skip navigation


Abstract Database

Register for Latest Research

Stay Informed
Register with NCJRS to receive NCJRS's biweekly e-newsletter JUSTINFO and additional periodic emails from NCJRS and the NCJRS federal sponsors that highlight the latest research published or sponsored by the Office of Justice Programs.

NCJRS Abstract

To download this abstract, check the box next to the NCJ number then click the "Back To Search Results" link. Then, click the "Download" button on the Search Results page. Also see the Obtain Documents page for direction on how to access resources online, via mail, through interlibrary loans, or in a local library.


NCJ Number: 249242 Find in a Library
Title: Evaluation of Direct Variance Estimation, Estimate Reliability, and Confidence Intervals for the National Crime Victimization Survey
Author(s): Rick Williams; David Heller; Lance Couzens; Bonnie Shook-Sa; Marcus Berzofsky; Hope Smiley-McDonald; Chris Krebs
Corporate Author: RTI International
United States of America
Project Director: Lynn Langton; Michael Planty
Date Published: 2015
Page Count: 55
Sponsoring Agency: Bureau of Justice Statistics (BJS)
Washington, DC 20531
RTI International
Research Triangle Park, NC 27709
Contract Number: 2011-NV-CX-K068
Sale Source: Bureau of Justice Statistics (BJS)
US Dept of Justice
Office of Justice Programs
810 Seventh Street NW
Washington, DC 20531
United States of America
Document: Agency Summary|PDF
Agency Summary: 
Type: Report (Grant Sponsored); Report (Study/Research); Research (Applied/Empirical)
Format: Document; Document (Online)
Language: English
Country: United States of America
Annotation: This study examined the feasibility of using direct variance estimation (DVE) on data obtained by the National Crime Victimization Survey (NCVS), and it offers recommendations regarding the calculation of confidence intervals (CIs) and the identification of unstable estimates.
Abstract: Currently, the Bureau of Justice Statistics uses generalized variance functions (GVFs) rather than DVE in calculating variances of estimates for the NCVS. GVF models enable users to calculate approximations of variances without knowledge of the sample design; however, the GVFs developed for the NCVS do not allow for complex analyses, such as regression modeling, and are cumbersome when multiple estimates are produced. DVE uses software that accounts for complex sample designs. The two direct variance techniques used in the current study are Taylor series linearization (TSL) and balanced repeated replication (BRR). The findings and recommendations from this study pertain to the areas of data analysis for the NCVS in which GVF, TSL, and BRR are best suited. Findings and recommendations are also provided for the flagging and suppression of unreliable estimation. Recommendations are also offered for where CIs should be used with NCVS data analysis. 18 exhibits and 13 references
Main Term(s): National Crime Victimization Survey (NCVS)
Index Term(s): BJS Grant-related Documents; BJS Resources; Bureau of Justice Statistics (BJS); Comparative analysis; Statistical analysis
To cite this abstract, use the following link:

*A link to the full-text document is provided whenever possible. For documents not available online, a link to the publisher's website is provided. Tell us how you use the NCJRS Library and Abstracts Database - send us your feedback.