Application of binomial and multinomial probability statistics to the sampling design process of a global grain tracing and recall system

被引:11
作者
Lee, Kyung-Min [1 ]
Armstrong, Paul R. [2 ]
Thomasson, J. Alex [3 ]
Sui, Ruixiu [4 ]
Casada, Mark [2 ]
Herrman, Timothy J. [1 ]
机构
[1] Texas A&M Univ Syst, Texas Agr Expt Stn, Off Texas State Chemist, College Stn, TX 77841 USA
[2] ARS, Ctr Grain & Anim Hlth Res, USDA, Manhattan, KS 66502 USA
[3] Texas A&M Univ, Biol & Agr Engn Dept, College Stn, TX 77843 USA
[4] ARS, Cotton Ginning Res Unit, USDA, Stoneville, MS 38776 USA
关键词
Grain traceability system; Sampling design process; Binomial probability distribution; Multinomial probability distribution; Sample size; SIMULTANEOUS CONFIDENCE-INTERVALS; WHEAT; SIZE;
D O I
10.1016/j.foodcont.2010.12.016
中图分类号
TS2 [食品工业];
学科分类号
0832 ;
摘要
Small, coded, pill sized tracers embedded in grains are proposed as a method to store a historical record of grains and retrieve coded information for grain traceability. This study aimed to develop and validate a statistical sampling procedure to securely collect sample sizes (kg) and number of tracers since the sampling accuracy is critical in the proposed traceability system for capturing information and data related to grain lots to trace the grain back through the route in a grain supply chain. The statistical results and observations showed similar concentrations and insignificant segregation of tracers in bin and truck operations. The number of tracers required for identification of grain sources fell within the confidence intervals and sample sizes (kg) estimated by statistical probability methods. Truck sampling appeared more feasible in collecting the secure number of tracers over bin sampling. The designed sampling process was empirically proven to be practically applicable and provide better scientific assurance of sampling accuracy, which may reduce economic risks and their consequent costs caused by unfavorable sampling in the propose traceability system. (C) 2011 Elsevier Ltd. All rights reserved.
引用
收藏
页码:1085 / 1094
页数:10
相关论文
共 34 条
[1]   Approximate is better than "exact" for interval estimation of binomial proportions [J].
Agresti, A ;
Coull, BA .
AMERICAN STATISTICIAN, 1998, 52 (02) :119-126
[2]   GRAPHICAL METHOD TO EVALUATE SAMPLE SIZES FOR MULTINOMIAL DISTRIBUTION [J].
ANGERS, C .
TECHNOMETRICS, 1974, 16 (03) :469-471
[3]  
Angers C, 1984, SIMULATION OCT, P175
[4]  
[Anonymous], 2004, SAS US GUID STAT VER
[5]  
[Anonymous], 2004, AGR EC REPORTS
[6]  
[Anonymous], 2002, OFFICIAL J EUROPEAN
[7]  
*CAN TRAC, 2003, AGR AGR CAN
[8]  
Coker Raymond D., 1995, Natural Toxins, V3, P257, DOI 10.1002/nt.2620030417
[9]  
*ERS, 2001, KNOW GOING BRIN FOOD
[10]   QUICK SIMULTANEOUS CONFIDENCE-INTERVALS FOR MULTINOMIAL PROPORTIONS [J].
FITZPATRICK, S ;
SCOTT, A .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1987, 82 (399) :875-878