I. Basic Requirements: Clearly specify the overall population, the randomness of sampling, and the representativeness of sampling. To make the sample taken as representative of the whole as possible. Only the sample is representative, then by the sample characteristics inferred from the overall characteristics of the general, the overall research results have the value of the official. The representativeness of the sample is the theoretical basis for the practice of inferring the whole from the part.
Second, the basic method: Sampling bias will lead to invalid conclusions. For sampling to be representative, the sampling error must also be correctly estimated. Sampling error is the deviation that exists between the indicator values of the sample and the indicator values of the whole.
The smaller the value of this difference, the more correctly the sample taken to reflect the overall. Therefore, in order to ensure the representativeness of the sampling, the researcher has to analyze the factors affecting the size of the error, determine the error value by calculating the standard error of sampling, and try to control the error to the lowest degree.
Features
Sampling surveys take a portion of individuals as samples from the overall population of the research object for investigation, from which the numerical characteristics of the overall population in question are inferred, good economy, high effectiveness, wide adaptability, and high accuracy.
Sample survey is based on the results of some of the actual survey to infer the total number of overall signs of a statistical survey method, belongs to the category of non-comprehensive survey. It is in accordance with scientific principles and calculations, from a number of units composed of things in the overall, take part of the sample unit to investigate, observation, with the data obtained from the survey mark to represent the overall, inferring the overall.
Reference: Baidu Encyclopedia-Sample Survey