湖北省稻谷质量调查、品质测报样品扦样量优化的研究

摘要

稻谷质量调查、品质测报是依据当年抽样稻谷样本数据来估计当年稻谷总体的质量水平.对2014年湖北省调查测报数据进行分析研究,结果表明:湖北省在市/县范围内进行扦样,品种选择按照"3+1"原则是合理的;由2014年样本数学模型计算出糙率、整精米率、直链淀粉含量、食味品质、垩白度的抽样误差分别为0.75、2.61、1.68、2.25、5.16;在95%的置信区间的条件下,以各市为扦样单位,2014年调查、测报的最小抽样个数分别以342、654个为宜;4、由2013年、2014年、2015年的数据分析知,3个年度5个指标的方差是随着年度变化而改变的,且显著不同.应先对当年样本方差进行估计,进而确定出当年的最小样本量,后按照规程合理分布各地样品量.%Paddy quality survey and forecast were based on the sampling data to estimate the overall level of quality in paddy. The survey data of Hubei Province in 2014 showed that:the province in the city/county sampler,variety selection in accordance with the"3 + 1"principle were reasonable;by the sample mathematics model of 2014,the sample error of husked rice yield, head rice yield,amylase content,edible quality and chalkiness were calculated as 0.75,2.61,1.68,2.25 and 5.16;at the 95%confidence interval,with each city as a sampler unit,the sampling number of 2014 quality survey and forecast were respectively 342,654;according to 2014,2013 and 2015 data analysis,the variance of five indicators in three annual reports were changed year by year,and were significantly different.We should estimate the sample variance,and then determine the minimum sam-ple size of the year,give a reasonable distribution throughout the sample according to the rules.