Experiment Design

What it does

Experiment Design helps the agent guide users through planning rigorous empirical studies. It covers selecting the right research design (RCTs, quasi-experiments, factorial designs, A/B tests, surveys, cohort studies), performing power analysis and sample-size calculations for t-tests, chi-square tests, and correlations, and drafting a complete study protocol. The skill includes practical templates and code snippets for sample-size calculations so the agent can provide concrete numeric examples.

When to use it

Use this skill when a user asks to design an experiment, calculate sample size or power, create randomization or blinding plans, draft a study protocol, or plan A/B tests and factorial experiments. It is appropriate for researchers, product teams running experiments, and students learning study design.

What's included

• Scripts: inline Python examples for power/sample-size calculations (ttest, chi-square, correlation)
• References: protocol template and practical tips for pre-registration and bias mitigation
• Instructions: step-by-step guidance for choosing design, mitigating bias, planning randomization/blinding, and preparing analysis plans.

Compatible agents

Best with agents that can run or present code snippets (Codex/GitHub Copilot-style, Claude Code, Gemini CLI) and agents that have access to research assets and docs.