TY - JOUR
AU - Rennolls, Keith
AB - The design of a complex sample survey is a difficult task, requiring considerable expertise and experience. Most experienced survey statisticians will try to keep things simple: stick to ‘epsem’ designs if possible; in multi-objective studies, try to define primary precision objectives for design purposes; consider only the more robust and simpler estimators; examine if the resource constraints determine the sample sizes, without the need to consider optimal allocation questions. However, the practical survey statistician will sometimes be forced into a more complex situation where he must attempt to determine optimal sample sizes. This paper considers what computer software is available to assist the survey statistician in his design efforts. The answer is ‘essentially none‘. It then considers what strategies should be used in the design of such computer software, and the theoretical framework within which the algorithms should be developed.
TI - Strategic Frameworks for a Sample Survey Design Package
JF - Journal of the Royal Statistical Society Series D: The Statistician
DO - 10.2307/2348281
DA - 2018-12-05
UR - https://www.deepdyve.com/lp/oxford-university-press/strategic-frameworks-for-a-sample-survey-design-package-PbA4DMuRqH
SP - 285
EP - 290
VL - 40
IS - 3
DP - DeepDyve
ER -