江苏省稻米重金属镉检测多级抽样最优试验方案的探讨

Discussion of Multistage Sampling Optimum Test Plans on Rice Cadmium Detection for Jiangsu Province

为了探讨江苏省稻米重金属镉检测多级抽样最佳试验方案,本文以一个三级抽样检测结果为例,经方差分析F检验,发现地市间和样点村间差异均极显著,各级误差方差以地市间最大。在此基础上,探讨了在一定精确度保证下的最优配置抽样方案。结果表明,为了提高调查的精确度,在经费许可的条件下可采用随机抽取6个地市,每个地市随机抽取20个样点村或农技推广站,每个样点村随机重复2次的抽样方案;若资金较为紧张,在保证当前精确度的基础上,亦可以采用随机抽取6个地市,每个地市随机抽取10个样点村或农技推广站,每个样点村随机重复2次的抽样方案。

A three stage sampling survey as a real example was used to investigate the optimal plan of the sampling survey for cadmium in rice in Jiangsu Province. F-test showed that differences of the survey among cities and sampling villages were both very significant, and the error variance among cities was the largest in the experiment. In this paper, considering the survey expenses, the optimal sampling plan was explored under a certain accuracy. The results indicated that six cities were randomly sampled and twenty villages or agricultural extension stations were selected at random from each city, with twice sampling in each village in order to increase the precision of the survey. If there is a lack of funds, on the basis of ensuring the current sampling accuracy, we suggest that ten villages or agricultural extension stations from each city can be sampled alternatively.