How strong must bias be to invalidate your research conclusions?
KonFound-It! provides interactive sensitivity analyses to help researchers, policy makers, and pragmatists evaluate the robustness of statistical inferences.
Why Sensitivity Analysis?
The assumptions underlying statistical analysis are rarely fully met. Pragmatists face the challenge of knowing when evidence is strong enough to justify action—and that is where sensitivity analysis helps by expressing the robustness of inferences relative to potential biases.
KonFound-It! quantifies how much of an observed effect would need to be bias-driven to invalidate your research conclusions, bridging the gap between rigorous statistical theory and actionable real-world policy.
Examples:
Robustness of Inference to Replacement (RIR): What proportion of our observed cases would we need to replace with “null effect” cases to change our minds?
Impact Threshold for a Confounding Variable (ITCV): How strongly must an omitted variable be correlated with the focal predictor (e.g., retention) and the outcome (e.g., reading achievement) to nullify an inference of an effect of the focal predictor on the outcome (e.g., retention on achievement)?
Explore the KonFound-It! Ecosystem
Project News
Read the latest release notes, upcoming workshop announcements, and research updates from the KonFound-It! team.
Read Latest News →Guides & FAQ
New to sensitivity analysis? Browse through our introductory tutorials, frequently asked questions, and practical research guides.
Read the Guide →Community Forum
Have questions about running a model or interpreting sensitivity indices? Connect with researchers and maintainers.
Join Google Group →WWC Benchmarks
Explore sensitivity analyses calculated for the What Works Clearinghouse (WWC) through our dedicated interactive benchmark tool.
Explore Benchmarks →Install Packages for R & Stata
Prefer working directly in your statistical script? Get the latest development versions of the konfound package for your workflow:
R Package (GitHub Dev)
Install the latest development release directly from GitHub:
install.packages("devtools")
devtools::install_github("konfound-project/konfound")
Stata Package (SSC)
Install the package and required dependencies directly from SSC:
ssc install konfound, replace
ssc install indeplist, replace
ssc install moss, replace
ssc install matsort, replace