Activity Number:
|
255
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Survey Research Methods*
|
Abstract - #301388 |
Title:
|
Sampling and Analysis Plan for a National Fluoride Study
|
Author(s):
|
Michael Bellow*+ and Charles Perry, Jr. and Pamela Pehrsson
|
Affiliation(s):
|
U.S. Department of Agriculture/NASS and U.S. Department of Agriculture/NASS and U.S. Department of Agriculture/ARS
|
Address:
|
3251 Old Lee Hwy., Rm. 305, Fairfax, Virginia, 22030, U.S.A.
|
Keywords:
|
PMR sampling ; mixed effects model ; time replicates
|
Abstract:
|
The U.S. Department of Agriculture's Nutrient Data Laboratory (NDL) is conducting a comprehensive study of the fluoride content of highly consumed beverages and a select number of fluoride-containing foods in the U.S. food supply. In preparation, we developed a sampling plan that ensured unbiased estimation of fluoride content of identified foods and beverages at national and regional levels, and accurate estimation of variability of estimates for all but low priority items at levels of interest. The sampling method involved: 1.) serpentinely ordering the U.S. population by census region, census division, and county; 2.) dividing the population into 72 roughly equal population zones; 3.) selecting one county from each zone; 4.) selecting two data collection sites (residential or retail outlets) in each sampled county; and 5.) collecting samples from all selected residential sites for tap water (twice over time) and a subset of the retail sites for commercial beverages and foods. We used mixed model analysis to find an optimal sample allocation and to estimate components of variance of fluoride content estimates. In this paper, we describe the sample design and analysis plan in detail.
|