(1) County Pumping Village. Taking the per capita net income of farmers in each village of a county (city) in recent three years as the relevant sign, queuing from low to high (or from high to low), taking the population of each village as the auxiliary data, accumulating in sequence according to the queuing order, making a queuing table (sampling box), and selecting 10 villages according to the principle of random starting point and symmetrical equidistant sampling. Calculate the sampling error and conduct a simple representative check, and the simple representative error should not exceed 3%.
Using the above method, we can compare multiple groups of samples and choose the best survey network.
(2) the village draws households. Queue with irrelevant signs, calculate the sampling distance, and select 10 households in each village by random starting point and symmetrical equidistant sampling method. After the survey households are selected, we should carefully check and evaluate the representativeness of the samples from the following aspects:
(1) income representative check. The per capita net income level of 10 households selected by each village should be basically similar to that of the village.
② Household structure representative inspection. The family structure of the surveyed households in the county 100 should be close to the local actual situation, and the proportion of rural cadres' families should not exceed 3%.
(3) Inspection of family production structure. The type structure of household production in the county 100 households should be basically similar to the actual situation in the county.
In order to solve the problems of survey households' boredom and sample aging, enhance the representativeness of sampling survey outlets, and reflect the rural social and economic situation more accurately and timely, the sample rotation period is set at 5 years, and the survey samples are re-selected after 5 years.
According to the principle of sampling survey, increasing samples is the only effective way to improve the representativeness of sampling survey under the premise of unchanged sampling methods. In 2004, in order to improve the representativeness of rural residents' income in different levels, different structures and different regions, our province expanded the sample of rural household survey on the basis of comprehensive and systematic analysis and calculation. The survey sample of farmers has expanded from 2,700 households in 27 cities and counties (cities, districts) to 4,700 households in 47 cities and counties (cities, districts). Thereby effectively ensuring a better representation of the province's overall data and income structure.