Project Introduction

This paper serves as an introduction and overview of the konfound() R package, which is a sensitivity analysis tool designed to quantify the robustness of causal inferences. It also provides an overview of two core functions within the package: konfound() and pkonfound(). The project’s primary objective is to advance, extend, and apply existing sensitivity analysis techniques in a manner most beneficial to education research and practice. Specifically, our research team aims to expand the Impact Threshold for a Confounding Variable (ITCV) and the Robustness of Inference to Replacement (RIR) approaches to various research designs and features, including those involving differential attrition, moderation analyses, and regression discontinuity designs. By doing so, we hope to provide researchers advocating for causal inference, as well as those challenging such inferences, with a more precise framework for interpreting the strength of evidence in relation to concerns about potential violations of inference assumptions.


Sarah Narvaiz
Joshua Rosenberg
Wei Wang


To be continued...