農林漁牧普查是台灣定期進行的重要調查之一，然而執行全國性的普查需要龐大的調查成本，因此這項調查每五年才會執行一次，但相關的農業調查必需每年都執行才能獲得有關產業的情況及資訊，同時才能適時的制定恰當的政策，本研究就是利用2010年農林漁牧普查的資料建構出台灣主力農家的抽樣策略，主力農家的定義為全年毛收入介於20萬元至5千萬元之間的農牧家戶且家戶內至少有一位65歲以下的人從事農牧業工作。在抽樣策略的設計階段，使用的是分層隨機抽樣，其中包含分層點以及最適樣本配置，除此之外，在推論階段使用輔助變數以得到更精確的估計，模擬的結果顯示輔助變數能在上述的分層設計下有效的增進推估精確度。除了分層變數定義的次母體之外，非分層變數定義的次母體在實用上也可能產生興趣，因此在本研究中也會討論如何推估未定義之次母體及其相關議題。

The census of Agriculture, Forestry, Fishery and Animal Husbandry is one of the important surveys conducted regularly in Taiwan. However, enormous amount of survey cost is required for such a nationwide census, hence it can be proceeded every five years. Therefore, it would be necessary to conduct certain agriculture survey annually for current industry situation and information, so that proper policy can be formulated timely. A sampling strategy for the Taiwanese primary farm household survey is constructed based on the 2010 census data in this research. Primary farm households is defined as the farm households whose gross income per year is higher than 200 thousand dollars and lower than 50 million dollars and at least one household member engaging in the agriculture production is under 65 years old. In the design stage of the sampling strategy, a stratified random sampling with optimal allocation and stratum boundary is constructed. In addition, auxiliary information is utilized in the inference stage for more efficient estimation results. Simulation result shows that auxiliary information in inference stage can effectively improve the estimation precision under the proposed stratified sampling design. Besides the subpopluation defined by the stratification variable, subpopulation not defined by stratification variable may be of interest in practice as well. Hence, estimation method of the subpopulation not defined by stratification variable and the related issues are studied as well in this research.

[1] Cochran, W.G., 1977, Sampling Technique. 3rd edition, John Wiley & Sons, New York
[2] Dalenius, T., 1957, Sampling in Sweden: Contributions to the methods and theories of sample survey practice. Almqvist & Wicksell, Stockholm, Sweden
[3] Dalenius, T. and Hodges, J., 1959, Minimum variance stratification. Journal of American Statistical Association, Vol.54, No.285, pp.88-101
[4] Gunning, P. and Horgan, J. M., 2004. A new algorithm for the construction of stratum boundaries in skewed populations. Survey Methodology, Vol.30, pp.159-166
[5] Kozak, M., 2004, Optimal stratification using random search method in agricultural surveys. Statistics in Transitions, Vol.6, pp.797-806
[6] Lavall´ee, P. and Hidiroglou, M., 1988, On the stratification of skewed population. Survey Methodology, Vol.14, pp.33-43
[7] Lednicki, B. andWieczorkowski, R., 2003. Optimal stratification and sample allocation between subpopulations and strata. Statistics in Transitions, Vol.6, pp.287-306
[8] Rivest, L.-P., 2002, A generalization of the Lavall´ee and Hidiroglou Algorithm for stratification in business surveys. Survey Methodology, Vol.28, pp.191-198
[9] Lohr, Sharon L., 2010, Sampling: Design and Analysis. 2nd edition, Brooks/Cole, Australia, Boston, Mass
[10] Thompson, S.K., 2012, Sampling. 3rd edition, John Wiley & Sons, Hoboken , NJ