Statistics by subject – Statistical methods

Other available resources to support your research.

Help for sorting results
Browse our central repository of key standard concepts, definitions, data sources and methods.
Loading
Loading in progress, please wait...
All (18)

All (18) (18 of 18 results)

  • Articles and reports: 12-001-X198700214605
    Description:

    The CPS uses raking ratio estimation in post-stratification estimation to adjust sample estimates of population to census-based estimates of the population. An alternative procedure, using generalized least squares, is compared to the current procedure.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214517
    Description:

    Because it is easy to use, the “AGEVEN” record makes it possible to date events more precisely and to classify retrospectively demographic events (births and deaths), changes in marital status and changes in place of residence. The data collected are used to accurately recreate the socio-economic conditions that were present when the demographic events being studied took place.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214607
    Description:

    Household surveys generally use separate procedures for estimating characteristics of persons and those of families. An integrated procedure is proposed and a least-squares estimator introduced to achieve this end. The estimator is shown to be unbiased under certain general conditions. Using data from the Canadian Labour Force Survey, variances for the estimator are calculated and shown to compare favourably to those from current procedures.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214514
    Description:

    This paper deals with the nonresponse problem in the estimation of the mean of a finite population, following an approach closely related to that of Cassel, Särndal and Wretman (1983). Two very simple methods are proposed for estimating the individual response probabilities; these are then used, in connection with a superpopulation model, to construct estimators for the population mean. A first evaluation of the properties of the proposed methods is given by a Monte Carlo experiment. The results shed some light on their effectiveness.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214516
    Description:

    The biases and stabilities of alternative variance estimators for the two stage random group design (Rao et al. 1962) are evaluated in a Monte Carlo study in the context of Canadian Labour Force Survey. The variance formula for raking ratio estimation procedure is derived using Taylor linearization method. The properties of the variance formula are investigated by a Monte Carlo simulation.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214606
    Description:

    A class of “constrained minimum distance” methods is considered for constraining household weights to be consistent with auxiliary information on the number of persons in various age x race x sex cells. The constrained weights are as close as possible to the initial weights based on the inverse probability of selection. This class of methods includes raking and generalized least square methods, as well as multinomial maximum likelihood, (where the cells of the distribution are household types.) The properties of the methods in the presence of systematic undercoverage of the household types are studied through some simple models for coverage. Comparisons with the principal person method are made and the paper concludes with the observation that it is necessary to know more about the nature of survey undercoverage before deciding on which of the constrained minimum distance or principal person methods is to be preferred in applications.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214515
    Description:

    Methods are given to estimate functions of the cell probabilities associated with a table of multinomial data that has been randomly rounded to multiples of a given number, say l. We show that: (i) random rounding causes only second order effects on bias and variance; (ii) the loss of efficiency in using the natural estimates of cell probability is negligible provided that the cell entry is large compared with (l^2 - 1) / (6R) where R is the number of cells in the table; and (iii) estimates of apparently exponentially small bias are available for moments of these natural estimates and for polynomials in the cell probabilities.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214609
    Description:

    The Hansen-Hurwitz (1943) strategy is known to be inferior to the Horvitz-Thompson (1952) strategy associated with a number of IPPS (inclusion probability proportional to size) sampling procedures. The present paper presents a simpler proof of these results and therefore has some pedagogic interest.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214608
    Description:

    A hybrid technique is described that employs both conventional and raking ratio estimation to handle the case when the population frequencies N_ij in a two-dimensional table are known, but some of the observed frequencies n_ij are small (or zero). Results are provided on the approach taken as it has evolved in the Corporate Statistics of Income Program over the last several years. Changes are still being considered and these will be discussed as well.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700114513
    Description:

    Singh and Srivastava (1973) proposed a linear unbiased estimator of the population mean when sampling on successive occasions using several auxiliary variables whose known population means remain unchanged for all occasions. In this paper, three composite estimators T_1, T_2 and T_3, each utilising an auxiliary variable whose known population mean changes from one occasion to the next, are presented for the estimation of the current population total. The proposed estimators are compared with the ordinary estimator, T_0, and the usual successive sampling estimator, T \prime, of the current population total without the use of auxiliary information. We find that using auxiliary information in conjunction with successive sampling does not always uniformly produce a gain in efficiency over T_0 or T \prime. However, when applied to a survey of teak plantations to estimate the mean height of teak trees, T_1, T_2 and T_3 proved more efficient than T_0 and T \prime.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114466
    Description:

    This paper presents results from methodological experiments comparing telephone and face-to-face interviewing in surveys of the general population. The relatively low level of telephone ownership in the United Kingdom, especially among the less privileged, argues the need for a dual-mode approach combining telephone interviews with face-to-face interviews for those without telephones. This approach depends on the absence of differential mode-effects on the answers obtained or on the ability to account for these effects when they occur.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114468
    Description:

    The National Agricultural Statistics Service, U.S. Department of Agriculture, conducts yield surveys for a variety of field crops in the United States. While field sampling procedures for various crops differ, the same basic survey design is used for all crops. The survey design and current estimators are reviewed. Alternative estimators of yield and production and of the variance of the estimators are presented. Current estimators and alternative estimators are compared, both theoretically and in a Monte Carlo simulation.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114511
    Description:

    A new unequal probability sampling scheme for selecting n(> 2) units without replacement from a finite population is proposed. This scheme ensures that the inclusion probabilities are proportional to sizes. It has the advantage of simplicity in selection and estimation and also provides a non-negative variance estimator. The variance of the Horvitz-Thompson (H-T) estimator under the proposed scheme is shown to be smaller than that of the customary estimator in probability proportional to size sampling with replacement. The proposed scheme also compares favourably with the without replacement scheme suggested by Sampford (1967) in an empirical study on a few natural populations.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114510
    Description:

    The method of minimum Q^(T) estimation for complex survey designs proposed by Singh (1985) provides asymptotically efficient estimates of model parameters analogous to Neyman’s (1949) min X^2 estimation procedure for simple random samples. The Q^(T) can be viewed as a X^2 type statistic for categorical survey data, and min Q^(T) estimates provide a robust alternative to Weighted Least Squares estimates, which often display unstable behaviour for complex surveys. In this paper, the min Q^(T) method is first described and then illustrated for the problem of estimating parameters of a logit model for survey estimates of unemployment rates which are obtained from the October 1980 Canadian LFS data cross-classified according to age-education covariate categories. It is seen that the trace efficiency of smoothed estimates obtained by Kumar and Rao (1986), who applied the method of pseudo maximum likelihood estimates (pseudo mle) to the same problem can be slightly improved by the min Q^(T) method. Interestingly enough, pseudo mle for individual cells behave much the same way as the efficient min Q^(T) estimates for the particular LFS example.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114509
    Description:

    This paper discusses three problems that have been a major preoccupation among researchers and practitioners of seasonal adjustment in statistical bureaus for the last ten years. These problems are: (l) the use of concurrent seasonal factors versus seasonal factor forecasts for current seasonal adjustment; (2) finding an optimal pattern of revisions for series seasonally adjusted with concurrent factors; and (3) smoothing highly irregular seasonally adjusted data.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114467
    Description:

    Demands for statistics on all aspects of our lives, our society and our economy continue to grow. At the same time statistical agencies share with many respondents a growing concern over the mounting burden of response to surveys. One result of the search for alternative methods of satisfying statistical demands has been an increased emphasis on the use of administrative records for statistical purposes. This paper reviews recent experience at Statistics Canada in this area and discusses obstacles to the greater use of administrative records. Approaches to rendering administrative systems more useful for statistical purposes are reviewed, together with some important concerns related to information protection and record linkage.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114512
    Description:

    The Health and Activity Limitation Survey is part of the program to establish a data base on the disabled population in Canada. The sample design used for the part of the survey covering the population not living in institutions is described. In addition, the methods used to determine the sizes of the samples and to select the samples are presented.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114465
    Description:

    The two-stage rejection rule telephone sample design described by Waksberg (1978) is modified to improve the efficiency of telephone surveys of the U.S. Black population. Experimental tests of sample design alternatives demonstrate that: a) use of rough stratification based on telephone exchange names and states; b) use of large cluster definitions (200 and 400 consecutive numbers) at the first stage; and c) rejection rules based on racial status of the household combine to offer improvements in the relative precision of a sample, given fixed resources. Cost and error models are examined to simulate design alternatives.

    Release date: 1987-06-15

Data (0)

Data (0) (0 results)

Your search for "" found no results in this section of the site.

You may try:

Analysis (18)

Analysis (18) (18 of 18 results)

  • Articles and reports: 12-001-X198700214605
    Description:

    The CPS uses raking ratio estimation in post-stratification estimation to adjust sample estimates of population to census-based estimates of the population. An alternative procedure, using generalized least squares, is compared to the current procedure.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214517
    Description:

    Because it is easy to use, the “AGEVEN” record makes it possible to date events more precisely and to classify retrospectively demographic events (births and deaths), changes in marital status and changes in place of residence. The data collected are used to accurately recreate the socio-economic conditions that were present when the demographic events being studied took place.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214607
    Description:

    Household surveys generally use separate procedures for estimating characteristics of persons and those of families. An integrated procedure is proposed and a least-squares estimator introduced to achieve this end. The estimator is shown to be unbiased under certain general conditions. Using data from the Canadian Labour Force Survey, variances for the estimator are calculated and shown to compare favourably to those from current procedures.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214514
    Description:

    This paper deals with the nonresponse problem in the estimation of the mean of a finite population, following an approach closely related to that of Cassel, Särndal and Wretman (1983). Two very simple methods are proposed for estimating the individual response probabilities; these are then used, in connection with a superpopulation model, to construct estimators for the population mean. A first evaluation of the properties of the proposed methods is given by a Monte Carlo experiment. The results shed some light on their effectiveness.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214516
    Description:

    The biases and stabilities of alternative variance estimators for the two stage random group design (Rao et al. 1962) are evaluated in a Monte Carlo study in the context of Canadian Labour Force Survey. The variance formula for raking ratio estimation procedure is derived using Taylor linearization method. The properties of the variance formula are investigated by a Monte Carlo simulation.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214606
    Description:

    A class of “constrained minimum distance” methods is considered for constraining household weights to be consistent with auxiliary information on the number of persons in various age x race x sex cells. The constrained weights are as close as possible to the initial weights based on the inverse probability of selection. This class of methods includes raking and generalized least square methods, as well as multinomial maximum likelihood, (where the cells of the distribution are household types.) The properties of the methods in the presence of systematic undercoverage of the household types are studied through some simple models for coverage. Comparisons with the principal person method are made and the paper concludes with the observation that it is necessary to know more about the nature of survey undercoverage before deciding on which of the constrained minimum distance or principal person methods is to be preferred in applications.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214515
    Description:

    Methods are given to estimate functions of the cell probabilities associated with a table of multinomial data that has been randomly rounded to multiples of a given number, say l. We show that: (i) random rounding causes only second order effects on bias and variance; (ii) the loss of efficiency in using the natural estimates of cell probability is negligible provided that the cell entry is large compared with (l^2 - 1) / (6R) where R is the number of cells in the table; and (iii) estimates of apparently exponentially small bias are available for moments of these natural estimates and for polynomials in the cell probabilities.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214609
    Description:

    The Hansen-Hurwitz (1943) strategy is known to be inferior to the Horvitz-Thompson (1952) strategy associated with a number of IPPS (inclusion probability proportional to size) sampling procedures. The present paper presents a simpler proof of these results and therefore has some pedagogic interest.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700214608
    Description:

    A hybrid technique is described that employs both conventional and raking ratio estimation to handle the case when the population frequencies N_ij in a two-dimensional table are known, but some of the observed frequencies n_ij are small (or zero). Results are provided on the approach taken as it has evolved in the Corporate Statistics of Income Program over the last several years. Changes are still being considered and these will be discussed as well.

    Release date: 1987-12-15

  • Articles and reports: 12-001-X198700114513
    Description:

    Singh and Srivastava (1973) proposed a linear unbiased estimator of the population mean when sampling on successive occasions using several auxiliary variables whose known population means remain unchanged for all occasions. In this paper, three composite estimators T_1, T_2 and T_3, each utilising an auxiliary variable whose known population mean changes from one occasion to the next, are presented for the estimation of the current population total. The proposed estimators are compared with the ordinary estimator, T_0, and the usual successive sampling estimator, T \prime, of the current population total without the use of auxiliary information. We find that using auxiliary information in conjunction with successive sampling does not always uniformly produce a gain in efficiency over T_0 or T \prime. However, when applied to a survey of teak plantations to estimate the mean height of teak trees, T_1, T_2 and T_3 proved more efficient than T_0 and T \prime.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114466
    Description:

    This paper presents results from methodological experiments comparing telephone and face-to-face interviewing in surveys of the general population. The relatively low level of telephone ownership in the United Kingdom, especially among the less privileged, argues the need for a dual-mode approach combining telephone interviews with face-to-face interviews for those without telephones. This approach depends on the absence of differential mode-effects on the answers obtained or on the ability to account for these effects when they occur.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114468
    Description:

    The National Agricultural Statistics Service, U.S. Department of Agriculture, conducts yield surveys for a variety of field crops in the United States. While field sampling procedures for various crops differ, the same basic survey design is used for all crops. The survey design and current estimators are reviewed. Alternative estimators of yield and production and of the variance of the estimators are presented. Current estimators and alternative estimators are compared, both theoretically and in a Monte Carlo simulation.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114511
    Description:

    A new unequal probability sampling scheme for selecting n(> 2) units without replacement from a finite population is proposed. This scheme ensures that the inclusion probabilities are proportional to sizes. It has the advantage of simplicity in selection and estimation and also provides a non-negative variance estimator. The variance of the Horvitz-Thompson (H-T) estimator under the proposed scheme is shown to be smaller than that of the customary estimator in probability proportional to size sampling with replacement. The proposed scheme also compares favourably with the without replacement scheme suggested by Sampford (1967) in an empirical study on a few natural populations.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114510
    Description:

    The method of minimum Q^(T) estimation for complex survey designs proposed by Singh (1985) provides asymptotically efficient estimates of model parameters analogous to Neyman’s (1949) min X^2 estimation procedure for simple random samples. The Q^(T) can be viewed as a X^2 type statistic for categorical survey data, and min Q^(T) estimates provide a robust alternative to Weighted Least Squares estimates, which often display unstable behaviour for complex surveys. In this paper, the min Q^(T) method is first described and then illustrated for the problem of estimating parameters of a logit model for survey estimates of unemployment rates which are obtained from the October 1980 Canadian LFS data cross-classified according to age-education covariate categories. It is seen that the trace efficiency of smoothed estimates obtained by Kumar and Rao (1986), who applied the method of pseudo maximum likelihood estimates (pseudo mle) to the same problem can be slightly improved by the min Q^(T) method. Interestingly enough, pseudo mle for individual cells behave much the same way as the efficient min Q^(T) estimates for the particular LFS example.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114509
    Description:

    This paper discusses three problems that have been a major preoccupation among researchers and practitioners of seasonal adjustment in statistical bureaus for the last ten years. These problems are: (l) the use of concurrent seasonal factors versus seasonal factor forecasts for current seasonal adjustment; (2) finding an optimal pattern of revisions for series seasonally adjusted with concurrent factors; and (3) smoothing highly irregular seasonally adjusted data.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114467
    Description:

    Demands for statistics on all aspects of our lives, our society and our economy continue to grow. At the same time statistical agencies share with many respondents a growing concern over the mounting burden of response to surveys. One result of the search for alternative methods of satisfying statistical demands has been an increased emphasis on the use of administrative records for statistical purposes. This paper reviews recent experience at Statistics Canada in this area and discusses obstacles to the greater use of administrative records. Approaches to rendering administrative systems more useful for statistical purposes are reviewed, together with some important concerns related to information protection and record linkage.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114512
    Description:

    The Health and Activity Limitation Survey is part of the program to establish a data base on the disabled population in Canada. The sample design used for the part of the survey covering the population not living in institutions is described. In addition, the methods used to determine the sizes of the samples and to select the samples are presented.

    Release date: 1987-06-15

  • Articles and reports: 12-001-X198700114465
    Description:

    The two-stage rejection rule telephone sample design described by Waksberg (1978) is modified to improve the efficiency of telephone surveys of the U.S. Black population. Experimental tests of sample design alternatives demonstrate that: a) use of rough stratification based on telephone exchange names and states; b) use of large cluster definitions (200 and 400 consecutive numbers) at the first stage; and c) rejection rules based on racial status of the household combine to offer improvements in the relative precision of a sample, given fixed resources. Cost and error models are examined to simulate design alternatives.

    Release date: 1987-06-15

Reference (0)

Reference (0) (0 results)

Your search for "" found no results in this section of the site.

You may try:

Date modified: