Abstract #301188

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 #301188
Activity Number: 27
Type: Contributed
Date/Time: Sunday, August 3, 2003 : 2:00 PM to 3:50 PM
Sponsor: General Methodology
Abstract - #301188
Title: The Best Linear Predictor for True Score from a Direct Estimate and a Derived Estimate
Author(s): Jiahe Qian*+ and Shelby J. Haberman
Companies: Educational Testing Service and Center for Statistical Theory & Practice Education Testing Service
Address: Rosedale Rd., MS 02-T, Princeton, NJ, 08541-0001,
Keywords: best linear predictor ; direct estimate ; derived estimate ; minimum mean square error ; e-rater
Abstract:

Statistical prediction problems often involve both direct and derived estimates of the same score. For example, in GMAT tests, a final essay score is based on a holistic score and an estimated score derived from the e-rater(TM); model, an automated essay scoring system invented at ETS. This study proposes best linear predictors to approximate the true score in the problem. We consider the direct estimate to be a random true score plus an independent random error, and consider the derived estimates to be random and correlated with the true score. The minimum mean square error is the criterion used in deriving the predictor. The best linear predictor allocates the direct estimate and the derived estimate with unequal weights. The weighting used depends on the size of the error variance of the direct estimate and on the size of the correlation of the true score and derived estimate. One application of the best linear predictor is to approximate the human true score from the observed holistic score and the e-rater score.

  • 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