Statistical inference often relies on approximations that may fail in practice, especially when working with small samples, biased data, or complex models common in biomedical research. The bootstrap offers a powerful, computationally based approach to assessing uncertainty without requiring closed-form solutions. This talk will introduce the basic principles of the bootstrap, develop intuition for its strengths and limitations, and illustrate how extensions such as the double bootstrap can improve bias correction, variance estimation, and confidence interval coverage. Through real-world examples—including applications to COVID-19 hospital stay data and HIV surveillance in developing countries—we will explore both the promise and pitfalls of resampling methods. Attendees will gain practical insights into when and how bootstrap methods can enhance statistical analyses in biomedical and public health research.