Sampling Methodology and Coverage Issue for a National School-Based Survey

• Activity Number: 530 - SPEED: Survey Research Methods
• Type: Contributed
• Date/Time: Wednesday, August 2, 2017 : 10:30 AM to 11:15 AM
• Sponsor: Survey Research Methods Section
• Author(s): S. Sean Hu* and Linda Neff and Ronaldo Iachan and Kate Flint and Alice Roberts and Karen Cullen
• Companies: CDC and Centers for Disease Control and Prevention and ICF International and ICF International and ICF International and Food and Drug Administration
• Keywords: Sampling ; Survey ; School-based

Abstract

The National Youth Tobacco Survey (NYTS) is a cross-sectional, school-based, self-administered, pencil-and-paper questionnaire administered to U.S. middle school and high school students. A three-stage cluster sampling procedure is used to generate a nationally representative sample of U.S. students attending public and private schools in grades 6-12. A probability based sample is used at each of the three stages: primary sampling units (a county, part of a county, or groups of small counties), schools, and classes. This paper presents the results of a review of the current NYTS sampling methodology, including a summary of potential refinements. The review assessed sampling design features such as sampling stages, sampling units, stratification, allocation, and subgroup sample sizes. Two identified areas for refinement include: 1) the ability to expand the sampling frame without negative impacts on school participation rates or the current budget; and 2) the feasibility of developing precise sub-national estimates by grade, sex, and race/ethnicity for middle and high school separately. These findings could help enhance the accuracy and utility of future waves of NYTS.