1.When the total variability,random smoke samples representative difference
2.From overall N a unit of arbitrary extraction N unit as the samples and make every possible sample are pumping the probability equal a sampling method
3.First the overall unit according to some features are divided into several secondary overall (layer),and then from each layer within the simple random sampling,
4.Will the overall units in a certain order,according to the requirement of sample size selection interval,then random determine the starting point,every certain interval extraction a unit of a sampling way;
5.Will the overall units merge into a no cross,no repetitive set,called the group; And then to group for sampling unit sample drawn a sampling method.