Experimental Design
- 17 installs
- 869 repo stars
- Updated June 8, 2026
- beita6969/scienceclaw
experimental-design is a Claude skill that plans rigorous scientific experiments with power analysis, controls, randomization, and blinding.
About
This skill designs rigorous scientific experiments, covering study-design selection, power analysis and sample-size calculation, variable control, randomization, and blinding. A developer or researcher uses it when planning an experiment or clinical trial. It explicitly does not run experiments or analyze collected data.
- Calculates sample sizes via power analysis with statsmodels
- Guides study design selection (RCT, factorial, crossover, quasi-experimental)
- Provides a reproducibility checklist and randomization/blinding strategies
Experimental Design by the numbers
- 17 all-time installs (skills.sh)
- Ranked #1,286 of 2,065 Data Science & ML skills by installs in the Skillselion catalog
- Data as of Aug 2, 2026 (Skillselion catalog sync)
experimental-design capabilities & compatibility
Free; uses open Python statistical libraries.
- Capabilities
- experiment design · economics analysis · exploratory data analysis
- Works with
- confluence
- Use cases
- research · data analysis
- Pricing
- Free
What experimental-design says it does
Design rigorous scientific experiments with power analysis and controls.
Required sample size per group
npx skills add https://github.com/beita6969/scienceclaw --skill experimental-designAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 17 |
|---|---|
| repo stars | ★ 869 |
| Last updated | June 8, 2026 |
| Repository | beita6969/scienceclaw ↗ |
What it does
Use it to plan rigorous experiments: sample-size calculation, controls, randomization, and blinding, before collecting data.
Who is it for?
Planning an experiment and calculating the sample size before any data is collected.
Skip if: Running experiments or analyzing collected data.
When should I use this skill?
The user needs to plan an experiment, calculate sample sizes, or set up controls.
What you get
Produces a valid experimental design with a justified sample size and specified controls, randomization, and blinding.
- Experimental design document
- Sample-size justification
- Reproducibility checklist
By the numbers
- 12-item reproducibility checklist
- 6-design selection table
Files
Experimental Design Skill
Design rigorous, reproducible experiments across scientific disciplines.
When to Use
- "Design an experiment to test..."
- "How many samples do I need?"
- "What controls should I include?"
- "Help me plan a clinical trial"
- "Is this experimental design valid?"
- Power analysis and sample size calculation
When NOT to Use
- Running the actual experiment (use code-execution)
- Analyzing collected data (use scipy-analysis + statsmodels-stats)
- Writing up results (use paper-writing)
- Literature review (use literature-search)
Design Components
1. Research Question and Hypotheses
- State clear, testable research question
- Formulate H0 and H1 (see hypothesis-gen skill)
- Define primary and secondary outcomes
2. Study Design Selection
| Design | When to Use | Strengths | Weaknesses |
|---|---|---|---|
| RCT | Causal inference needed | Gold standard causality | Expensive, ethical limits |
| Factorial | Multiple factors | Tests interactions | Complex analysis |
| Crossover | Within-subject comparison | Reduced variability | Carryover effects |
| Quasi-experimental | Randomization impossible | Practical feasibility | Weaker causality |
| Observational (cohort) | Long-term outcomes | Natural setting | Confounding |
| Case-control | Rare outcomes | Efficient for rare events | Recall bias |
3. Power Analysis
# Sample size calculation template (using scipy/statsmodels)
from statsmodels.stats.power import TTestIndPower
analysis = TTestIndPower()
n = analysis.solve_power(
effect_size=0.5, # Cohen's d (small=0.2, medium=0.5, large=0.8)
alpha=0.05, # Significance level
power=0.80, # Statistical power (commonly 0.80 or 0.90)
ratio=1.0, # Ratio of group sizes (n2/n1)
alternative='two-sided'
)
print(f"Required sample size per group: {int(n) + 1}")Key parameters:
- Effect size: Expected magnitude of difference
- Alpha: Type I error rate (usually 0.05)
- Power: 1 - Type II error rate (usually 0.80-0.95)
- Attrition: Add 10-20% for expected dropout
4. Variable Control
- Independent variables: What you manipulate
- Dependent variables: What you measure
- Confounding variables: What could bias results
- Control strategies: Randomization, blocking, matching, blinding
5. Randomization
- Simple randomization (coin flip)
- Block randomization (balanced groups)
- Stratified randomization (balance key covariates)
- Cluster randomization (group-level assignment)
6. Blinding
- Single-blind: Participants unaware of assignment
- Double-blind: Participants and researchers unaware
- Triple-blind: Including data analysts
Reproducibility Checklist
- [ ] Protocol pre-registered (OSF, ClinicalTrials.gov, PROSPERO)
- [ ] All materials/reagents specified with catalog numbers
- [ ] Detailed step-by-step procedure written
- [ ] Statistical analysis plan pre-specified
- [ ] Data management plan documented
- [ ] Raw data sharing plan established
- [ ] Code availability ensured
- [ ] Sample size justified with power analysis
- [ ] Randomization method specified
- [ ] Blinding procedures documented
- [ ] Inclusion/exclusion criteria defined
- [ ] Primary endpoint pre-specified
Output Format
## Experimental Design: [Title]
**Research Question**: [Clear question]
**Design Type**: [RCT/Factorial/etc.]
### Participants/Samples
- Population: [target population]
- Inclusion: [criteria]
- Exclusion: [criteria]
- Sample Size: N=[total] ([n] per group) — Power=[X], alpha=[X], effect=[X]
### Groups
- Experimental: [treatment description]
- Control: [control description]
- Blinding: [single/double/triple/none]
### Variables
- IV: [variables]
- DV: [primary + secondary outcomes]
- Controls: [confounds and how addressed]
### Procedure
1. [Step-by-step protocol]
### Analysis Plan
- Primary: [statistical test]
- Secondary: [additional analyses]
- Multiple comparison correction: [method]
### Timeline
- [Phase 1]: [duration]
- [Phase 2]: [duration]
### Ethics
- IRB/IACUC requirements: [details]
- Consent procedure: [details]Related skills
Microsoft FoundryDeploy, evaluate, and continuously improve Microsoft Foundry agents from a single agent interface.490k1.3k
Ai Research ReproductionOrchestrate trustworthy, auditable reproduction of deep learning repositories directly from their READMEs.176k511Run TrainSafely execute selected deep learning training commands with standardized evidence capture.176k511Explore RunSafely run isolated exploratory experiments with clear recording and conservative selection before committing changes.176k511Paper Context ResolverFetch precise reproduction-critical details like dataset splits, preprocessing steps, or evaluation protocols from the original academic paper when the repo README leav141k511Repo Intake And PlanScan unfamiliar AI research repositories and receive a minimal, trustworthy reproduction target before investing significant time.140k511FAQ
How does it calculate sample size?
Via power analysis using statsmodels' TTestIndPower, taking effect size, alpha, power, and group ratio.
What is it not for?
Running the actual experiment or analyzing collected data, which are handled by other skills.
Data Science & MLanalytics