Submitted to: Meeting Abstract
Publication Type: Abstract only
Publication Acceptance Date: 9/1/2007
Publication Date: 10/17/2007
Citation: Holden, J.M., Bhagwat, S.A. 2007. Development of Sampling Strategies for Foods. FoodComp 2007, October 1-17, 2007, Wageningen, Netherlands. Meeting Abstract. Interpretive Summary:
Technical Abstract: National nutrient databases rely, in part, on the generation of original analytical data to estimate nutrient values for commonly consumed foods. The generation of representative analytical values requires the collection of units (e.g., packages of cereal, heads of cabbage) of a food according to predetermined sampling plans. The development of this plan includes consideration of multiple factors (e.g., brand, climate, maturity locations) which have an impact on the mean and statistical variability of values for a given food and component. Foods can be sampled in various ways according to the specific objective of the study. The strategy for sampling must include a demographic plan coupled with characteristics about the food itself to define the locations for sampling, the type of sample units to select, and the number of units required. USDA has conducted food sampling research through the National Food and Nutrient Analysis Program to determine the optimal sampling scheme for each food. A probability based sampling strategy has been developed to assure the representativeness of food samples collected. Sampling plans for more than 400 agricultural products (e.g., plant products, meats, flours) and processed foods (e.g., pizza, snack foods) have been developed. Basic statistical concepts of sampling will be discussed and will be demonstrated by developing sampling plans for four foods (poultry, bread, margarine, and one processed product) with application to developed and developing countries.