Course Description :

One of the strengths of the Six Sigma Black Belt program is a strong grounding in the fundamental tools of statistics. This course builds on that foundation and adds two new important components: distribution-free methods that "free" the user from assuming the data are normally distributed and robust methods methods that are resistant to outliers. Classical rank-based methods such as the Wilcoxon Rank Sum test are by their nature both distribution-free and robust and will be a main focus. However, the ideas behind these methods will allow the use of a much broader array of techniques.

The course will include:

• A review of classical procedures: sampling distributions, t-tests, one-way ANOVA, two-way ANOVA, block and factorial designs.
• Rank and robust procedures: Wilcoxon Signed Rank and Rank Sum tests; Kruskal-Wallis, Jonckheere, and Friedman tests.
• The role of data transformation.
• Robust Estimators: Hodges-Lehmann rank-based estimators, trimmed means, Huber M-estimators.
• Resampling principles: permutation, bootstrap, and jackknife methods.

Participants will be encouraged to bring real problems to class. They will learn how to apply the above methods to their problems as well as to other examples using appropriate software.

About the Instructor:

Dennis D. Boos has been a professor in the Department of Statistics at North Carolina State University since obtaining his Ph.D. from Florida State University in 1977. He was elected a Fellow of of the American Statistical Association in 2001. Active in teaching, research, and consulting, Dr. Boos has created three new undergraduate courses including an innovative course (with an online version) for engineering students that features a computing manual in R, Splus, and MATLAB. He has also created three new graduate courses and has taught short courses for post-graduates that include 1) "Robust Methods in Statistics" presented at the FDA and pharmaceutical companies, and 2) "Statistics for Geneticists I" as part of the annual "Summer Institute in Statistical Genetics" offered by the NC State Statistics Department 1997-2005. His research covers a wide range of topics with current focus on variable selection in regression. He has directed 13 Ph.D. students and consults with faculty at NC State and with companies in Research Triangle Park. He is currently finishing a book with Len Stefanski entitled "Modern Statistical Inference: Theory and Methods" under contract with Springer.

Register Now! (May 20-23, 2008)