Author Information

Paavo Väisänen

Abstract

The Finnish Agricultural Income Statistics. published yearly by Statistics Finland, are based on a survey in which the data are collected from the farms in connection with taxation. The sampling design is stratified simple random sampling, in which Neyman allocation is used to calculate the sample sizes for the strata. The Farm Register is used as the sampling frame where variables such as region, production sector and arable land are available for stratification. The total incomes of farms from the previous survey serve as the allocation variable. Stratification and Neyman allocation rendered the estimates of most income variables more effective When measured by the design effect (DEFF) values which ranged from 0.3 to 0.7. Ratio estimation was studied by using arable land as an auxiliary variable. The sample was also evaluated by calculating estimates for variables available from administrative records and by companng them with the true values. The estimated values were systematically bigger than the true values. Non-response among small farms was one reason for this systematic error. A comparison by production sector revealed that the biggest differences were in cattle farming and in the production of cereals. An examination of the correlations in these sectors revealed a linear dependence between the survey and auxiliary variables. Ratio estimation was used in these sectors to reduce the error of estimates and to balance the variables known from other sources.

Keywords

Income of agriculture, multipurpose survey, ratio estimation, stratified sampling

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
Apr 23rd, 4:00 PM

SAMPLE DESIGN IN THE FINNISH AGRICULTURAL INCOME STATISTICS

The Finnish Agricultural Income Statistics. published yearly by Statistics Finland, are based on a survey in which the data are collected from the farms in connection with taxation. The sampling design is stratified simple random sampling, in which Neyman allocation is used to calculate the sample sizes for the strata. The Farm Register is used as the sampling frame where variables such as region, production sector and arable land are available for stratification. The total incomes of farms from the previous survey serve as the allocation variable. Stratification and Neyman allocation rendered the estimates of most income variables more effective When measured by the design effect (DEFF) values which ranged from 0.3 to 0.7. Ratio estimation was studied by using arable land as an auxiliary variable. The sample was also evaluated by calculating estimates for variables available from administrative records and by companng them with the true values. The estimated values were systematically bigger than the true values. Non-response among small farms was one reason for this systematic error. A comparison by production sector revealed that the biggest differences were in cattle farming and in the production of cereals. An examination of the correlations in these sectors revealed a linear dependence between the survey and auxiliary variables. Ratio estimation was used in these sectors to reduce the error of estimates and to balance the variables known from other sources.