1. Upload
2. Model
3. Method
4. Results
Upload Your Data
Start by uploading a CSV file with your observational or experimental data
Upload a CSV file to get started
Treatment variable should be binary (0/1)
Define Causal Model
Specify your treatment, outcome, and any confounding variables
Variable Selection
Choose Method
Select an estimation strategy for the average treatment effect
Difference in Means
Simple comparison of average outcomes between treated and control groups. Best for randomized experiments.
Regression Adjustment (OLS)
Linear regression controlling for covariates. Assumes additive, linear confounding effects.
Inverse Probability Weighting
Reweights observations by propensity scores to balance treatment groups. Requires covariates.
Results
Estimated treatment effect and statistical inference
Analysis Complete
Average Treatment Effect
—
Interpretation
Key Identifying Assumptions
- Unconfoundedness: All variables affecting both treatment and outcome are observed
- SUTVA: No interference between units; stable treatment values
- Positivity: Every unit has positive probability of receiving treatment