Abstract #301744

This is the preliminary program for the 2003 Joint Statistical Meetings in San Francisco, California. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 2-5, 2003); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2003 Program page
JSM 2003 Abstract #301744
Activity Number: 129
Type: Contributed
Date/Time: Monday, August 4, 2003 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract - #301744
Title: Small Sample Performance of GEE-Ordinal Regression Estimators of ROC Curves
Author(s): Alicia Y. Toledano*+
Companies: Brown University
Address: Center for Statistical Sciences, Providence, RI, 02912-0001,
Keywords: GEE ; ordinal regression ; ROC curve
Abstract:

We performed Monte Carlo studies of GEE-ordinal regression estimators of ROC curves (Toledano and Gatsonis 1996) in small datasets. Data were generated considering test readers either as fixed or as a random sample from a population of potential readers. Working independence (WI) and a design-induced (3PC) correlation structure were applied. In a bivariate setting (e.g., 2 imaging modalities), a quadrivariate setting (e.g., 4 readers), and an octavariate setting (e.g., the combination thereof), estimation of areas was unbiased. For fixed readers, WI model-based naive SEs (NSEs) are too large for differences and too small for averages; 3PC NSEs, WI robust SEs (RSEs), and 3PC RSEs are close to the empirical SDs of the estimated differences and averages. In the random reader setting, results for SEs conditional on the reader sample mirror those for fixed effects. Unconditionally, a simple and intuitive correction to the SEs from the GEE-OR method provides SEs for the difference in average area between the modalities that are slightly too large, and approximately 96% coverage of nominal 95% CIs for this difference, using 3PC NSEs, WI RSEs, or 3PC RSEs.

  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2003 program

JSM 2003 For information, contact meetings@amstat.org or phone (703) 684-1221. If you have questions about the Continuing Education program, please contact the Education Department.
Revised March 2003