Resampling Approaches to Data Analysis

Resampling methods are approaches to conducting statistical inference (e.g., standard error estimation, confidence interval construction, hypothesis testing) that rely on the power and speed of computers to construct sampling distributions for statistics of interest. These methods are appealing for natural resource sciences because they offer a way to conduct inference without having to assume underlying distributions for collected data or statistics of interest. In this class, students are exposed to common resampling approaches including jackknifing, bootstrapping, and randomization/permutation testing. Particular attention is paid to bootstrapping with coverage including multiple approaches for constructing bootstrap confidence intervals and different bootstrap data generating methods. Although the course delves into some of the underlying theory for the various approaches, the primary focus of the course is application of the methods.

This class uses R and is designed for students who have at least a basic background in programming -- the equivalent of one semester of R, or any similar programming language (e.g., JavaScript, C...).

Instructor: Dr. Travis Brenden

Class Format and Sections

This is an asynchronous-online class that is non-credit and self-paced -- the class is about the equivalent of a 3 credit course. Students can start the class at any time and have six months to complete the course.