What problem does it solve?
Helps researchers and product teams design valid experiments, choose appropriate statistical tests, calculate sample size and power, and interpret results so decisions are based on sound evidence rather than misapplied statistics or p-hacking.
Core Features & Use Cases
- Hypothesis formulation & test selection: Define null and alternative hypotheses, choose one-tailed vs two-tailed, and recommend parametric or non-parametric tests (t-test, chi-square, ANOVA, z-test for proportions).
- Power analysis & sample size: Perform a priori and post-hoc power calculations to determine required sample sizes and trade-offs between detectable effect and cost.
- Execution & interpretation: Calculate p-values, confidence intervals, effect sizes, provide APA-style reporting, and offer Bayesian and causal inference options for advanced analyses.
- Use Case: Plan and analyze an A/B test for a website redesign, compute required visitors per variant, run the selected statistical test on collected counts, and produce an actionable recommendation.
Quick Start
Design a one-tailed A/B test to detect a 20% relative uplift from a 3.2% baseline with 80% power and return the required sample size, test choice, and analysis plan.
