How to choose sampling method in market survey?

Three. The types of sampling methods and their significance can be divided into two categories: 1. Probability-Sampling, that is, when sampling, the probability that each sampling unit in the parent population is selected as a sample is the same. Random sampling has a sound statistical theoretical basis and can be explained by probability theory. It is an objective and scientific sampling method, which is usually used in market research. 2. Non-probabilistic sampling, when sampling, the probability that the sampling unit is selected as the sample is unknown. There are four types of non-probabilistic sampling: (1) convenience sampling only considers the proximity to the sample or the convenience of measurement. One example is visiting passers-by. (2) Fixed sampling (a) Select "control characteristics" as the standard for subdividing the matrix. B subdivide the matrix into several sub-matrices, and allocate the sample size of each sub-matrix according to the comparison. C the interviewer has great freedom in selecting sample individuals in the sub-matrix. As long as the quota survey is completed, it is completed. This method loses its accuracy because of its investigation preference and convenience. Sampling quota allocation table, selected by visitors, without any modification. (3). Judgment sampling When the structure of the matrix is very different and the number of samples is very small, the sampling designer must have a good understanding of the relevant characteristics of the matrix when selecting individual samples according to the judgment of the sampling designer. When compiling the price index, the selection of product items and the decision of sample area are determined by judgment sampling. (4) Snowball sampling adopts random method or social survey to select the original interviewee. And then obtain other interviewee according to that information provide by the original interviewee. The purpose of this law is that matrices are difficult to find or very rare. For example, single-parent families are sampled. The types of random sampling are as follows: 1. Simple random sampling is used to sample all individuals with uniform probability distribution in the matrix, so that the probability of each individual being extracted is known and equal. Simple random sampling is the basis of other random sampling methods. The samples obtained by simple random sampling method are extracted by probability with a random number table after the matrix is numbered. Suppose a random number table is used to randomly select 150 samples from 2000 respondents. The sampling steps are as follows: (1) 2000 respondents are numbered from 000 1 to 2000. (2) From the random number table, select the starting point of the number by drawing lots. For example, it is selected as the fourth column of row 15. (3) From the set starting point, select the number, and select the number so that the number of digits of the respondents is the same: 1475, 9938, 4460, 0628, ..., and the number of valid samples is less than 2000. (4) If the sampling unit is the same as the sampling number in the random number table, it is a sample; if it is greater than the survey number, skip. (5) If there are duplicate numbers, skip them. (6) According to the above method, use 150 numbers continuously, that is, complete the sample selection. The timing of adopting simple random sampling: (1) matrix is small and the matrix list is satisfactory, which is the only source of matrix information. (2) The unit access cost is not affected by the sample unit distance.