Evaluating the Impact of Bootstrap Resampling on the Stability and Reliability of Regression Model Estimates

Posted: Aug 06, 2008

• Sophia Harris
• Mia Lee
• Theodore Rodriguez

Abstract

Bootstrap resampling, since its introduction by Bradley Efron in 1979, has revolutionized statistical practice by providing a computationally intensive but conceptually straightforward approach to estimating sampling distributions and constructing confidence intervals. The method's appeal lies in its ability to make minimal assumptions about the underlying data generating process while offering robust inference procedures. In the context of regression modeling, bootstrap techniques have been extensively employed for variance estimation, bias correction, and model validation. However, a critical but often overlooked aspect of bootstrap methodology concerns its impact on the stability and reliability of the very parameter estimates it seeks to evaluate.