Methods Lab · Trial Lab
Explore uncertainty in binary trials.
Simulate independent control and treatment groups at a fixed horizon. Inspect single trials and compare up to five synthetic scenarios.
What it explores
Estimate treatment minus control risk with a 95% Newcombe–Wilson interval. Use a two-sided exact Fisher test at 0.05. Examine power or null rejection, coverage, Monte Carlo standard errors and Clopper–Pearson intervals. Sample size is fixed before each simulation.
Assumptions and limits
Newcombe coverage is approximate and can fall below 95%; Fisher can be conservative. The interval and test are distinct procedures. No dropout, censoring, non-adherence, clustered groups, multiple outcomes or clinical protocol validation are included.
Reproduce and export
Trials and summary CSV, metadata, standalone base-R replay script, three plot PNGs and comparison CSV/script. Edited controls remain pending; results, scenario seeds and downloads retain their executed snapshot.
Synthetic simulation in your browser
These apps generate synthetic participants only. Calculations run in browser memory; no R installation is required. Initial loading downloads the R runtime and packages. Language changes preserve controls and executed results.
For teaching and methodological exploration. Results do not establish clinical efficacy or recommend treatments.