Estimation in Survey Sampling: Robustness and Optimality
本文在统一理论框架下，将效率与稳健性整合为单一最优性准则，通过扩展模型估计范围并利用适当抽样设计，结合数值研究讨论了调查实践中的常见情形。
Abstract It is shown how the criteria of efficiency and robustness can be integrated fruitfully in a single optimality criterion within the framework of the Unified Theory (Godambe 1955). This is achieved by extending the scope of the usual model-based estimation using appropriate sampling designs. Some situations of common occurrence in survey practice are discussed with a numerical study. Key Words: RobustnessPurposive samplingIndividual labelUnbiased minimum variance estimationNonexistence theoremsSuperpopulation modelDepartures from the model