Survey design

All (2) ((2 results))

• 1. A moving stratification algorithm Archived

  Articles and reports: 12-001-X199600114382

  Description:

  A general algorithm with equal probabilities is presented. The author provides the second order inclusion probabilities that correspond to the algorithm, which generalizes the selection-rejection method, so that a sample may be drawn using simple random sampling without replacement. Another particular case of the algorithm, called moving stratification algorithm, is discussed. A smooth stratification effect can be obtained by using, as a stratification variable, the serial number of the observation units. The author provides approximations of first and second order inclusion probabilities. These approximations lead to a population mean estimator and to an estimator of the variance of this mean estimator. The algorithm is then compared to a classical stratified plan with proportional allocation.

  Release date: 1996-06-14

  ---

• 2. Applying the Lavallée and Hidiroglou method to obtain stratification boundaries for the Census Bureau’s Annual Capital Expenditures Survey Archived

  Articles and reports: 12-001-X199600114384

  Description:

  The Lavallée-Hidiroglou (L-H) method of finding stratification boundaries has been used in the Census Bureau’s Annual Capital Expenditures Survey (ACES) to stratify part of its universe in the pilot study and the subsequent preliminary survey. This iterative method minimizes the sample size while fixing the desired reliability level by constructing appropriate boundary points. However, we encountered two problems in our application. One problem was that different starting boundaries resulted in different ending boundaries. The other problem was that the convergence to locally-optimal boundaries was slow, i.e., the number of iterations was large and convergence was not guaranteed. This paper addresses our difficulties with the L-H method and shows how they were resolved so that this procedure would work well for the ACES. In particular, we describe how contour plots were constructed and used to help illustrate how insignificant these problems were once the L-H method was applied. This paper describes revisions made to the L-H method; revisions that made it a practical method of finding stratification boundaries for ACES.

  Release date: 1996-06-14

Data (0) (0 results)

No content available at this time.

Analysis (2) ((2 results))

• 1. A moving stratification algorithm Archived

  Articles and reports: 12-001-X199600114382

  Description:

  A general algorithm with equal probabilities is presented. The author provides the second order inclusion probabilities that correspond to the algorithm, which generalizes the selection-rejection method, so that a sample may be drawn using simple random sampling without replacement. Another particular case of the algorithm, called moving stratification algorithm, is discussed. A smooth stratification effect can be obtained by using, as a stratification variable, the serial number of the observation units. The author provides approximations of first and second order inclusion probabilities. These approximations lead to a population mean estimator and to an estimator of the variance of this mean estimator. The algorithm is then compared to a classical stratified plan with proportional allocation.

  Release date: 1996-06-14

  ---

• 2. Applying the Lavallée and Hidiroglou method to obtain stratification boundaries for the Census Bureau’s Annual Capital Expenditures Survey Archived

  Articles and reports: 12-001-X199600114384

  Description:

  The Lavallée-Hidiroglou (L-H) method of finding stratification boundaries has been used in the Census Bureau’s Annual Capital Expenditures Survey (ACES) to stratify part of its universe in the pilot study and the subsequent preliminary survey. This iterative method minimizes the sample size while fixing the desired reliability level by constructing appropriate boundary points. However, we encountered two problems in our application. One problem was that different starting boundaries resulted in different ending boundaries. The other problem was that the convergence to locally-optimal boundaries was slow, i.e., the number of iterations was large and convergence was not guaranteed. This paper addresses our difficulties with the L-H method and shows how they were resolved so that this procedure would work well for the ACES. In particular, we describe how contour plots were constructed and used to help illustrate how insignificant these problems were once the L-H method was applied. This paper describes revisions made to the L-H method; revisions that made it a practical method of finding stratification boundaries for ACES.

  Release date: 1996-06-14

Reference (0) (0 results)

No content available at this time.