[SoftwareX 2026] FastRerandomize: Rerandomization Using Accelerated Computing
Monte Carlo audit of naive, cluster-robust, and randomization-based experiment inference
EDI (Experimental Design and Inference): R & Python packages for fixed and sequential randomized experimental designs (blocking, optimal, matching-on-the-fly) with matched exact, asymptotic, and randomization inference for continuous, count, proportion, survival, and ordinal responses. C++ core. Not Electronic Data Interchange.
Validation swaps
R package implementing Lee & Wooldridge (2025, 2026) rolling difference-in-differences estimator for panel data with staggered adoption, multiple estimators (RA/IPW/IPWRA/PSM), and exact small-sample inference.(Public Preview)(Disclaimer: CURRENTLY WIP)
Exact randomization inference with uncertain record identities
Design-aware reanalysis of a household-randomized voter-turnout field experiment
Academic analysis of randomized-experiment data using treatment-effect estimation and randomization inference.
Real-data evidence on intraday volatility periodicity across cryptocurrency assets, periods, and exchanges.
Replication code for "Can Fintech Build Trade Resilience Under Tariff Shocks?" — three-way FE PPML gravity on an East Asian trade panel, with N_eff (a treatment-concentration diagnostic), randomization inference, cluster jackknife, synthetic control, specification curve and MDE-based design requirements.
Replication package for spatial spillover analysis of Azerbaijan (RI placebo tests, spatial econometrics)
Can public card data see the spending drop from ending pandemic unemployment benefits early? Staggered DiD across 48 states, 140 border-county pairs, randomization inference, power, and an audit folder.
Out-of-sample evaluation of geopolitical risk as a carbon-price predictor, with a dimension-matched placebo design. Replication code for "Geopolitical Risk and Carbon Price Predictability: An Effect That Disappears."
Do LLM-judge biases compose? A label-free, scale-aware test (SSIT) for position×length interaction in LLM-as-a-judge pipelines, with proofs, tie-coding and curvature diagnostics, 7-judge/6-family data, and the TMLR manuscript.
R tools for the Solomon four-group design: historical analyses, modern regression, randomization inference, maximum likelihood, SEM, power and sample-size planning, and Solomon-specific figures, each traced to its published source.
To associate your repository with the randomization-inference topic, visit your repo's landing page and select "manage topics."