• Journal of the Korean Statistical Society
• /
• v.35 no.4
• /
• pp.343-353
• /
• 2006

The two-way balanced one-level rotation design, $r_1^m-r_2^{m-1}$, and the three-way balanced multi-level rotation design, $r_1^m(\iota)-r_1^{m-1}$, were discussed (Park et al., 2001, 2003). Although these rotation designs enjoy balancing properties, they have a restriction of $r_2=c{\cdot}r_1$ (c should be a integer value) which interferes with applying these designs freely to various situations. To overcome this difficulty, we extend the $r_1^m(\iota)-r_1^{m-1}$ design to new one under the most general rotation system. The new multi-level rotation design also satisfies tree-way balancing which is done on interview time, rotation group and recall time. We present the rule and rotation algorithm which guarantee the three-way balancing. In particular, we specify the necessary condition for the extended three-way balanced multi-level rotation sampling design.

• Proceedings of the Korean Statistical Society Conference
• /
• 2002.05a
• /
• pp.19-24
• /
• 2002

The two-way balanced one-level rotation design has been discussed (Park, Kim and Choi, 2001), where the two-way balancing is done on interview time in monthly sample and rotation group. We extend it to three-way balanced multi-level design under the most general rotation system. The three-way balancing is accomplished on interview time not only in monthly sample and rotation group but also in recall time. We present the necessary condition and rotation algorithm which guarantee the three-way balancing. We propose multi-level composite estimators (MCE) from this design and derive their variances and mean squared errors (MSE), assuming the correlation from the measurements of the same sample unit and three types of biases in monthly sample.

• Journal of the Korean Statistical Society
• /
• v.32 no.2
• /
• pp.175-191
• /
• 2003

An empirical Bayesian approach is discussed for estimation of characteristics from the two-way balanced rotation sampling design which includes U.S. Current Population Survey and Canadian Labor Force Survey as special cases. An empirical Bayesian estimator is derived for monthly effect under presence of two types of biases and correlations It is shown that the marginal distribution of observation provides more general correlation structure than that frequentist has assumed. Consistent estimators are derived for hyper-parameters in Normal priors.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2002.11a
• /
• pp.255-274
• /
• 2002

We classify rotation sampling designs into two classes. The first class replaces sample units within the same rotation group while the second class replaces sample units between different rotation groups. The first class is specified by the three-way balanced design which is a multi-level version of previous balanced designs. We introduce an extended generalized composite estimator (EGCE) and derive its variance and mean squared error for each of the two classes of design, cooperating two types of correlations and three types of biases. Unbiased estimators are derived for difference between interview time biases, between recall time biases, and between rotation group biases. Using the variance and mean squared error, since any rotation design belongs to one of the two classes and the EGCE is a most general estimator for rotation design, we evaluate the efficiency of EGCE to simple weighted estimator and the effects of levels, design gaps, and rotation patterns on variance and mean squared error.

• Proceedings of the Korean Statistical Society Conference
• /
• 2003.10a
• /
• pp.123-130
• /
• 2003

We define a broad class of rotation designs whose monthly sample is balanced in interview time, level of recall, and rotation group, and whose rotation scheme is time-invariant. The necessary and sufficient conditions are obtained for such designs. Using these conditions, we derive a minimum variance unbiased generalized composite estimator (MVUGCE). To examine the existence of time-in-sample bias and recall bias, we also propose unbiased estimators and their variances. Numerical examples investigate the impacts of design gap, non-sampling error sources, and two types of correlations on the variance of MVUGCE.