The study of homicide invokes a scientific investigation of the frequency, nature, and causes of one human being killing another. As researchers explore criminal homicide, they tend to examine murder and nonnegligent manslaughter as defined by official sources (such as the UCR), which excludes justifiable homicide, manslaughter caused by negligence, suicide, or attempted murders.
Criminal homicide is classified by the Uniform Crime Reporting (UCR) Program as both murder and nonnegligent manslaughter or as manslaughter by negligence. Homicide research generally examines murder and nonnegligent manslaughter, defined as the willful killing of one by another. Justifiable homicide, manslaughter caused by negligence, suicide, or attempts to murder are not included in this definition. For UCR purposes, justifiable homicide is limited to the killing of a felon by an officer in the line of duty or the killing of a felon, during the commission of a felony, by a private citizen. Manslaughter by negligence is the killing of a human being by gross negligence.
When researching homicide, scholars generally utilize two national sources of homicide data—the Uniform Crime Reporting Program of the Federal Bureau of Investigation (FBI) and mortality files from the Vital Statistics Division of the National Center for Health Statistics (NCHS). These two data sources vary greatly in the information collected.
The UCR is an official source of crime statistics based on reported crimes. That is, it is based on the number of arrests voluntarily reported to the FBI by law enforcement agencies. These crimes include murder, rape, robbery, aggravated assault, burglary, larceny-theft, vehicle theft, and arson. In addition to monthly criminal offense information compiled for UCR purposes, law enforcement agencies submit supplemental data to the FBI on homicide. Supplemental Homicide Reports (SHRs) contain supplemental information on homicide incidents. SHRs include detailed, incident-level data on nearly all murders and nonnegligent manslaughters that have occurred in the United States in a given year. These reports contain information for each homicide incident, including information on trends, demographics of persons arrested, and the characteristics of the homicide (i.e., demographics of victims, victim–offender relationships, weapon used, and circumstances surrounding the homicide).
Though a rich source of homicide data, the UCR Program has weaknesses of which researchers are well aware. Missing information regarding the homicide incident is problematic in the UCR and probably its main weakness. This may be due to the fact that participation by police agencies in the UCR Program is completely voluntary. Therefore, some law enforcement agencies fail to report their homicide incidents to the FBI or fail to fully record all relevant information. Despite this fact, official sources like the Bureau of Justice Statistics (BJS) have found that the SHRs are just over 90% complete. Though the coverage is high, there are still a number of homicides that go unaccounted for. Some researchers have taken steps to correct for underreporting by law enforcement agencies by statistically adjusting for the total number of homicide incidents reported to the UCR (see Fox, 2004). The ability to adjust for missing data based on known homicide cases has increased the popularity of this data source among researchers. However, one growing problem, particularly with homicide offender information, is the increase in the number of unsolved or uncleared murders by police agencies. Stranger homicides take longer to clear by arrest and therefore often get submitted as “unknown.” Ignoring homicides with missing offender information understates homicide offending. Thus, there is greater dependency on researchers to use weighting strategies that statistically adjust for missing offender data (see Fox, 2004, for a detailed discussion of this specific issue).
The mortality reporting system is far simpler compared to the UCR. In this system, homicide information is gathered by medical examiners in the completion of standardized death certificates. Once verified, death certificates are entered into a national mortality dataset by the NCHS. According to the NCHS, these data represent at least 90% of all homicides that have occurred in the United States. This type of data contains information on the victim of the homicide. Victim information includes demographics, occupation, education, time of death, place of death, and cause of death. Just as seen for the UCR, there are weaknesses in mortality data such as omissions and underreporting. Unfortunately, for a variety of reasons not all death certificates are received by NCHS, and in some incidents information on death certificates is not entered into mortality figures. In addition, unlike with the UCR, offender information is not available and unable to be collected. Even with these notable limitations, homicide remains the most accurately measured and reported offense relative to other types of criminal offenses.
One of the reasons these data issues are so critical is that researchers and policymakers are interested in documenting and understanding the changes in homicide offending over time. That is, researchers and policy makers alike want to know how much homicide offending is occurring, why homicides occur and if the level of homicide offending is increasing or decreasing in certain areas (i.e., states, cities, or counties) over time.