Why use sampling in statistics?

The goal is the whole of the object to be studied. For example, if a manufacturer checks whether a batch of products is qualified, the overall goal is the batch of products. Sampling population is the population from which samples are drawn.

Theoretically, the sampling population should be consistent with the target population, but in practice, there are often inconsistencies. For example, scientists test the overall impact of drugs on humans through mouse experiments.

Its basic requirement is to ensure that the sampling unit can fully represent all samples. The purpose of sampling is to estimate and infer all the sample characteristics from the analysis and research results of the sampled sample units, which is an economical and effective work and research method commonly used in scientific experiments, quality inspection and social surveys.

Extended data

superiority

(1) reduces the sampling error, and increases the uniformity in the stratified layer, thus reducing the variability of observation values and the sampling error of each layer. With the same sample content, the total standard error of stratified sampling is generally less than that of simple random sampling, systematic sampling and cluster sampling.

(2) The sampling method is flexible, and different sampling methods can be adopted for different layers according to the specific conditions of each layer. For example, the prevalence of a disease among residents in a certain place can be divided into two levels: urban and rural. Urban population is concentrated. Systematic sampling method can be considered; The rural population is scattered, and the cluster sampling method can be used.

(3) Different levels can be analyzed independently. The disadvantage of stratified sampling is that if the stratified variables are not selected properly, the intra-layer variation is large and the inter-layer mean value is similar, stratified sampling will lose its significance.