The difference between stratified sampling and cluster sampling

The hierarchical sampling style divides the surveyed market matrix into sub-primitives with different characteristics, which are generally called layers, and then randomly draws samples from each layer.

Cluster sampling, on the other hand, is a method of dividing the population into several different groups according to certain signs according to the overall characteristics, and then investigating the units in the selected groups.

To put it simply, one is to stratify and then draw samples from each layer, and the other is to divide and then draw a group survey.

Just read a book on market research.

Similarities: Both stratified sampling and cluster sampling require random sampling that divides the population by a certain marker beforehand.

Differences: 1. The logo is different.

The division markers of stratified sampling are closely related to the survey markers, while the division markers of cluster sampling are not necessarily related to the survey markers.

2. The extracted parts are different.

Stratified sampling is a random sample of the strata, while cluster sampling randomly selects a subset of the population from the entire population.

3. The error is different.

The sampling error of stratified sampling depends on the mean of the variance of the population in each layer, and the sampling error of cluster sampling depends on the inter-group variance of the population.

4. The purpose is different.

The purpose of stratified sampling is mainly to reduce the sampling error and meet the needs of inferring the quantitative characteristics of subpopulations, while the purpose of cluster sampling is mainly to expand the sampling units and simplify the organizational work.

Applicable occasions: Stratified sampling is used for large inter-layer differences and small intra-layer differences, and to meet the needs of hierarchical management decisions. Cluster sampling is used when the differences between groups are small and the differences within groups are large, or when only the population is the sampling unit.

Although cluster sampling is similar to stratified sampling, for example, the first step is to divide the population into clusters or strata according to some criteria, but there are significant differences between the two. For cluster sampling, the individuals of the whole vertical part of the group are all sample units, and the sample units of the non-group are not counted in the survey. Stratified sampling involves taking a small sample from all strata, which together make up the population sample. In other words, for stratified sampling, the respondents are from all strata, and the results are naturally more representative.

Therefore, when the differences between different subgroups are large and the differences within each subgroup are small, the residue is suitable for stratified sampling. Cluster sampling is particularly appropriate when there are small differences between different subgroups and large differences within each subgroup.

Stratified sampling. Stratified sampling is also known as categorical sampling or type sampling. It is suitable for situations where the overall quantity is large and the degree of difference is large.

The population units are first classified and stratified according to their degree of difference or a certain characteristic, and then the sample units are randomly selected in each category or layer.

Advantages and disadvantages: The advantages of cluster sampling are that it is easy to implement and cost-saving; The disadvantage is that the sampling error caused by this is often greater than that of simple random sampling due to the large differences between different groups, and the sample distribution is not wide, and the sample is relatively poor in the representativeness of the population.

Application: The application of the cluster sampling method needs to be different from the stratified sampling method.

When a population is composed of several subgroups (or categories or levels) with natural boundaries and distinctions, and at the same time, different subgroups are very different from each other, and the differences within each subgroup are not large, the stratified sampling method is suitable. Conversely, when there is little difference between different subgroups and there is a large heterogeneity within each subgroup, the whole cluster sampling method is particularly suitable.

Stratified sampling is the method of randomly selecting samples (individuals) from different layers in a specified proportion from a population that can be divided into different subpopulations (or strata). The advantage of this method is that the sample is better representative and the sampling error is smaller. The disadvantage is that the sampling procedure is simpler and the random sampling is more complicated.

Cluster random sampling.

In areas where the survey units are sparsely distributed, or where the overall heterogeneity is high and it is difficult to establish a unified standard for stratification, the method of surveying several regions can only be used, which is the cluster random sampling method. When sampling, there should be commonalities between groups, such as population size, ethnic composition, etc.; However, there are differences within each group, and the targets of the survey are broader. Therefore, it is suitable to adopt the group random sampling method in which the selected population is randomly selected and then the selected population is surveyed.

Stratified sampling is to divide the sampling unit into different layers according to a certain characteristic or a certain rule, and then independently and randomly draw samples from different layers. Samples from each layer are combined to estimate the target amount of the population. The advantage is that the sample contains sampling units with various characteristics, and the structure of the sample is similar to that of the population, so that the accuracy of estimation can be effectively improved.

Cluster sampling, in which several units in a population are combined into groups, such groups are called clusters. The sampling method of sampling is called cluster sampling, and then all units in the selected group are surveyed. The advantage is that only the sampling frame of the group is required when the sample is taken, and it is not necessary to include the sampling frame of all units, which greatly simplifies the workload of compiling the sampling frame.

The difference between these two sampling methods is that the number of layers of stratified sampling is the sample capacity; The number of units in a cluster sample is the sample size.

1. The survey samples are different.

Cohort sampling is to divide the survey matrix into several groups, and then select several groups as survey samples by simple random sampling. Stratified sampling is a survey sample that randomly draws samples (individuals) from different strata in a specified proportion from a population that can be divided into different subpopulations (or strata).

2. The survey states have different methods.

Group sampling first divides the population into groups, then randomly selects several groups to form a sample, and finally investigates all the groups in the group. Stratified random sampling first divides the population into several categories according to a certain characteristic that has a greater impact on the observation index, and then randomly selects a certain number of observation units of silver seepage from each layer, and then composes a sample together, and finally investigates.

Systematic sampling, cluster sampling, and stratified sampling all belong to probability sampling. Scattered spine.

When the number of the population is relatively large, the population is first divided into several balanced parts, and then some individuals are selected from each part according to the predetermined rules to obtain the required sample.

In stratified sampling, the population is divided into layers that do not intersect with each other, and then a certain number of individuals are independently selected from each layer according to a certain proportion to obtain the required sample, which is suitable for the population to be composed of several parts with obvious differences.