Sort Help
entries

Results

All (21)

All (21) (0 to 10 of 21 results)

  • Articles and reports: 12-001-X199400214417
    Description:

    Without-replacement list sampling with probability proportional to some measure of element size has not enjoyed much application in forestry because of the difficulty of implementing such sample strategies, that have been termed \pi ps designs to distinguish without-replacement sampling from the well-known with-replacement pps designs. In this contribution, an exact \pi ps strategy (Sunter’s variant 2), an approximate \pi ps design (Sunter’s variant 1) and the Rao-Hartley-Cochran random group method are examined and the variances of the respective estimators for total bole volume are computed for four tree populations. The results indicate that compared to the Rao-Hartley-Cochran design Sunter’s variant 1 in general leads to higher precision if the relationship between auxiliary information x_k and target characteristic y_k is loose but is sensitive to the ordering of the sampling frame, whereas the Rao-Hartley-Cochran design does not require the sampling frame to be ordered at all and appears to be superior if strong linear relationships between x_k and y_k are present.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214418
    Description:

    We deal with the nonresponse problem by drawing on the model of selection in phases that was proposed by Särndal and Swenson (1987). To estimate response probabilities, we use the nonparametric approach first advanced by Giommi (1987). We define estimators according to the nonparametric estimation (NPE) model, and we study their general properties empirically. Inference is based on the concept of quasi-randomization (Oh and Scheuren 1983). The emphasis is on estimating the variance and constructing confidence intervals. We find, by way of a Monte Carlo study, that it is possible to improve the quality of the estimators considered by using a variant of the NPE approach. The latter also serves to confirm the performance of regression estimators in terms of variance estimation.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214419
    Description:

    The study was undertaken to evaluate some alternative small areas estimators to produce level estimates for unplanned domains from the Italian Labour Force Sample Survey. In our study, the small areas are the Health Service Areas, which are unplanned sub-regional territorial domains and were not isolated at the time of sample design and thus cut across boundaries of the design strata. We consider the following estimators: post-stratified ratio, synthetic, composite expressed as linear combination of synthetic and of post-stratified ratio, and sample size dependent. For all the estimators considered in this study, the average percent relative biases and the average relative mean square errors were obtained in a Monte Carlo study in which the sample design was simulated using data from the 1981 Italian Census.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214420
    Description:

    The statistical literature contains many methods for disclosure limitation in microdata. However, their use by statistical agencies and understanding of their properties and effects has been limited. For purposes of furthering research and use of these methods, and facilitating their evaluation and quality assurance, it would be desirable to formulate them within a single framework. A framework called matrix masking - based on ordinary matrix arithmetic - is presented, and explicit matrix mask formulations are given for the principal microdata disclosure limitation methods in current use. This enables improved understanding and implementation of these methods by statistical agencies and other practitioners.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214421
    Description:

    This paper discusses testing a single hypothesis about linear regression coefficients based on sample survey data. It suggests that when the design-based linearization variance estimator for a regression coefficient is used it should be adjusted to reduce its slight model bias and that a Satterthwaite-like estimation of its effective degrees of freedom be made. A very important special case of this analysis is its application to domain means.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214422
    Description:

    Dual system estimation (DSE) has been used since 1950 by the U.S. Bureau of Census for coverage evaluation of the decennial census. In the DSE approach, data from a sample is combined with data from the census to estimate census undercount and overcount. DSE relies upon the assumption that individuals in both the census and the sample can be matched perfectly. The unavoidable mismatches and erroneous nonmatches reduce the accuracy of the DSE. This paper reconsiders the DSE approach by relaxing the perfect matching assumption and proposes models to describe two types of matching errors, false matches of nonmatching cases and false nonmatches of matching cases. Methods for estimating population total and census undercount are presented and illustrated using data from 1986 Los Angeles test census and 1990 Decennial Census.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214423
    Description:

    Most surveys suffer from the problem of missing data caused by nonresponse. To deal with this problem, imputation is often used to create a “completed data set”, that is, a data set composed of actual observations (for the respondents) and imputations (for the nonrespondents). Usually, imputation is carried out under the assumption of unconfounded response mechanism. When this assumption does not hold, a bias is introduced in the standard estimator of the population mean calculated from the completed data set. In this paper, we pursue the idea of using simple correction factors for the bias problem in the case that ratio imputation is used. The effectiveness of the correction factors is studied by Monte Carlo simulation using artificially generated data sets representing various super-populations, nonresponse rates, nonresponse mechanisms, and correlations between the variable of interest and the auxiliary variable. These correction factors are found to be effective especially when the population follows the model underlying ratio imputation. An option for estimating the variance of the corrected point estimates is also discussed.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214424
    Description:

    This paper provides an overview of important considerations that should be taken into account when developing and designing questionnaires for business surveys. These considerations include the determination of objectives and data requirements, consultation with data users and respondents, and methods for testing questionnaires. In developing and designing business survey questionnaires, focus groups and cognitive research methods help the researcher to identify potential sources of measurement error and to understand the response process that respondents go through in completing the questionnaires. Examples of focus groups and cognitive research undertaken by Statistics Canada are provided.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214425
    Description:

    We present a formal model based sampling solution to the problem of estimating list frame size based on capture-recapture sampling which has been widely used for animal populations and for adjusting the US census. For two incomplete lists it is easy to estimate total frame size using the Lincoln-Petersen estimator. This estimator is model based with a key assumption being independence of the two lists. Once an estimator of the population (frame) size has been obtained it is possible to obtain an estimator of a population total for some characteristic if a sample of units has that characteristic measured. A discussion of the properties of this estimator will be presented. An example where the establishments are fishing boats taking part in an ocean fishery off the Atlantic Coast of the United States is presented. Estimation of frame size and then population totals using a capture-recapture model is likely to have broad application in establishment surveys due to practicality and cost savings but possible biases due to assumption violations need to be considered.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214426
    Description:

    In the MARS Project (Monitoring Agriculture with Remote Sensing) of the E.C. (European Community), area frames based on a square grid are used for area estimation through ground surveys and high resolution satellite images. These satellite images are useful, though expensive, for area estimation: their use for yield estimation is not yet operational. To fill this gap the sample elements (segments) of the area survey are used as well for sampling farms with a template of points overlaid on the segment. Most often we use a fixed number of points per segment. Farmers are asked to provide global data for the farm, and estimates are computed with a Horvitz-Thompson approach. Major problems include locating farmers and checking for misunderstanding of instructions. Good results are obtained for area and for production of the main crops. Area frames need to be complemented with list frames (multiple frames) to give reliable estimates for livestock.

    Release date: 1994-12-15
Stats in brief (0)

Stats in brief (0) (0 results)

No content available at this time.

Articles and reports (21)

Articles and reports (21) (0 to 10 of 21 results)

  • Articles and reports: 12-001-X199400214417
    Description:

    Without-replacement list sampling with probability proportional to some measure of element size has not enjoyed much application in forestry because of the difficulty of implementing such sample strategies, that have been termed \pi ps designs to distinguish without-replacement sampling from the well-known with-replacement pps designs. In this contribution, an exact \pi ps strategy (Sunter’s variant 2), an approximate \pi ps design (Sunter’s variant 1) and the Rao-Hartley-Cochran random group method are examined and the variances of the respective estimators for total bole volume are computed for four tree populations. The results indicate that compared to the Rao-Hartley-Cochran design Sunter’s variant 1 in general leads to higher precision if the relationship between auxiliary information x_k and target characteristic y_k is loose but is sensitive to the ordering of the sampling frame, whereas the Rao-Hartley-Cochran design does not require the sampling frame to be ordered at all and appears to be superior if strong linear relationships between x_k and y_k are present.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214418
    Description:

    We deal with the nonresponse problem by drawing on the model of selection in phases that was proposed by Särndal and Swenson (1987). To estimate response probabilities, we use the nonparametric approach first advanced by Giommi (1987). We define estimators according to the nonparametric estimation (NPE) model, and we study their general properties empirically. Inference is based on the concept of quasi-randomization (Oh and Scheuren 1983). The emphasis is on estimating the variance and constructing confidence intervals. We find, by way of a Monte Carlo study, that it is possible to improve the quality of the estimators considered by using a variant of the NPE approach. The latter also serves to confirm the performance of regression estimators in terms of variance estimation.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214419
    Description:

    The study was undertaken to evaluate some alternative small areas estimators to produce level estimates for unplanned domains from the Italian Labour Force Sample Survey. In our study, the small areas are the Health Service Areas, which are unplanned sub-regional territorial domains and were not isolated at the time of sample design and thus cut across boundaries of the design strata. We consider the following estimators: post-stratified ratio, synthetic, composite expressed as linear combination of synthetic and of post-stratified ratio, and sample size dependent. For all the estimators considered in this study, the average percent relative biases and the average relative mean square errors were obtained in a Monte Carlo study in which the sample design was simulated using data from the 1981 Italian Census.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214420
    Description:

    The statistical literature contains many methods for disclosure limitation in microdata. However, their use by statistical agencies and understanding of their properties and effects has been limited. For purposes of furthering research and use of these methods, and facilitating their evaluation and quality assurance, it would be desirable to formulate them within a single framework. A framework called matrix masking - based on ordinary matrix arithmetic - is presented, and explicit matrix mask formulations are given for the principal microdata disclosure limitation methods in current use. This enables improved understanding and implementation of these methods by statistical agencies and other practitioners.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214421
    Description:

    This paper discusses testing a single hypothesis about linear regression coefficients based on sample survey data. It suggests that when the design-based linearization variance estimator for a regression coefficient is used it should be adjusted to reduce its slight model bias and that a Satterthwaite-like estimation of its effective degrees of freedom be made. A very important special case of this analysis is its application to domain means.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214422
    Description:

    Dual system estimation (DSE) has been used since 1950 by the U.S. Bureau of Census for coverage evaluation of the decennial census. In the DSE approach, data from a sample is combined with data from the census to estimate census undercount and overcount. DSE relies upon the assumption that individuals in both the census and the sample can be matched perfectly. The unavoidable mismatches and erroneous nonmatches reduce the accuracy of the DSE. This paper reconsiders the DSE approach by relaxing the perfect matching assumption and proposes models to describe two types of matching errors, false matches of nonmatching cases and false nonmatches of matching cases. Methods for estimating population total and census undercount are presented and illustrated using data from 1986 Los Angeles test census and 1990 Decennial Census.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214423
    Description:

    Most surveys suffer from the problem of missing data caused by nonresponse. To deal with this problem, imputation is often used to create a “completed data set”, that is, a data set composed of actual observations (for the respondents) and imputations (for the nonrespondents). Usually, imputation is carried out under the assumption of unconfounded response mechanism. When this assumption does not hold, a bias is introduced in the standard estimator of the population mean calculated from the completed data set. In this paper, we pursue the idea of using simple correction factors for the bias problem in the case that ratio imputation is used. The effectiveness of the correction factors is studied by Monte Carlo simulation using artificially generated data sets representing various super-populations, nonresponse rates, nonresponse mechanisms, and correlations between the variable of interest and the auxiliary variable. These correction factors are found to be effective especially when the population follows the model underlying ratio imputation. An option for estimating the variance of the corrected point estimates is also discussed.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214424
    Description:

    This paper provides an overview of important considerations that should be taken into account when developing and designing questionnaires for business surveys. These considerations include the determination of objectives and data requirements, consultation with data users and respondents, and methods for testing questionnaires. In developing and designing business survey questionnaires, focus groups and cognitive research methods help the researcher to identify potential sources of measurement error and to understand the response process that respondents go through in completing the questionnaires. Examples of focus groups and cognitive research undertaken by Statistics Canada are provided.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214425
    Description:

    We present a formal model based sampling solution to the problem of estimating list frame size based on capture-recapture sampling which has been widely used for animal populations and for adjusting the US census. For two incomplete lists it is easy to estimate total frame size using the Lincoln-Petersen estimator. This estimator is model based with a key assumption being independence of the two lists. Once an estimator of the population (frame) size has been obtained it is possible to obtain an estimator of a population total for some characteristic if a sample of units has that characteristic measured. A discussion of the properties of this estimator will be presented. An example where the establishments are fishing boats taking part in an ocean fishery off the Atlantic Coast of the United States is presented. Estimation of frame size and then population totals using a capture-recapture model is likely to have broad application in establishment surveys due to practicality and cost savings but possible biases due to assumption violations need to be considered.

    Release date: 1994-12-15

  • Articles and reports: 12-001-X199400214426
    Description:

    In the MARS Project (Monitoring Agriculture with Remote Sensing) of the E.C. (European Community), area frames based on a square grid are used for area estimation through ground surveys and high resolution satellite images. These satellite images are useful, though expensive, for area estimation: their use for yield estimation is not yet operational. To fill this gap the sample elements (segments) of the area survey are used as well for sampling farms with a template of points overlaid on the segment. Most often we use a fixed number of points per segment. Farmers are asked to provide global data for the farm, and estimates are computed with a Horvitz-Thompson approach. Major problems include locating farmers and checking for misunderstanding of instructions. Good results are obtained for area and for production of the main crops. Area frames need to be complemented with list frames (multiple frames) to give reliable estimates for livestock.

    Release date: 1994-12-15
Journals and periodicals (0)

Journals and periodicals (0) (0 results)

No content available at this time.

Date modified: