experimental Design 1925-Present
- Instrumental Variables 1928-Present
- Regression Discontinuity Design 1960-Present
- Bradford Hill criteria 1965-1990
- Natural experiment 1980-Present
- Propensity Score and Matching Methods 1983-Present
- Causal graph 1986-Present
- G-Methods for Time-Varying Treatments 1986-Present
- Sensitivity Analysis for Unmeasured Confounding 1987-Present
- Partial Identification 1990-Present
- Doubly Robust and Semiparametric Estimation 1994-Present
- Causal Machine Learning 2010-Present
Overview
Causal inference is the part of statistics that tries to answer not just what is associated with what but what would happen if we intervened1. Does a drug save lives, or do healthier people simply happen to take it? Would a minimum wage raise unemployment, or do states that raise it differ from those that don't in ways that matter? These are causal questions, and the central difficulty is that we never observe the same unit both treated and untreated. Causal inference builds the concepts, designs, and estimators that let researchers reason about such counterfactuals from data—and, just as importantly, state honestly what the data cannot tell them.
From Correlation to Cause: The Prehistory
For most of the nineteenth and early twentieth centuries, statistics was wary of causation. The dominant tradition, associated with Karl Pearson and the biometricians, treated correlation as the proper scientific object and causation as metaphysics2. John Stuart Mill's methods of agreement and difference (1843) had given philosophers a logic of causal reasoning, and in 1899 the epidemiologist Wade Hampton Frost and others were already reasoning informally about confounding, but statistics as a discipline had no formal machinery.
The germ of one appeared in 1923, when the geneticist and statistician Sewall Wright published a method of path coefficients—diagrams of arrows between variables whose strengths could be estimated from correlations. Wright used it to disentangle the contributions of heredity and environment to guinea-pig coat color and, later, to interpret corn and livestock breeding data. His work was largely ignored by statisticians for decades, but it was the first serious attempt to represent causal structure in a formal, estimable model2.
Randomization and the Potential Outcomes Idea
The decisive break came from experimental design. In the 1920s at Rothamsted Experimental Station, Ronald Fisher developed the randomized experiment: assign treatments to plots by chance, and the treatment groups become comparable in expectation, so differences in outcomes can be attributed to the treatment. Fisher's The Design of Experiments (1935) codified randomization, blocking, and the analysis of variance. Randomization solved confounding by construction—but it was a design principle, not a theory of what a causal effect is.
That theory arrived from two directions. In 1923, in a Polish-language paper on agricultural experiments, Jerzy Neyman introduced what we now call potential outcomes: each unit has a pair of outcomes, one under treatment and one under control, and the causal effect is their difference32. Neyman used this to justify the analysis of randomized experiments and to derive confidence intervals. In the 1970s, Donald Rubin rediscovered and generalized the idea—now often called the Rubin causal model—extending it to observational studies by treating the treatment assignment itself as something to be modeled. The framework's core object is the counterfactual: the outcome a unit would have had under a treatment it did not receive. Since only one potential outcome per unit is ever observed, causal inference is fundamentally a missing-data problem.
Observational Studies and the Problem of Confounding
Randomization is often impossible—you cannot randomly assign people to smoke4. In 1950, Austin Bradford Hill and Richard Doll published the first strong epidemiological evidence linking cigarette smoking to lung cancer. The finding was attacked on the grounds that smokers might differ from non-smokers in other ways. Hill's response, elaborated in his 1965 essay on "The Environment and Disease: Association or Causation?", was a set of causal criteria—strength, consistency, temporality, biological gradient, plausibility, and others—for judging when an association is likely causal. These criteria were heuristics, not a formal theory, but they shaped epidemiology for a generation1.
The formal theory of observational studies came from William Cochran and, above all, Donald Rubin, whose 1974 paper "Estimating Causal Effects of Treatments in Randomized and Nonrandomized Studies" laid out the potential-outcomes framework for observational data. Rubin's key assumption was strong ignorability: treatment assignment is independent of potential outcomes once we condition on observed covariates. Under that assumption, adjusting for the right variables identifies the causal effect. The framework made precise what "controlling for confounders" had always meant informally.
Matching, Stratification, and the Propensity Score
If adjustment requires conditioning on many covariates, the practical problem is that exact matching on many variables becomes impossible as their number grows. Paul Rosenbaum and Donald Rubin solved this in 1983 with the propensity score: the probability of receiving treatment given observed covariates4. They proved that, under ignorability, conditioning on this single scalar balances the covariates between treated and control groups—so matching or stratifying on the propensity score is enough. The result turned a high-dimensional adjustment problem into a one-dimensional one and spawned a large matching literature (Rosenbaum and Rubin's 1985 paper on bias reduction, and later work by Guido Imbens, Donald Rubin, and others on optimal matching and its properties).
Natural Experiments: IV, Regression Discontinuity, and Difference-in-Differences
A parallel tradition asked what can be learned when treatment is not assigned by a researcher but by circumstance. Instrumental variables originated in economics: Philip Wright's 1928 work on supply and demand for butter and flax used an external variable to identify a demand curve, and the method was formalized in the 1940s by Trygve Haavelmo and the Cowles Commission. An instrument is a variable that affects treatment but affects the outcome only through treatment; it identifies a causal effect for the subpopulation whose treatment it moves. In the 1990s, Joshua Angrist, Guido Imbens, and others connected IV to the potential-outcomes framework, clarifying that it estimates a local average treatment effect—the effect for "compliers" whose treatment responds to the instrument5.
Regression discontinuity exploits a threshold: if a scholarship is awarded above a test score cutoff, students just above and just below the cutoff are comparable, and the jump in outcomes at the cutoff estimates the effect6. The design was introduced by Donald Thistlethwaite and Donald Campbell in 1960 in a study of scholarship awards, and was revived and formalized in the late 1990s and 2000s by Guido Imbens, Thomas Lemieux, Justin McCrary, and others.
Difference-in-differences compares the change in outcomes in a treated group to the change in a control group, differencing out fixed differences between them7. It became a workhorse of applied economics after John Snow's 1855 analysis of a London cholera outbreak is often cited as an early instance, and its modern form was developed in labor economics from the 1980s onward (notably by Orley Ashenfelter and David Card). The key assumption—parallel trends absent treatment—is untestable in general, which has driven a large recent literature on its diagnostics and on staggered adoption designs.
Causal Graphs and the Structural Turn
A third tradition grew from Sewall Wright's path diagrams, revived by computer scientists and philosophers. In the 1980s and 1990s, Judea Pearl developed causal graphical models—directed acyclic graphs (DAGs) in which an arrow means "directly causes." Pearl's Causality (2000) gave the framework its mature form, including the do-calculus, a set of rules for determining whether a causal effect is identifiable from observational data given the graph. The graph makes confounding visible: a backdoor path is a non-causal path between treatment and outcome, and adjusting for a set of variables that blocks all backdoor paths identifies the effect. Peter Spirtes, Clark Glymour, and Richard Scheines's Causation, Prediction, and Search (1993) addressed the inverse problem—learning graph structure from data. The graphical and potential-outcomes traditions were shown to be closely related, and both are now standard.
Sensitivity Analysis
Because ignorability is an assumption, not a fact, a central question is how robust a conclusion is to its violation. Paul Rosenbaum's 1987 work on sensitivity analysis introduced a formal way to ask: how strong would an unmeasured confounder have to be to overturn the finding? The approach, developed with Rosenbaum's Observational Studies (1995), quantifies the departure from randomization needed to explain away an effect, and remains the standard tool for assessing unmeasured confounding.
Doubly Robust and Targeted Estimation
In the 2000s, attention turned to estimators that combine outcome modeling and treatment modeling. Heuristics for combining them had circulated, but James Robins, Andrea Rotnitzky, and Lue Ping Zhao (1994) gave the rigorous theory of doubly robust estimators: consistent if either the outcome model or the treatment model is correctly specified, not necessarily both. This work grew out of Robins's broader development of g-methods—g-computation, inverse probability weighting, and g-estimation—for time-varying treatments, where standard regression fails because treatment and confounders feed back on each other over time. In the 2010s, Mark van der Laan and Sherri Rose developed targeted maximum likelihood estimation (TMLE), a doubly robust, semiparametric-efficient approach built on machine-learning fits of the nuisance models, and Aad van der Vaart, Peter Bickel, and others supplied the underlying semiparametric efficiency theory.
Heterogeneous Effects and Causal Machine Learning
Classical methods estimated an average effect. But a drug may help some patients and harm others, and a policy may work in some regions and not others. The modern literature asks how effects vary—conditional average treatment effects—and how to estimate them with flexible models. Charles Manski's work on partial identification (1990) established what can be learned when assumptions are weak. Susan Athey and Guido Imbens's causal forests (2016) and the broader causal machine learning program adapted tree ensembles, neural networks, and other tools to estimate heterogeneous effects and to select covariates for adjustment, while preserving valid inference. Victor Chernozhukov, Denis Chetverikov, and colleagues' double/debiased machine learning (2018) showed how to use high-dimensional or nonparametric nuisance estimates without contaminating the causal estimate—a direct descendant of the doubly robust tradition.
The Present
Causal inference today is a meeting point of several lineages: the potential-outcomes framework (Rubin, Neyman) supplies the definition of a causal effect; graphical models (Pearl) supply the logic of identification; design-based methods (randomization, IV, regression discontinuity, difference-in-differences) supply credible strategies when randomization is unavailable; and semiparametric and machine-learning theory supplies estimators that are flexible yet valid1. The field's central open questions are the ones its history kept returning to: how to justify ignorability in a given application, how to detect and bound the effects of unmeasured confounding, how to handle time-varying treatments and interference between units, and how to make heterogeneous-effect estimates trustworthy enough to guide individual decisions. The 2021 Nobel Prize in Economic Sciences, awarded to David Card, Joshua Angrist, and Guido Imbens for work on natural experiments and methods for causal inference, marked how far the field had traveled from the days when causation was considered outside the statistician's remit.
References
- Causal inference (introduction, Approaches in epidemiology, General), Wikipedia
- Causal model (History), Wikipedia
- Rubin causal model, Wikipedia
- Propensity score matching, Wikipedia
- Local average treatment effect, Wikipedia
- Regression discontinuity design, Wikipedia
- Difference in differences, Wikipedia
The frameworks of Causal Inference
Each framework is a way of working the field has used: when it held, and the question it answers.
- 1923-Present
- 1925-Present
Randomized Experimental Design
How does random assignment identify a treatment effect?
- 1928-Present
Instrumental Variables
How can an external variable identify a causal effect?
- 1960-Present
Regression Discontinuity Design
When does a cutoff create a usable comparison?
- 1965-1990
- 1980-Present
- 1983-Present
Propensity Score and Matching Methods
How can many covariates be balanced at once?
- 1986-Present
Causal graph
Which variables must be adjusted for to identify an effect?
- 1986-Present
G-Methods for Time-Varying Treatments
How can effects be estimated with time-varying treatments?
- 1987-Present
Sensitivity Analysis for Unmeasured Confounding
How strong must a hidden confounder be to overturn a finding?
- 1990-Present
Partial Identification
What range of effects is consistent with the data?
- 1994-Present
Doubly Robust and Semiparametric Estimation
How can estimators remain valid when models are wrong?
- 2010-Present
Causal Machine Learning
How can flexible algorithms estimate causal effects?
How the ideas connect
Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.
Propensity Score and Matching Methods influenced by Rubin causal model
Rosenbaum and Rubin built the propensity score directly on Rubin's potential-outcomes framework and strong ignorability, proving that conditioning on the score balances covariates under that assumption.
Carried by Paul R. Rosenbaum; Donald B. Rubin.
Instrumental Variables influenced by Rubin causal model
Angrist and Imbens recast instrumental variables in potential-outcomes terms, defining the local average treatment effect for compliers and giving the old econometric method a counterfactual interpretation.
Carried by Joshua D. Angrist; Guido W. Imbens.
G-Methods for Time-Varying Treatments influenced by Rubin causal model
G-computation and the g-formula were defined within the potential-outcomes framework, extending counterfactual reasoning to longitudinal treatments and time-varying confounders.
Carried by James M. Robins.
Sensitivity Analysis for Unmeasured Confounding influenced by Rubin causal model
Rosenbaum's sensitivity analysis formalized how far treatment assignment could depart from the ignorability assumption of the Rubin causal model before an estimated effect would be overturned.
Carried by Paul R. Rosenbaum.
Doubly Robust and Semiparametric Estimation influenced by Rubin causal model
Doubly robust estimators were developed to estimate the average treatment effect under the Rubin model's ignorability assumption while remaining consistent if either the outcome or the treatment model is correct.
Carried by Andrea Rotnitzky; James M. Robins.
Regression Discontinuity Design influenced by Randomized Experimental Design
Thistlethwaite and Campbell developed regression discontinuity in the context of randomized scholarship experiments, treating the cutoff as a quasi-random assignment device in the spirit of experimental design.
Carried by Donald T. Campbell.
Natural experiment influenced by Instrumental Variables
The natural-experiments movement in economics adopted instrumental variables as its central identification strategy, using exogenous variation from policy or circumstance as instruments.
Carried by Joshua D. Angrist; Guido W. Imbens.
Partial Identification influenced by Instrumental Variables
Work on weak instruments and on bounds for treatment effects when instruments are imperfect showed that IV estimates identify ranges rather than points, feeding the partial-identification literature.
Carried by Charles F. Manski.
Doubly Robust and Semiparametric Estimation influenced by Propensity Score and Matching Methods
Semiparametric efficiency theory for the propensity-score-adjusted treatment effect, culminating in doubly robust estimators, took the propensity score as its central nuisance function.
Carried by James M. Robins; Andrea Rotnitzky.
Partial Identification influenced by Causal graph
Pearl's do-calculus and the backdoor criterion made identifiability a graph-theoretic question, and his later work on bounds for effects under unobserved confounding connected graphical models to partial identification.
Carried by Judea Pearl.
Causal Machine Learning influenced by Propensity Score and Matching Methods
Causal machine learning methods for treatment-effect estimation typically use propensity scores or weighting as components of their estimators, inheriting the balancing logic of Rosenbaum and Rubin.
Carried by Susan Athey; Guido W. Imbens.
Beyond the field
Where Causal Inference's ideas came from in other fields, and what they led to there: the strongest 12 of 65 links.
Rubin causal model influenced Causal Inference from Observational Data in Evidence Based Medicine
The Rubin causal model's potential-outcomes formulation was imported into epidemiology as the counterfactual basis for causal inference from observational data. Rubin's framework and its later extension by Robins provided the formal machinery epidemiologists adopted.
Carried by Donald Rubin; James Robins.
Randomized Experimental Design influenced Randomized Controlled Trial in Clinical Epidemiology
Randomized experimental design, developed in agricultural statistics by Fisher and his successors, supplied the randomization and control methodology that the clinical trial adopted. The 1948 streptomycin trial used random allocation and statistical comparison drawn from this tradition.
Carried by Ronald A. Fisher; Austin Bradford Hill.
Randomized Experimental Design influenced Counterfactual Causal Inference in Epidemiology
Randomized experimental design, codified in the 1920s and 1930s, provided the gold-standard comparison that the counterfactual framework formalised as the ideal experiment behind every causal estimand. Randomisation became the benchmark against which observational identification strategies are judged.
Carried by Ronald A. Fisher.
Randomized Experimental Design led to Randomized controlled trial in Experimental Design
The randomized controlled trial grew directly out of Fisher's randomized experimental design, applying random allocation and significance testing to clinical comparisons. The 1948 streptomycin trial used these design principles.
Carried by Austin Bradford Hill; Medical Research Council.
Randomized Experimental Design influenced Experimental and Quasi-Experimental Design in Public Health Program Evaluation
Randomized experimental design, developed in agricultural and clinical statistics, supplied the random assignment and comparison-group logic that experimental and quasi-experimental evaluation adopted.
Carried by Ronald A. Fisher.
Randomized Experimental Design influenced Randomized controlled trial in Epidemiology
Fisher's randomized experimental design, set out in his 1925 Statistical Methods for Research Workers and 1935 Design of Experiments, supplied randomization, replication and the analysis of variance that the 1948 streptomycin trial adopted. The MRC trial's chief statistician Austin Bradford Hill applied Fisher's randomization scheme to clinical comparison.
Carried by Ronald A. Fisher; Statistical Methods for Research Workers (1925); Austin Bradford Hill.
Randomized Experimental Design influenced Design-Based Causal Inference in Econometrics
Randomized experimental design, from Fisher's agricultural trials through clinical trials, is the methodological foundation of design-based causal inference, which treats randomization as the gold standard for identification.
Carried by Ronald Fisher.
Randomized Experimental Design influenced Causal inference and experimental design in Political Science
Randomised experimental design, codified by Fisher in the 1920s and 1930s, is the direct ancestor of the field-experiment turn in political science. The later programme imported Fisher's randomisation, blocking and analysis-of-variance logic into political research.
Carried by Ronald A. Fisher; The Design of Experiments (1935).
Randomized Experimental Design influenced Randomized Controlled Trial in Epidemiological Methods
The randomized controlled trial took its randomization and control-group logic from Fisher's randomized experimental design, applied to patients in the 1948 streptomycin trial.
Carried by Ronald A. Fisher; Austin Bradford Hill.
Instrumental Variables influenced Local Average Treatment Effects and Instrumental Variables in Microeconometrics
Instrumental variables estimation, developed from the 1920s work on supply and demand and later two-stage least squares, supplied the estimator that Imbens and Angrist reinterpreted as identifying a local average treatment effect in 1994.
Carried by Guido Imbens; Joshua Angrist.
Instrumental Variables influenced Colonial Origins and Historical Instrumental Variables in Institutional Development
Colonial-origins research adopted instrumental variables estimation, using settler mortality as an instrument for institutions. The technique descends from the econometric instrumental-variables tradition and was applied here to historical data.
Carried by Daron Acemoglu; Simon Johnson; James A. Robinson.
Regression Discontinuity Design influenced Spatial-Empirical Causal Design in Land Economics
Regression discontinuity design supplied the boundary-based identification strategy that spatial-empirical causal designs use to estimate effects of land-use regulations across jurisdictional borders. The later designs apply RDD to spatial discontinuities.
Carried by Donald Thistlethwaite; Donald Campbell.

Causal Inference is a planet in Nousfarer
Its city has a tower for each of these 13 frameworks. Land, walk in, and climb one concept at a time.
Land on Causal InferenceRelated fields
Fields its ideas are linked to
- Epidemiology 6 links
- Biostatistics 6 links
- Statistical Inference 5 links
- Econometrics 4 links
- Experimental Design 4 links
- Health Capital And Demand For Health 3 links
- Data Science 2 links
- Microeconometrics 2 links
- Evidence Based Medicine 2 links
- Epidemiological Methods 2 links
- Political Science 2 links
- Regression Analysis 2 links
More in Statistics
- Bayesian Statistics
- Experimental Design
- Multivariate Statistics
- Nonparametric Statistics
- Regression Analysis
- Spatial Statistics
- Statistical Computing
- Statistical Inference
- Statistical Learning
- Statistics
- Survey Sampling
- Survival Analysis
- Time Series Analysis
See where Causal Inference sits on the Map of Human Knowledge