What changed after the policy?
Difference-in-differencesΔYᵀ − ΔYᶜ
Read the policy effect
- Compare groups
- Check trends
- Estimate change
THE CAUSAL INFERENCE WORKSPACE
Good questions.
Credible answers.
Find a method for your question. Follow the analysis, inspect its evidence, and discuss every step.
THE METHOD STUDIO
01 / 03
What happens at the cutoff?
Regression discontinuityμ₊(c) − μ₋(c)
Compare the two limits
- Find the cutoff
- Compare nearby
- Estimate the jump
Who benefits most?
Causal forestτ(x) = E[Y₁ − Y₀ | X=x]
Inspect effect variation
- Describe features
- Learn effects
- Check heterogeneity
A world of methods.
COMMUNITY ATLAS- Method guides
- 68
- Research fields
- 10
THE OPEN NOTEBOOK
Explore the methods.
68 workflows/10 fields of inquiry
Design & analyze randomized experiments (experiment)
The idea
Randomize treatment (complete, blocked, cluster) and estimate average effects with design-based variance — including cluster-randomized trials.
Design-based estimators done fast (estimatr)
The idea
Lin's covariate-adjusted estimator and Neyman/HC2 robust standard errors for randomized experiments — one fast function, design-based inference.
Random assignment by design (randomizr)
The idea
Reproducible random assignment — simple, complete, block, cluster, stratified — with the exact assignment probabilities design-based inference needs.
Design & diagnose a randomized experiment (DeclareDesign)
The idea
Specify a study as model–inquiry–data–answer, simulate it, and read its diagnosands — bias, power, coverage — before you run it.
Randomization inference, packaged (ri2)
The idea
Exact Fisher randomization tests and sharp-null confidence intervals for any randomization scheme — the packaged version of the design-based test.
