순환표본에 기초한 기간모평균 추정량의 통계적 성질 = Statistical Properties of the Period Mean Estimator Based on Rolling Samples

다국어 초록 (Multilingual Abstract)

In a rolling sample survey, whereas point-in-time estimation is made based on each rolling sample, period estimation is based on the total sample combined by rolling samples. Different from point-in-time estimation, period estimation may be affected by time, so we should pay attention to time-effect in period estimation. In the paper, we investigated the concept of period mean estimation and considerations of survey variables in period estimation. If a survey variable is related with survey-time, accumulated survey- time, or currency-price then those effects should be considered in period mean estimation process in order to avoid misleading inference. In addition, when there is a geographical change in the estimation area, it is necessary for such a change to be considered. The period mean estimator has higher precision but less currency than the point-in-time mean estimator and two estimators are not comparable in the time-trend population. We hope that illustration of the properties of the period mean estimator in the paper may help giving higher understanding of period estimation in rolling surveys.