Filter results by
Search HelpKeyword(s)
Subject
Author(s)
Results
All (4)
All (4) ((4 results))
- Articles and reports: 12-001-X20060029548Description:
The theory of multiple imputation for missing data requires that imputations be made conditional on the sampling design. However, most standard software packages for performing model-based multiple imputation assume simple random samples, leading many practitioners not to account for complex sample design features, such as stratification and clustering, in their imputations. Theory predicts that analyses of such multiply-imputed data sets can yield biased estimates from the design-based perspective. In this article, we illustrate through simulation that (i) the bias can be severe when the design features are related to the survey variables of interest, and (ii) the bias can be reduced by controlling for the design features in the imputation models. The simulations also illustrate that conditioning on irrelevant design features in the imputation models can yield conservative inferences, provided that the models include other relevant predictors. These results suggest a prescription for imputers: the safest course of action is to include design variables in the specification of imputation models. Using real data, we demonstrate a simple approach for incorporating complex design features that can be used with some of the standard software packages for creating multiple imputations.
Release date: 2006-12-21 - Articles and reports: 12-001-X20060029555Description:
Researchers and policy makers often use data from nationally representative probability sample surveys. The number of topics covered by such surveys, and hence the amount of interviewing time involved, have typically increased over the years, resulting in increased costs and respondent burden. A potential solution to this problem is to carefully form subsets of the items in a survey and administer one such subset to each respondent. Designs of this type are called "split-questionnaire" designs or "matrix sampling" designs. The administration of only a subset of the survey items to each respondent in a matrix sampling design creates what can be considered missing data. Multiple imputation (Rubin 1987), a general-purpose approach developed for handling data with missing values, is appealing for the analysis of data from a matrix sample, because once the multiple imputations are created, data analysts can apply standard methods for analyzing complete data from a sample survey. This paper develops and evaluates a method for creating matrix sampling forms, each form containing a subset of items to be administered to randomly selected respondents. The method can be applied in complex settings, including situations in which skip patterns are present. Forms are created in such a way that each form includes items that are predictive of the excluded items, so that subsequent analyses based on multiple imputation can recover some of the information about the excluded items that would have been collected had there been no matrix sampling. The matrix sampling and multiple-imputation methods are evaluated using data from the National Health and Nutrition Examination Survey, one of many nationally representative probability sample surveys conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention. The study demonstrates the feasibility of the approach applied to a major national health survey with complex structure, and it provides practical advice about appropriate items to include in matrix sampling designs in future surveys.
Release date: 2006-12-21 - Articles and reports: 12-001-X20060019257Description:
In the presence of item nonreponse, two approaches have been traditionally used to make inference on parameters of interest. The first approach assumes uniform response within imputation cells whereas the second approach assumes ignorable response but make use of a model on the variable of interest as the basis for inference. In this paper, we propose a third appoach that assumes a specified ignorable response mechanism without having to specify a model on the variable of interest. In this case, we show how to obtain imputed values which lead to estimators of a total that are approximately unbiased under the proposed approach as well as the second approach. Variance estimators of the imputed estimators that are approximately unbiased are also obtained using an approach of Fay (1991) in which the order of sampling and response is reversed. Finally, simulation studies are conducted to investigate the finite sample performance of the methods in terms of bias and mean square error.
Release date: 2006-07-20 - Articles and reports: 12-001-X20050029040Description:
A large part of sample survey theory has been directly motivated by practical problems encountered in the design and analysis of sample surveys. On the other hand, sample survey theory has influenced practice, often leading to significant improvements. This paper will examine this interplay over the past 60 years or so. Examples where new theory is needed or where theory exists but is not used will also be presented.
Release date: 2006-02-17
Stats in brief (0)
Stats in brief (0) (0 results)
No content available at this time.
Articles and reports (4)
Articles and reports (4) ((4 results))
- Articles and reports: 12-001-X20060029548Description:
The theory of multiple imputation for missing data requires that imputations be made conditional on the sampling design. However, most standard software packages for performing model-based multiple imputation assume simple random samples, leading many practitioners not to account for complex sample design features, such as stratification and clustering, in their imputations. Theory predicts that analyses of such multiply-imputed data sets can yield biased estimates from the design-based perspective. In this article, we illustrate through simulation that (i) the bias can be severe when the design features are related to the survey variables of interest, and (ii) the bias can be reduced by controlling for the design features in the imputation models. The simulations also illustrate that conditioning on irrelevant design features in the imputation models can yield conservative inferences, provided that the models include other relevant predictors. These results suggest a prescription for imputers: the safest course of action is to include design variables in the specification of imputation models. Using real data, we demonstrate a simple approach for incorporating complex design features that can be used with some of the standard software packages for creating multiple imputations.
Release date: 2006-12-21 - Articles and reports: 12-001-X20060029555Description:
Researchers and policy makers often use data from nationally representative probability sample surveys. The number of topics covered by such surveys, and hence the amount of interviewing time involved, have typically increased over the years, resulting in increased costs and respondent burden. A potential solution to this problem is to carefully form subsets of the items in a survey and administer one such subset to each respondent. Designs of this type are called "split-questionnaire" designs or "matrix sampling" designs. The administration of only a subset of the survey items to each respondent in a matrix sampling design creates what can be considered missing data. Multiple imputation (Rubin 1987), a general-purpose approach developed for handling data with missing values, is appealing for the analysis of data from a matrix sample, because once the multiple imputations are created, data analysts can apply standard methods for analyzing complete data from a sample survey. This paper develops and evaluates a method for creating matrix sampling forms, each form containing a subset of items to be administered to randomly selected respondents. The method can be applied in complex settings, including situations in which skip patterns are present. Forms are created in such a way that each form includes items that are predictive of the excluded items, so that subsequent analyses based on multiple imputation can recover some of the information about the excluded items that would have been collected had there been no matrix sampling. The matrix sampling and multiple-imputation methods are evaluated using data from the National Health and Nutrition Examination Survey, one of many nationally representative probability sample surveys conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention. The study demonstrates the feasibility of the approach applied to a major national health survey with complex structure, and it provides practical advice about appropriate items to include in matrix sampling designs in future surveys.
Release date: 2006-12-21 - Articles and reports: 12-001-X20060019257Description:
In the presence of item nonreponse, two approaches have been traditionally used to make inference on parameters of interest. The first approach assumes uniform response within imputation cells whereas the second approach assumes ignorable response but make use of a model on the variable of interest as the basis for inference. In this paper, we propose a third appoach that assumes a specified ignorable response mechanism without having to specify a model on the variable of interest. In this case, we show how to obtain imputed values which lead to estimators of a total that are approximately unbiased under the proposed approach as well as the second approach. Variance estimators of the imputed estimators that are approximately unbiased are also obtained using an approach of Fay (1991) in which the order of sampling and response is reversed. Finally, simulation studies are conducted to investigate the finite sample performance of the methods in terms of bias and mean square error.
Release date: 2006-07-20 - Articles and reports: 12-001-X20050029040Description:
A large part of sample survey theory has been directly motivated by practical problems encountered in the design and analysis of sample surveys. On the other hand, sample survey theory has influenced practice, often leading to significant improvements. This paper will examine this interplay over the past 60 years or so. Examples where new theory is needed or where theory exists but is not used will also be presented.
Release date: 2006-02-17
Journals and periodicals (0)
Journals and periodicals (0) (0 results)
No content available at this time.
- Date modified: