Analysis
Results
All (19)
All (19) (0 to 10 of 19 results)
- Articles and reports: 12-001-X197500254826Description: Exact formulae for bias and mean square error of an estimator of process average in single sampling with rectification for finite lots are obtained. Efficiency of the estimator as compared to an unbiased estimator based on the first sample is obtained for a number of values of lot size, sample size, acceptance number and process average used in sampling plans in quality control of data processing.Release date: 1975-12-15
- 2. Method Test Panel Phase II ‒ Data analysis ArchivedArticles and reports: 12-001-X197500254827Description: In the Methods Test Panel Phase II it was required to do analysis of variance on proportions. Since such analysis gives only approximate results, two models were used in order to be able to draw safe conclusions. Analysis of variance was performed with the proportions as variable and also with the arc sine of the square root of the proportions. The two models are outlined in the present paper and empirical comparisons are made using the MTP Phase II data.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254828Description: The Canadian Travel Survey, 1971 was the largest survey on travel of Canadian residents. This paper describes some important aspects of the methodology. Particular emphasis is given to the development of definitions in relation to the methodology, the sampling technique and interview strategy.Release date: 1975-12-15
- 4. Some variance estimators for multistage sampling ArchivedArticles and reports: 12-001-X197500254829Description: J.N.K. Rao (1975) derived a general formula for estimating the variance in multistage sample designs. This general formula extends the previous results by Des Raj (1966) to the case where the conditional variance from a given primary sampling unit is a random variable. The authors reviewed Rao's paper for its application to Horvitz-Thompson and Yates-Grundy variance estimators as well as the variance estimator for the random group method by Rao, Hartley and Cochran (1962). The authors present an altered version of the Yates-Grundy variance estimators as a result of Rao's paper.Release date: 1975-12-15
- 5. On the improvement of sample survey estimates ArchivedArticles and reports: 12-001-X197500254830Description: This paper focuses on the improvement of sample survey estimates in the particular situation where the survey sample, or part of it, is included in a larger sample from which auxiliary information is available. The properties of a method of estimation - sometimes applied in specific circumstances - are investigated and the limitations of its application are found. The application of the method to rotation designs in continuing surveys is more closely studied in the context of composite estimation.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254831Description: This paper summarizes the results of a telephone experiment conducted in conjunction with the Canadian Labour Force Survey over the period June 1972 to November 1973. Included in the paper is a detailed outline of the purpose and design of the experiment. A discussion of the impact telephone interviewing had on the cost of enumeration, non-response and participation and unemployment rates is given. In addition, interviewer and respondent attitudes toward telephone interviewing are described. Finally, the paper summarizes the experiences gained from this experiment and indicates some areas where further examinations related to telephone interviewing can be carried out.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254832Description: A ratio estimate based on an auxiliary variable is considered for the case when the sample is post-stratified using information on another auxiliary variable. The variance of the ratio estimate is derived by the method of linearization [3,4]. An application to subprovincial estimation in the Canadian Labour Force Survey is discussed.Release date: 1975-12-15
- 8. Analytic studies of sample survey data ArchivedArticles and reports: 12-001-X197500300001Description: Most sample surveys in the past have been "descriptive" in the sense that the main objective is the computation of means or totals of a number of characters of interest along with their standard errors. However, in recent years data produced from "descriptive" surveys are also being increasingly used for "analytical" purposes, i.e., for investigating relationships among variables. Also some sample surveys might have primary "analytical goals" in which case the "optimal" designing of such "analytical surveys" becomes important. These lecture notes present an account of some recent developments in the analytical studies of sample survey data. Many challenging problems remain to be solved and I hope these notes will provide stimulation for further research in this important area.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254824Description:
Madow [1968] has proposed a two-phase sampling scheme under which response bias can be eliminated from sample surveys by obtaining “true” values for a subsample of the original sample. Often in cases of Censuses or ongoing surveys, the subsample data are not used to correct the main survey estimates but to assess their reliability. The main purpose of this paper is to present methods by which reliability estimates can be obtained when true values can be determined for a subsample of units.
Release date: 1975-12-15 - 10. Controlled random rounding ArchivedArticles and reports: 12-001-X197500254825Description:
Random rounding is a technique to ensure confidentiality of aggregate statistics. By randomly rounding all the components of a total, independently, together with the random rounding of the total itself, substantial discrepancies may arise when aggregating the published data. This paper presents a procedure which avoids substantial discrepancies while still protecting the concept of confidentiality.
Release date: 1975-12-15
Articles and reports (19)
Articles and reports (19) (0 to 10 of 19 results)
- Articles and reports: 12-001-X197500254826Description: Exact formulae for bias and mean square error of an estimator of process average in single sampling with rectification for finite lots are obtained. Efficiency of the estimator as compared to an unbiased estimator based on the first sample is obtained for a number of values of lot size, sample size, acceptance number and process average used in sampling plans in quality control of data processing.Release date: 1975-12-15
- 2. Method Test Panel Phase II ‒ Data analysis ArchivedArticles and reports: 12-001-X197500254827Description: In the Methods Test Panel Phase II it was required to do analysis of variance on proportions. Since such analysis gives only approximate results, two models were used in order to be able to draw safe conclusions. Analysis of variance was performed with the proportions as variable and also with the arc sine of the square root of the proportions. The two models are outlined in the present paper and empirical comparisons are made using the MTP Phase II data.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254828Description: The Canadian Travel Survey, 1971 was the largest survey on travel of Canadian residents. This paper describes some important aspects of the methodology. Particular emphasis is given to the development of definitions in relation to the methodology, the sampling technique and interview strategy.Release date: 1975-12-15
- 4. Some variance estimators for multistage sampling ArchivedArticles and reports: 12-001-X197500254829Description: J.N.K. Rao (1975) derived a general formula for estimating the variance in multistage sample designs. This general formula extends the previous results by Des Raj (1966) to the case where the conditional variance from a given primary sampling unit is a random variable. The authors reviewed Rao's paper for its application to Horvitz-Thompson and Yates-Grundy variance estimators as well as the variance estimator for the random group method by Rao, Hartley and Cochran (1962). The authors present an altered version of the Yates-Grundy variance estimators as a result of Rao's paper.Release date: 1975-12-15
- 5. On the improvement of sample survey estimates ArchivedArticles and reports: 12-001-X197500254830Description: This paper focuses on the improvement of sample survey estimates in the particular situation where the survey sample, or part of it, is included in a larger sample from which auxiliary information is available. The properties of a method of estimation - sometimes applied in specific circumstances - are investigated and the limitations of its application are found. The application of the method to rotation designs in continuing surveys is more closely studied in the context of composite estimation.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254831Description: This paper summarizes the results of a telephone experiment conducted in conjunction with the Canadian Labour Force Survey over the period June 1972 to November 1973. Included in the paper is a detailed outline of the purpose and design of the experiment. A discussion of the impact telephone interviewing had on the cost of enumeration, non-response and participation and unemployment rates is given. In addition, interviewer and respondent attitudes toward telephone interviewing are described. Finally, the paper summarizes the experiences gained from this experiment and indicates some areas where further examinations related to telephone interviewing can be carried out.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254832Description: A ratio estimate based on an auxiliary variable is considered for the case when the sample is post-stratified using information on another auxiliary variable. The variance of the ratio estimate is derived by the method of linearization [3,4]. An application to subprovincial estimation in the Canadian Labour Force Survey is discussed.Release date: 1975-12-15
- 8. Analytic studies of sample survey data ArchivedArticles and reports: 12-001-X197500300001Description: Most sample surveys in the past have been "descriptive" in the sense that the main objective is the computation of means or totals of a number of characters of interest along with their standard errors. However, in recent years data produced from "descriptive" surveys are also being increasingly used for "analytical" purposes, i.e., for investigating relationships among variables. Also some sample surveys might have primary "analytical goals" in which case the "optimal" designing of such "analytical surveys" becomes important. These lecture notes present an account of some recent developments in the analytical studies of sample survey data. Many challenging problems remain to be solved and I hope these notes will provide stimulation for further research in this important area.Release date: 1975-12-15
- Articles and reports: 12-001-X197500254824Description:
Madow [1968] has proposed a two-phase sampling scheme under which response bias can be eliminated from sample surveys by obtaining “true” values for a subsample of the original sample. Often in cases of Censuses or ongoing surveys, the subsample data are not used to correct the main survey estimates but to assess their reliability. The main purpose of this paper is to present methods by which reliability estimates can be obtained when true values can be determined for a subsample of units.
Release date: 1975-12-15 - 10. Controlled random rounding ArchivedArticles and reports: 12-001-X197500254825Description:
Random rounding is a technique to ensure confidentiality of aggregate statistics. By randomly rounding all the components of a total, independently, together with the random rounding of the total itself, substantial discrepancies may arise when aggregating the published data. This paper presents a procedure which avoids substantial discrepancies while still protecting the concept of confidentiality.
Release date: 1975-12-15