The sample size depends on many factors, including the type of research, the accuracy of predetermined analysis, the size of allowable error, the homogeneity of the whole population, sampling methods, research conditions and so on. , and can't generalize.

Sample representativeness is also related to randomness (sample preparation, sampling and sample measurement) and stratified sampling. Sample representativeness or sample size needs to be appropriate, that is, balanced with other economic-related indicators, such as considering misjudgment rate, missed judgment rate and total sampling cost.

Extended data

Formula for determining sample size

According to the formula for calculating the sample size, the size of the sample size does not depend on the number of the population, but depends on:

(1) the degree of change of the research object;

(2) Required or allowable error (i.e. accuracy requirement);

(3) Infer the required confidence.

That is to say, when the phenomenon studied is more complex and the difference is greater, the requirement for sample size is greater; When the required precision is high and the requirement for inferability is high, the sample size is large. Therefore, if different cities infer separately, it is wrong in principle that big cities smoke more and small cities smoke less. Too much sampling in big cities is a waste, and too little sampling in small cities has no inferential value.

In short, when determining the sampling method and sample size, we should not only consider the purpose, nature and precision requirements (sampling error) of the survey, but also consider the feasibility of actual operation, the control of non-sampling error, budget and so on. In this regard, professional survey companies will make a comprehensive balance according to your situation and the nature of the survey to achieve an optimal sample size.