Please provide a description of the project below. This description will help students better understand the most important and defining components of the project. It will be used in informational materials and on the application page where students apply.
In the field of statistics many methods are used because they “work.” But what does that mean? How do we know that we can trust a randomization test or a bootstrap confidence interval? How far from normal does a population have to be before we should worry about using a t-test? Do we really need 5 observations per cell before we can trust a chi-square test? How well does the “plus 4” method for constructing a confidence interval for a proportion work? In this project we will use computer simulation, by programming in the R language, or other tools to investigate questions such as these.
Before Winter Term, participants will consult with the project sponsors about statistics questions they might choose to pursue. During Winter Term, participants will further refine these questions. The sponsors will support student efforts to carry out computational experiments to address chosen questions. We will also support learning background in R, which we expect will be used in most projects.
Many participants will work in pairs or small groups. The entire group will meet twice a week to discuss general topics, share progress, and brainstorm. Each participant will also meet with a sponsor regularly about their own project focus.
We hope that students will have taken at least one of STAT 113, STAT 114, STAT 205 or a similar course, but please talk with the sponsors if you have any questions about background. This project is appropriate for first- and second-year students.
Interested students are encouraged to contact the sponsors, Colin Dawson and Jeff Witmer, as soon as possible.
Number of students:
January 3rd - January 28th, 2020