Stability of variance estimators under complex sampling designs /
Stability of a variance estimator V(o) is an important practical consideration in the analysis of sample survey data. In some cases, the stability of V(o) is a matter of intrinsic interest. In other cases, stability issues are an intermediate consideration in the construction of appropriate confid...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1996.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=739364011&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | Stability of a variance estimator V(o) is an important practical consideration in the analysis of sample survey data. In some cases, the stability of V(o) is a matter of intrinsic interest. In other cases, stability issues are an intermediate consideration in the construction of appropriate confidence intervals for the principal parameters of interest. Specifically, for a given point estimator O and variance estimator V(O), at-based confidence interval for a parameter 0 depends on a t-critical value based on the degrees-of-freedom stability measure d. Consequently, a stability measure, d, has a direct impact on confidence interval construction. We develop three methods for assessing the stability of V(o) allowing for heterogeneity among stratum-level variances Vh. First, the standard Satterthwaite estimator ds generally has a substantial negative bias if the number of strata, L, is large and the numbers of primary sample units, nh, are small. Examination of the expectations of the leading terms of a Taylor expansion of ds leads to a modified estimator. Second, in some practical cases, within-primary-utlit variance estimators provide auxiliary information which is closely related to Vh. We can use this information in conjunction with an errors-in-variables model to produce an improved estimator of ds. Third, we use stratum-collapse methods to produce some conservative stability measures. The proposed methods are applied to data from the U.S. Third National Health and Nutrition Examination Survey and the U.S. Consumer Expenditure Survey. |
|---|---|
| Item Description: | Vita. "Major Subject: Statistics". |
| Physical Description: | xiii, 127 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilms Inc. |
| Bibliography: | Includes bibliographical references: pages 124-125. |