SPES Sessions at the Joint Statistical Meetings in Indianapolis, August 13-17, 2000

Robert N. Rodriguez, 2000 JSM Program Chair

In support of the JSM 2000 theme, the Section on Physical and Engineering Sciences will be celebrating the Meetings with a variety of traditional and forward-looking sessions that reflect the rich diversity of its members and their interests.

SPES will be the primary sponsor for four invited sessions:

• Statistical Challenges in the Semiconductor Industry. This session explores the challenges facing the statistical community in its support of the semiconductor industry, which is increasingly applying statistical techniques to its diverse problems. The organizer is Barry Hembree, and the participants are Joe Lebowitz, Richard Kittler, Paul Tobias, Karen Horrell, Terri Moore, George Milliken, and George Runger.
• New Ideas for Old Problems in Chemometrics. This session reinforces the interface between statistics and chemistry by addressing topics of major influence at this interface: the issue of calibration, the use of latent variable and partial least squares models, and the reconciliation of information from multiple experiments. The organizer is David Banks, and the participants are Will Guthrie, Cliff Spiegelman, Mark Levenson, Alison Burnham, John MacGregor, Paul Nomikos, and David Duewer.
• Sample Size Issues. Adequate sample size and power are vitally important in making sure that studies will contain adequate data to meet scientific goals. This panel discussion provides various points of view on these issues, which are encountered in many industries. The organizer is Russell Lenth, and the participants include Thomas Louis, Janet Elashoff, Pamela Getson, and Keith Muller.
• Experiments for Optimization and Robustness. System optimization and robustness are two main objectives in many scientific and engineering investigations, and traditionally these objectives have been addressed independently with response surface methodology and robust parameter designs. This session deals with the interface between these two objectives. The organizer is Jeff Wu, and the participants include Randy Sitter, Derek Bingham, David Steinberg, Yisrael Parmet, and Shao-Wei Cheng.

SPES is also co-sponsoring the following invited sessions:

• The Technometrics Invited Session
• Moving Statistics Into the New Millennium, sponsored by the Council of Chapters
• Developments in Complex Split-Plot Experiments, sponsored by the Section on Quality & Productivity
• Some Statistical Challenges for the Automobile Industry, sponsored by Q&P
• Contributions of Statistical Science to Weapons Treaty Verification and National Security, sponsored by the Seismological Society of America and organized by SPES member Dale Anderson. A number of distinguished scientists and statisticians will showcase the application of statistical sciences to national security and verification of treaties, such as the Comprehensive Nuclear-Test-Ban Treaty, the Chemical Weapons Convention, and the Biological and Toxin Weapons Convention.

There are two SPES topic contributed sessions:

• The Isolated Statistician: Challenges and Opportunities for the Statistics Profession. Organized by Robert Rodriguez and Jose Ramirez (Q&P Program Chair), this panel discussion describes problems faced by isolated statisticians working in industry and identifies various solutions.
• Dealing with Large Data Sets. This session addresses the design and analysis of large data sets, which are becoming important in areas such as drug discovery as the result of automation and robotics. The organizer is Stan Young, and the participants include William Welch, Raymond Lam, Aijun An, Christophe Lambert, and Doug Hawkins.

The SPES regular contributed sessions cover advances in a broad range of areas, including design of experiments, reliability and fatigue data analysis, and statistical modeling in engineering and physical science. SPES is also co-sponsoring Q&P contributed sessions in the areas of statistical process control and adjustment.

Join us in Indianapolis for the celebration!

For More Information Contact:

Robert N. Rodriguez
SAS Institute Inc.
SAS Campus Drive
Cary, NC 27513-8000
(919) 677-8000 x7650
Bob.Rodriguez@sas.com

Joanne Wendelberger, 2001 JSM Program Chair

SPES will be sponsoring 5 Roundtable Luncheons at the 2000 Joint Statistical Meetings in Indianapolis in August. Luncheon moderators include

• Karen Kafadar, University of Colorado-Denver - "Publication of Statistical Applications"
  A discussion of applications papers as a motivator for new statistical methodology will address issues such as key ingredients, benefits to a variety of readers, and stimulation of future developments, both in generality and with reference to specific papers having applications in physical, chemical, and engineering sciences (e.g., 1993 Technometrics 256-267; 1995 Technometrics 176-184; 1989 JASA 945-957; participants please suggest others to kk@math.cudenver.edu).
• Cliff Spiegelman, Texas A&M University - "State-of-the-Art Calibration Methods"
  The statistics profession has made important contributions in the area of chemometrics; state-of-the-art techniques for calibration provide an opportunity for an enlarged role for statisticians in the future of chemometrics.
• Alyson Wilson, Los Alamos National Laboratory - "Data Integration"
  When a decision has to be made in the presence of multiple sources of information, including designed experiments, results from computer models, and expert opinion, what are the appropriate ways to integrate all of these sources of information?
• Donald McCormack, Sematech - "Teaching Statistics in an Industrial Setting"
  Statisticians teaching in industry face unique challenges including classroom and nonclassroom instructional settings, nonhomogeneous audiences of varying ages and educational backgrounds, and conflicting student priorities; this luncheon will discuss/debate how to best develop, deliver, and evaluate statistics instruction in industrial settings.
• Luis Escobar, Louisiana State University - "Degradation Data Analysis and Modeling"
  Analysis and modeling of degradation data provides valuable insights into the condition of a system over time and provides information which can in some cases extend the user's ability to predict reliability behavior beyond what can be gleaned from traditional approaches based on analysis of failure data.

New for 2000, there will be 10 SPES Young Investigator student luncheon awards available to pay for lunches for students participating in the SPES Roundtable Luncheons. Individuals interested in applying for these awards should contact

Joanne Wendelberger
Los Alamos National Laboratory
Statistical Sciences Group
MS F600
Los Alamos, NM 87545
joanne@lanl.gov

SPES is pleased to announce the offering of two short courses at the 2000 JSM in Indianapolis. The date and time of each course will be published in the JSM registration packets which will be mailed to all ASA members in May/June 2000.

• Statistical Methods for Failure Time and Accelerated Test Data

  Dr. William Meeker and Dr. Luis Escobar will present this one-day course on modern methods for analyzing failure-time and degradation data. This course will describe and illustrate the use of a mix of proven traditional techniques, enhanced and brought up-to-date with modern computer-based methodology. Topics to be covered include censored data, nonparametric estimation, probability plotting, maximum likelihood estimation, likelihood-based confidence intervals, acceleration models, accelerated life testing, and accelerated degradation testing. Most of the examples in the course will come from applications of product reliability, but some biological examples will also be presented to illustrate the breadth of application.

  Dr. Meeker is Professor of Statistics and Distinguished Professor of Liberal Arts and Sciences at Iowa State University. Dr. Escobar is a Professor in the Department of Experimental Statistics at Louisiana State University.

  This course is being co-sponsored by the SPES and Q&P sections and was succesfully presented at the 1998 JSM in Dallas, Texas.

• Wavelets in the Real World: Concepts and Applications

  Dr. R. Todd Ogden's presentation of this half-day course is designed to give the participants enough background information on wavelets to allow them to begin using wavelets in data analysis. The first part of the course will introduce participants to wavelet functions, starting with the simplest wavelet function, the Haar wavelet. This will be followed by multiresolution analysis, computational aspects of the discrete wavelet transformation, and a comparison of the strengths and weaknesses of a variety of other wavelet families.

  The second part of the course will treat a variety of standard statistical problems and demonstrate how the strengths of wavelet methods lend themselves well to these problems. Applications such as histograms, density estimation, nonparametric regression, thresholding, and time series will be addressed. The course closes wil an overview of some of the wavelet software that is available, with a discussion of their relative strengths.

  Dr. Ogden is an Assistant Professor at the University of South Carolina.