Sampling

Sampling a survey of all potential consumers (called population) of company products is unrealistic, time consuming, and expensive. However, companies still need to collect enough critical data to clearly understand consumer's viewpoint. This can be done by collecting demographic samples. This sample group consists of consumers representing all potential purchasers of products. You can select samples in various ways.

Survey samples can be roughly divided into two types, probability samples and non-probability samples. A probability-based sample implements a sampling plan with a specified probability (possibly the adaptive probability specified by the adaptation process). Probability - based sampling allows design - based reasoning about the target population. Inference is based on the known objective probability distribution specified in the research protocol. Inferences from probability-based surveys can still be affected by many types of bias.

Sampling methods can be divided probabilistically or non-stochastically. In probability samples, each member of the population has a known nonzero probability that is chosen. Probabilistic methods include random sampling, systematic extraction, and stratified extraction. For non-stochastic sampling, members are chosen randomly from the group. These include sampling acceleration, sampling judgment, quarter sampling, and snowball sampling. The advantage of stochastic sampling is that you can calculate the sampling error. Sampling error is the extent to which the sample may differ from the population. When estimating a population, the results of the report are added or subtracted from the sampling error. With non-stochastic sampling, the degree of difference between sample and population is still unknown.

The sampling error is one of two reasons for the difference between population parameter estimates and population parameter actual values ​​but unknown values. Another reason is non-sampling error. Even though the sampling process has no non-sampling errors, estimates from different random samples (same size) will vary from sample to sample, and each estimate may be different from the actual value of the population parameter. Students need many exercises to identify the cause of potential errors in their work and criticize the report. In addition, I discovered that the truth issue is very effective when practicing proper use of these terms. Whether students consciously decide which term to use over a period of time helps to understand the concept and to recognize it. Then a strange term no longer has its original chaotic meaning