statistical-analysis

Perform pairwise t-tests, Cohen's d, and Bonferroni corrections on SWARM experiment CSV data.

39|4|Updated Feb 3, 2026
One-click install
npx skills add https://github.com/swarm-ai-safety/swarm --skill statistical-analysis-swarm-ai-safety
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: statistical-analysis
Source: https://github.com/swarm-ai-safety/swarm/tree/main/bench/skills/statistical-analysis
Command: npx skills add https://github.com/swarm-ai-safety/swarm --skill statistical-analysis-swarm-ai-safety

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires scipy, pandas, numpy, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates rigorous statistical analysis of SWARM experiment data, making it easy to determine the significance and impact of different experimental parameters.

Core Features & Use Cases

  • Hypothesis Testing: Performs pairwise Welch's t-tests to compare groups.
  • Effect Size Calculation: Computes Cohen's d to quantify the magnitude of differences.
  • Multiple Comparison Correction: Applies Bonferroni correction to control for false positives.
  • Normality Testing: Assesses data distribution using the Shapiro-Wilk test.
  • Use Case: After running a SWARM experiment with varying transaction tax rates, use this Skill to analyze the welfare metric, identify which tax rates significantly impact welfare, and understand the size of those effects.

Quick Start

Analyze the results in '/path/to/your/swarm_results.csv' and save the summary to '/path/to/output/directory'.

Frequently Asked Questions about statistical-analysis

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform statistical analysis on SWARM experiment data from a CSV file?

Statistical analysis on SWARM experiment data requires a CSV file containing your experimental results. The Skill processes the file to identify significant parameter impacts on specified metrics and saves a summary output to your chosen directory.

What statistical tests are used for hypothesis testing and effect size calculation?

Hypothesis testing uses pairwise Welch's t-tests to compare experimental groups. Effect size is quantified using Cohen's d, while multiple-comparison corrections are applied via the Bonferroni method to control for false positives.

Do I need pandas, scipy, and numpy installed to analyze SWARM data?

Yes, scipy, pandas, and numpy are required dependencies for robust statistical computations and data handling. These libraries provide the foundational framework for executing t-tests, effect size calculations, and data processing.

How does normality testing work for SWARM experiment results?

Normality testing assesses data distribution using the Shapiro-Wilk test. This determines whether your experimental metric data meets normality assumptions before proceeding with parametric statistical comparisons.

What is the best way to identify which experimental parameters significantly impact a welfare metric?

To identify significant parameter impacts on a welfare metric, run the Skill on your SWARM results CSV. It automatically performs pairwise comparisons and applies Bonferroni correction to isolate genuinely significant parameters.

Are there limitations when using Bonferroni correction for multiple comparisons in SWARM data?

The Bonferroni correction controls false positives during multiple comparisons but can be conservative, potentially reducing statistical power. This Skill balances this by also calculating Cohen's d to ensure effect sizes remain practically meaningful.