What constitutes a 'robust' finding?

A finding is considered robust when conclusions remain consistent and significant across various analytical decisions, different tests, and data exclusions.

Can I use this for meta-analyses?

Yes, it specifically supports 'one-study-removed' analysis and the exclusion of high risk-of-bias studies to test the stability of meta-analytic findings.

What is sensitivity analysis in this context?

It is a method used to determine how different values of an independent variable or different analytical choices impact a particular result under a given set of assumptions.

How does this skill handle missing data?

The skill can perform complete case analysis, best/worst-case scenarios, and multiple imputation to see how missing data affects the final conclusions.

Sensitivity Analysis

Name: Sensitivity Analysis
Author: astoreyai

byastoreyai

•

データサイエンスとML

Evaluates the robustness of research findings by testing how results change under different analytical assumptions and data conditions.

概要

The Sensitivity Analysis skill empowers Claude to rigorously validate research findings by examining the impact of varying analytical decisions. It automates complex checks such as exclusion analyses (removing outliers or high-bias studies), testing different statistical methods, and evaluating the effects of missing data or alternative measurement definitions. By identifying whether results are robust or sensitive to specific conditions, it ensures that data-driven conclusions are reliable and high-quality for scientific, academic, or business reporting.

主な機能

Automated outlier and high-risk-of-bias study exclusion
Comparative statistical testing across parametric and non-parametric methods
Missing data impact evaluation using multiple imputation or worst-case scenarios
Standardized reporting of findings consistency and uncertainty
2 GitHub stars
Robustness checks for alternative outcome definitions and scoring methods

ユースケース

Verifying if a meta-analysis result holds when removing specific high-bias studies
Testing if a model's performance is sensitive to data imputation techniques
Validating business insights across different statistical transformations

概要

主な機能

Automated outlier and high-risk-of-bias study exclusion
Comparative statistical testing across parametric and non-parametric methods
Missing data impact evaluation using multiple imputation or worst-case scenarios
Standardized reporting of findings consistency and uncertainty
2 GitHub stars
Robustness checks for alternative outcome definitions and scoring methods

ユースケース

Verifying if a meta-analysis result holds when removing specific high-bias studies
Testing if a model's performance is sensitive to data imputation techniques
Validating business insights across different statistical transformations