This study examined the effect of the National Incident-Based Reporting System (NIBRS) on item missing data in homicide cases.
Homicide researchers often confront the problem of item missing data or murders in which particular characteristics are unknown, such as victim-offender relationship for circumstance. Two options are available to address missing data: reduce the amount by improving data collection during the design stage and/or estimate the values using imputation methods during the analysis stage. Many researchers rely on the Supplementary Homicide Report (SHR) rather than the Uniform Crime Reports (UCR) for United States homicide data. The Federal Bureau of Investigation (FBI) oversees collection of the SHR data therefore researchers have little control over the design and instead focus on the imputation strategies to compensate for item missing data. The study used tables to create a comparison of selected SHR and NIBRS design elements and the expected effect on NIBRS missing data, the effect in each case being less missing data. The data used to conduct this study were from the 1999 SHR and the 1999 NIBRS, which constitute the most recently available data for both the SHR and NIBRS data. The cases examined were all murders and nonnegligent manslaughters. The findings demonstrate the affect data collection design has on murder information available to researchers. As more States and agencies convert from the summary and SHR systems to NIBRS, more attention is needed to understand how this new design affects the quality of homicide data. One limitation is that the UCR does not capture information known to the police. A certain amount of missing data can be attributed to how the data are collected as opposed to reflecting truly unknown information. A second limitation concerns the continued difficulty to capture certain incident details. One explanation is the lack of definitional guidance and it is also possible that new problems have arisen based on how the data are collected in NIBRS. Additional work is needed to more fully understand the effect of NIBRS reporting on other aspects of the quality of homicide and overall crime data. Tables, references