|
Activity Number:
|
33
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #309627 |
|
Title:
|
Recovering the Estimated Slope of an Unobservable Predictor in a Simple Regression
|
|
Author(s):
|
Wenyaw Chan and Hung-Wen Yeh*+ and Elaine Symanski
|
|
Companies:
|
The University of Kansas Medical Center and The University of Texas at Houston and University of Texas at Houston
|
|
Address:
|
Dept of Preventive Medicine and Public Health, Kansas City, KS, 66160,
|
|
Keywords:
|
attenuation ; measurement error ; regession ; surrogate variable
|
|
Abstract:
|
Attenuation of regressor effects in simple regression when a surrogate variable is used as a substitute for a latent predictor has long been a concern in public health. Recognizing that the surrogate variable is an imperfect measure of the true predictor, prior work has focused on estimating attenuation in the slope caused by using the surrogate variable. In addition, there is interest in recovering the estimated slope of the latent predictor. Thus, this study proposes an estimator of the attenuation using conditional expectation of the two slopes estimated from the latent and surrogate predictors given the outcome and surrogate variables. Simulation studies were conducted to examine the performance of this estimator in terms of recovering the estimated slope of the latent predictor. An application of the method is illustrated using data from the occupational health arena.
|
