Teaching Robust Methods for Exploratory Data Analysis.
Abstract
This paper is an introduction to some of the ideas of robust statistical methods as was presented to the Fourth International Congress for Mathematical Education, session on Exploratory Data Analysis. Most statistical methods taught and used today are very sensitive to bad or atypical data and can give meaningless results in their presence. Robust methods protect against these undesirable effects and can be incorporated into the teaching of statistics at all levels of complexity. We discuss the need for robust methods to supplement (not replace) standard procedures, suggest some considerations regarding teaching, and review some of the fundamental concepts of robust estimation. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1980
- Accession Number
- ADA092661
Entities
People
- Andrew F. Siegel
Organizations
- Princeton University