We thank Nafisa Halim and Sarah Strong for invaluable assistance with data management, and Steven Messner for comments on a previous draft. This research was supported by the National Institute of Justice, Grant #2005-IJ-CX-0007. Direct correspondence to Tim Wadsworth, Department of Sociology, UCB 327, University of Colorado at Boulder, Boulder, CO 80309 (e-mail: email@example.com).
WHEN MISSING DATA ARE NOT MISSING: A NEW APPROACH TO EVALUATING SUPPLEMENTAL HOMICIDE REPORT IMPUTATION STRATEGIES†
Article first published online: 3 DEC 2008
© 2008 American Society of Criminology
Volume 46, Issue 4, pages 841–870, November 2008
How to Cite
WADSWORTH, T. and ROBERTS, J. M. (2008), WHEN MISSING DATA ARE NOT MISSING: A NEW APPROACH TO EVALUATING SUPPLEMENTAL HOMICIDE REPORT IMPUTATION STRATEGIES. Criminology, 46: 841–870. doi: 10.1111/j.1745-9125.2008.00129.x
- Issue published online: 3 DEC 2008
- Article first published online: 3 DEC 2008
- Supplemental Homicide Reports (SHR);
- data imputation;
- crime data
The Supplemental Homicide Reports (SHR) are widely used in criminological research and inform a broad range of research topics and subsequent policy applications. A serious issue with the SHR is missing information about the offender and incident in many recorded homicides. Although it is convenient to discard cases with missing data before analysis, such discarding is not theoretically justified and can lead to incorrect substantive conclusions. Recently, several techniques for imputing missing SHR data have been proposed, but it is difficult to evaluate their effectiveness. This research presents a new approach to testing and evaluating SHR imputation techniques.
Offender data that are missing in the SHR are often found in police records available for individual cities. We examine similarities and differences among cases with known offender characteristics in the SHR, cases with such information missing in the SHR but available in police records, and cases with such information missing in both sources. We then use these data sets to evaluate four different imputation techniques suggested in the literature (Fox, 2004; Messner, Deane, and Beaulieu, 2002; Pampel and Williams, 2000; Regoeczi and Riedel, 2003). We apply each imputation technique to the SHR, and for cases with information missing in the SHR but known in the police records, we see how well the imputed values correspond both with the individual known values and with the overall distributions in the police records. We discuss the outcome of our assessment of these strategies, and we outline important implications this assessment has for research using SHR data.