scientific-causal-inference

Use a modular causal inference pipeline with methods including instrumental variables, regression discontinuity design, and difference-in-differences. It requires a dataset with treatment, outcome, and covariates to estimate effects and runs sensitivity analyses for robust inference.

Select confounders for causal inference by building Directed Acyclic Graphs (DAGs) to map variable relationships. This skill integrates DAG-based confounder selection directly into the pipeline to identify and adjust for variables that bias treatment effect estimates.

Yes, you can use propensity score matching and inverse probability weighting in the same analysis pipeline. Both methods are supported as distinct modules to balance treated and control groups, allowing you to compare treatment effect estimates across different adjustment techniques.

Yes, the causal inference pipeline supports both regression discontinuity design and difference-in-differences. These methods are included alongside instrumental variables and propensity score matching to handle various observational study scenarios in economics and epidemiology.

Validate the robustness of causal inference estimates by running optional sensitivity analyses provided within the pipeline. Additionally, the workflow supports literature search via OpenAlex to contextualize your findings and ensure reproducible inference.

To run instrumental variables or regression discontinuity analysis, provide a dataset containing treatment, outcome, and covariate variables. The pipeline processes this observational data to estimate causal effects while requiring a reproducible workflow structure.