Online shopping survey method description

Considering that the randomness of telephone surveys is easier to guarantee and the accuracy is higher, this project is conducted by telephone survey in June 2019. The target audience is netizens who have bought online and bought things online.

In order to make the sample more representative, the sample design is considered from the urban level and the regional perspective, and the total design sample is 2,600. Analysis of factors such as the economic population of various cities in China and the reality of online shopping in China, it is believed that the current Internet shopping is still more economically developed cities. The sampling frame of this survey is divided into 4 municipalities and 15 sub-provincial cities, a total of 19 cities, from which 10 representative cities are selected for visit.

Beijing, Shanghai and Guangzhou, as three representative cities in the north and the south, are cities that are of great interest to all online shopping websites. These three cities are sampled separately, with 400 samples per city and 16 other cities as a whole. Sampling. In order to improve the sample representation of other cities, the district is divided into seven layers according to administrative divisions: North China, Northeast China, East China, Central China, South China, Southwest China, and Northwest China. Each region randomly selects one municipality or sub-provincial city, and each city sample is designed to be 200. The specific cities and sample distributions are as follows:

This sampling design satisfies the case where the confidence is 95% and the design effect is 1.8. The maximum allowable absolute error of the main data is less than 3%.

The survey method is to randomly generate telephone numbers based on all telephone numbers in each city, and access the residential telephones, PHS and dormitory telephones. Survey randomness and accuracy control methods:

1) Use each station number of each city to randomly generate a phone number for access.

2) In order to avoid the sampling deviation that the office workers may cause during the daytime work, the method of making a call at 18:00-21:00 on the working day and 9:00-21:00 on the weekend is adopted.

3) In order to avoid the impact of the connection rate on the randomness, if the number cannot be connected, take at least three times.

5) After the telephone survey is completed, the data is preprocessed, the logical relationship between the value of the variable and the variable is checked, and the unqualified sample is deleted.