Measurement error in nonlinear models course announced

Monday, March 17, 2003 North American Users Group meeting Program Tuesday, March 18, 2003 Measurement error in nonlinear models course 8:30 – 9:30, registration (coffee and doughnuts provided) 9:30 – 4:30, with 1 hour break for lunch, plus coffee breaks lunch and refreshments will be provided for participants Longwood Galleria Conference Center 342 Longwood Avenue Boston, Massachusetts   Professional Student North American Stata Users Group meeting and Measurement error in nonlinear models course $85 $55 Measurement error in nonlinear models course only $45 $45 North American Stata Users Group meeting only $45 $20

Raymond Carroll and James Hardin will be presenting a one-day short course on Measurement error in nonlinear models. It will follow the second North American Stata Users Group meeting.

The Measurement error in nonlinear models short course is based on the book

Measurement Error in Nonlinear Models R. J. Carroll, D. Ruppert and L.A. Stefanski Chapman and Hall/CRC Press 1995 ISBN 0 412 04721 7

and is an outgrowth of recent work with StataCorp by the presenters along with Joseph Newton, Henrik Schmiediche, and Tamara Stoner to develop new statistical techniques along with software for estimation in the presence of measurement error.

This course concerns analysis strategies for nonlinear regression problems in which predictors are measured with error. Such problems are commonly known as measurement-error or errors-in-variables modeling, a topic for which there exists a large amount of literature in the case of linear models.

Nonlinear errors-in-variables modeling began in earnest in the early 1980s with the publication of a series of papers on diverse topics: Prentice (1982) on survival analysis; Carroll, Spiegelman, Lan, Bailey and Abbott (1984) and Stefanski and Carroll (1985) on binary regression; Armstrong (1985) on generalized linear models; Amemiya (1985) on instrumental variables; and Stefanski (1985) on estimating equations. In 1987, David Byar and Mitchell Gail organized a workshop on the topic at the National Institutes of Health. In 1989, the results of the workshop were published as a special issue in Statistics in Medicine. Since these early papers, the field has grown dramatically.

The course will cover the following:

• Applications in which measurement error is of concern
  
  
• Definition of basic terminology of the error structure
  
  
• Illustration of consequences of measurement error in linear regression
  
  
• Effects of measurement error on hypothesis testing
  
  
• Regression calibration and simulation-extrapolation
  
  
• Instrumental variables
  
  
• Likelihood-based approaches to estimation and inference

In addition to showing how to obtain estimates, methods for obtaining bootstrap, jackknife, and robust ("sandwich") standard errors will be discussed.

A Stata module for estimating these models will be introduced and made available.

Registration and cost

The North American Users Group meeting will be held on March 17, 2003, and the Measurement error in nonlinear models course will be held the next day at

Longwood Galleria Conference Center
342 Longwood Avenue
Boston, Massachusetts

A block of rooms has been reserved at the Best Western Boston – the Inn at Longwood Medical for March 15–19, 2003. The hotel is offering a limited number of sleeping rooms at a discount rate of $149.00 plus 12.45% tax per room per night. Please call the hotel direct at 617-731-4700 or toll free at 800-468-2378 and ask for in-house reservations. Note that the last day to make reservations at the discounted rate is February 22, 2003. For more information, visit the hotel's web site at http://www.bestwestern.com/prop_22028 or Best Western's web site at http://www.bestwestern.com.

About the presenters

Dr. Carroll is the author of 3 books and over 200 professional papers, including papers on measurement error modeling, regression variance functions and transformations, nutrition, toxicology, and bioinformatics. Dr. Carroll received his Ph.D. in Statistics from Purdue University in 1974.

James Hardin is Lecturer and Assistant Research Scientist at Texas A&M University and previously was a Senior Statistician at StataCorp, where he developed Stata's cross-sectional time-series capabilities. He is also the author of Stata's current GLM command. He is the author of two books and ten refereed papers, and he has recently been working with Henrik Schmiediche, developing the software, in Stata, for the estimation of measurement-error models. Dr. Hardin received his Ph.D. in Statistics from Texas A&M University in 1992.

Purchase the book