|
Activity Number:
|
41
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 3, 2003 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #301026 |
|
Title:
|
School Survey Issues: Design and Estimation
|
|
Author(s):
|
William D. Kalsbeek*+ and Robert Agans and Amanda Pyles and Eric S. Pevzner and J. Michael Bowling and Kurt M. Ribisl
|
|
Companies:
|
University of North Carolina and University of North Carolina and University of North Carolina and University of North Carolina and University of North Carolina and University of North Carolina
|
|
Address:
|
Survey Research Unit, Chapel Hill, NC, 27599-0001,
|
|
Keywords:
|
nonsampling error ; nonresponse bias ; question context effects
|
|
Abstract:
|
The National Youth Tobacco Survey (NYTS) and National Youth Risk Behavior Survey (NYRBS) are biennial surveys that are important sources of national youth tobacco use information. The NYTS and NYRBS utilize comparable sampling approaches, identical modes of data collection (school-based and self-report), and the same wording for the 12 tobacco questions that are included in both surveys. Despite these similarities, estimates of smoking prevalence derived from contemporary data in these two surveys have been found to differ somewhat. This paper presents some early findings from a study to explain differences between estimates obtained from the spring 1999 NYRBS and the fall 1999 NYTS. These findings focus on how variation in study design and implementation associated with sample nonresponse and question sequencing in these surveys may help to understand the observed differences in reported prevalence. Wherever possible, these study effects are quantified. Other potentially expository study effects are discussed.
|
