Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

Experimental Design

Plan studies before data collection — pick the right design, randomize, block, and generate reproducible DOE layouts.

Data & AnalyticsAdvanced★ 33,030⑂ 3,248AI score 8/10Last updated: Aug 9, 2026

What it does

Covers the decisions made before any data exists, because no analysis can rescue a confounded or pseudoreplicated study.

  • Design selection via a decision tree: completely randomized, randomized block, crossover/repeated measures, split-plot, Latin square, cluster-randomized
  • Multi-factor experiments: full 2^k factorial, fractional factorial, Plackett-Burman screening, response-surface (central composite, Box-Behnken), Latin hypercube
  • randomization.py — seeded allocation schedules (simple, permuted block, stratified block, cluster) plus arm-balance checks, exportable to CSV
  • doe_designs.py — DOE matrices returned in real factor units (°C, mM, pH) with run order randomized by default to defeat drift
  • A checklist of eight structural killers: pseudoreplication, batch confounding, plate edge effects, ignored aliasing, missing controls, and more

Sample size and power are delegated to the statistical-power skill; analysis of collected data goes to statistical-analysis.

Who it's for

  • Lab and clinical researchers planning trials, animal studies, or omics runs
  • R&D / process and quality engineers using DOE for optimization
  • Data scientists designing cluster-randomized field experiments or A/B tests
  • Grad students and reviewers auditing whether a design can answer the question

Example uses

  1. "Assign 60 mice to drug vs. control, balanced by cage" → produces a seeded stratified block schedule and saves allocation_schedule.csv.
  2. "I have 7 process factors and limited runs" → builds a Plackett-Burman screening design in real units and explains the alias structure before you declare a factor inert.
  3. "How should I lay samples out on a 96-well plate?" → recommends randomized/blocked positions to avoid edge and batch effects, plus randomized processing order.

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/experimental-design folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/experimental-design/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/K-Dense-AI/scientific-agent-skills.git /tmp/scientific-agent-skills && mkdir -p ~/.claude/skills && cp -r /tmp/scientific-agent-skills/skills/experimental-design ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Confirm Python 3.10 or newer: python3 --version.
  2. Clone the repository: git clone https://github.com/K-Dense-AI/scientific-agent-skills.git
  3. Copy the skill into your Claude Code skills folder: mkdir -p ~/.claude/skills && cp -r scientific-agent-skills/skills/experimental-design ~/.claude/skills/
  4. Install dependencies: uv pip install "numpy>=1.26" "pandas>=2.0" pyDOE3 (use pip install if you don't have uv).
  5. Restart Claude Code and ask something like "how should I set up this experiment to avoid confounding?" — the skill triggers automatically.
  6. Optionally install the companion statistical-power and statistical-analysis skills the same way to cover sample size and post-collection analysis.
View source on GitHub ↗License: MIT