skip navigation

PUBLICATIONS

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

The document referenced below is part of the NCJRS Virtual Library collection. To conduct further searches of the collection, visit the Virtual Library. 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: 200054 Find in a Library
Title: Missing Data Problems in Criminological Research: Two Case Studies
Journal: Journal of Quantitative Criminology  Volume:19  Issue:1  Dated:March 2003  Pages:55-78
Author(s): Robert Brame; Raymond Paternoster
Date Published: March 2003
Page Count: 24
Publisher: http://www.wkap.nl/journalhome.htm/0748-4518 
Type: Case Study
Format: Article
Language: English
Country: United States of America
Annotation: Using two case studies, this paper considers the problem of missing data in two circumstances commonly confronted by criminologists: missing data due to subject attrition and missing data on an independent variable of interest for typical reasons (the respondent did not wish to answer a question or could not be located).
Abstract: A reading of the relevant literature suggests that criminologists often deal with missing data problems by performing an analysis on those observations for which complete data are available (i.e., complete-case analysis). In some circumstances, this approach will yield valid inferences, but in other circumstances, it will not. This paper discusses the issue by considering two case studies that typify problems commonly encountered in criminological research. In the first case study, missing data were due to a loss of observations in a longitudinal research project. The problem was that researchers could not observe the outcome variable for all individuals because some cases dropped out of the study between the two time points. The second case study considered the problem that developed when the objective was to estimate the association between an independent variable and an outcome variable when some of the observations had missing data on the independent variable. The methods used in analyzing these case studies were based on likelihood functions and models that had been described in previous work by Little and Rubin (1987), Little and Schenker (1995), Vach (1994), and Wainer (1986). The case studies analyzed indicated that when the proportion of cases with missing data were substantial, complete-case analysis might produce misleading results. Using two simple problems from actual criminological data sets, this paper illustrated the potential consequences of complete-case analysis by using simple models with a single independent variable and a single dependent variable. The results of the complete-case analyses did not stand up to the scrutiny of a basic sensitivity analysis that critically evaluated the assumptions upon which the complete-case model was based. This analysis suggests that there was some value in the development of methods that would allow researchers to investigate the robustness of their conclusions to different assumptions about missing data mechanisms. Further, this research concludes that missing data can inject considerable ambiguity into criminological research, and this ambiguity can significantly impact what appears at first to be straightforward conclusions. 7 tables and 19 references
Main Term(s): Criminology
Index Term(s): Data analysis; Data collections; Data integrity; Research design; Research methods
To cite this abstract, use the following link:
http://www.ncjrs.gov/App/publications/abstract.aspx?ID=200054

*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.