|
Activity Number:
|
341
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #309688 |
|
Title:
|
Diagnostic Process to Assess the Effects of Truncating Extreme BRFSS Sampling Weights
|
|
Author(s):
|
Henry Roberts*+ and Elizabeth Hughes and Robert Woldman and Ruth Jiles
|
|
Companies:
|
Centers for Disease Control and Prevention and Northup Gruman Inc and North Carolina Department of Health and Human Services and Centers for Disease Control and Prevention
|
|
Address:
|
4770 Buford Hwy NE MSK66, Atlanta, GA, 30341,
|
|
Keywords:
|
BRFSS ; Weighting-Class Variables ; Nonresponse Bias ; Nonsampling Error ; Standard Error
|
|
Abstract:
|
This research focuses on developing and implementing a diagnostic process to study the effects of truncating extreme BRFSS sampling weights; and the extent that extreme sampling weights affects the robustness and precision of BRFSS prevalence estimates. This diagnostic process sequentially assesses: variability, robustness, and precision. The truncation of extreme sampling weights resulted in reducing the variance of the pre-truncation sampling weights for Hispanics and adults ages 18-34 by 298% and 181%, respectively. Estimates that were robust or non-robust prior to the truncation of extreme sampling weights remained robust or non-robust. Overall, the impact of truncating BRFSS sampling weights were: no effect on robustness; large deceases in the variability of the sampling weights; and large decreases in standard errors resulting in enhanced precision.
|
