Please Note: Due to the large size of the map files included in the PDF version of this report, it may take a while for your computer to download them completely. You may be able to view complete copies of the maps easier and more quickly by saving the document to your hard drive and opening it from your computer. If you have any questions or experience difficulty in accessing this report, please call Dave Williams at 202-261-5542 or e-mail the webmaster (webmaster@ui.urban.org).
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, its funders.
Introduction
Violent crime continues to be a major concern in the District of Columbia. In 1999, the District recorded 14,871 violent incidents against 16,372 victims. Nearly 12,000 individuals were assaulted and almost 4,000 robbed. To inform efforts to reduce violence in the city, the Mayor's Office and the Criminal Justice Coordinating Council (CJCC) convened a
small "data team" in mid-July, 2000. Composed of representatives from the Metropolitan Police Department of the District of Columbia (MPDC), the Urban Institute, the Mayor's Office, and the CJCC, the team was charged with analyzing violent incidents in the city by identifying trends and patterns that characterize violent incidents, victims, and offenders. The team focused on a small number of straightforward yet important questions:
- What types of violent incidents occur most frequently?
- Who is impacted most by the violence? Who is most at-risk for being victimized, and who is committing the offenses?
- Where is the violence occurring? Which neighborhoods and Police Service Areas are disproportionately affected by violence?
- When do the violent incidents occur? At what time of day? On what days of the week?
Under a contract from the Criminal Justice Coordinating Council, the Urban Institute used existing data to address these questions. In the report that follows, we describe the aggregate characteristics of violent crime and identify patterns and trends—the places, times and
individuals most at risk of violence—to inform policymakers and law enforcement officials in the development of public safety strategies.
For this report, violent crimes include the following: homicide, sexual assault, robbery, aggravated assault, and simple assault.1 Homicide includes murder (first and second degree),
involuntary manslaughter and any other felony murder, including manslaughter (neglect). Sexual assault includes all categories of rapes, including statutory rape and attempted rape; and all categories of sexual abuse. Robbery includes all robbery categories as well as purse snatching, carjacking and attempted robberies. Aggravated assaults are more serious assaults, including assaults with a dangerous weapon, threats involving a weapon, assaults resulting in injury, and assault with intent to kill, rape, or rob. Simple assaults include assaults resulting in minor injury and those without injury. For this report, offenses are not broken down into subcategories (e.g., armed robbery, attempted robbery, purse snatching) of the major offenses.2
THE DATA
Offense Data. Offense report data were provided by the Metropolitan Police Department through the Criminal Justice Coordinating Council. The data file provided to the Urban Institute in August 2000 contains all offense reports filed by the MPDC in 1999. Offense data include age, race, and gender of victim, time of day, day of week, and location of the incident including address, Police Service Area, and police district. These data were analyzed in two ways. Most analyses that follow describe patterns in the number of victims, recognizing that a single incident may have involved multiple victims. In some places, we also describe the number of incidents. Throughout the report, we denote whether we are referring to victims or incidents.3
It should be noted that offense data reflect only offenses that were both reported to the police and filed by the police (it is MPDC's policy to file a report on every incident for which the police officer determines a crime has occurred). Still, these data undercount the actual level of victimization and violence in the District because they only include reported crimes. For example, according to a Department of Justice report, "Criminal Victimization and Perceptions of Community Safety in 12 Cities, 1998," it is estimated that there were more than 26,000 violent victimizations in the District of Columbia in 1998, yet only 50 percent were reported to the police. In addition to not knowing how many violent crimes were not reported to the police in
1999, we also do not know what percentage of the reported incidents were actually filed in an offense report.
Arrest Data. Arrest data were also provided by the Metropolitan Police Department through the Criminal Justice Coordinating Council. These data were delivered as two files. One file contained a list of each person arrested in 1999 in the District of Columbia, and the other file
contained a complete list of charges and charge descriptions for each arrested individual. Additional information was recorded on the characteristics of the offender (e.g., age, race, and gender) and the arrest event (e.g., location, time, and date of arrest).
The arrest data were analyzed in two ways. First, to highlight the demographic characteristics of violent crime offenders, we examined the race, gender, and age distributions of arrestees by offense type. Second, we described the victim-offender relationship in terms of the age group
offenders are likely to attack. This analysis is a straightforward presentation of offenders' age, race, and gender compared with the age, race, and gender of their victims.
Population Data. Population data for the District were obtained from the U.S. Census Bureau Web site and racial distribution data were obtained from the D.C. Office of Planning. In 1999, it is estimated that there were 519,000 persons residing in the District. Of this total, blacks accounted for 60 percent of the population and nearly 30 percent were non-Hispanic whites. Where possible, data in tables and figures are shown as a rate (the number of violent crime victims per 1,000 persons).4 For the offense report data, race is entered as either black or white, meaning all non-black minority victims are coded as white in the offense data files kept by the Metropolitan Police Department.
Methodology for Capturing Violence in D.C.
For this report, we quantify violent victimization in terms of the number of violent crime victims in the District of Columbia in 1999. This approach should not be confused with the reporting of violent crime incidents that the Metropolitan Police Department submits to the UCR, or use for MPD official reporting on violence (i.e., Web site statistics). Our focus was to paint a detailed picture of victimization by counting every victim.
To accomplish this task, we counted each victim of a violent crime incident by offense type. Because the data file from MPD does not have a personal identification number, we identify multiple victims per offense by the number of records that have the same complaint/complainant number or "CCN." We suspected that 688 (4.2 percent of 16,372) records may have been duplicate records and not evidence of a multiple-victims incident. For these 688 we did not have details on age, gender, and race. More specifically, the event-specific information was listed (e.g., offense code, day of event, time of event) for each victim, but the demographic information was provided for only a single victim.
After a number of discussions with members of the MPD's Research and Resource Development unit, we decided that, given our limited resources, we would leave the possible duplicates in our counts. If we were to drop them, we could not be certain we were not excluding real victims. We decided to err on the side of inclusion. However, the potential duplicates are not reflected in the tables on age, race, and gender, because demographic information was not available on the 688 records. Overall patterns described in the report should remain unaffected by
the inclusion of the small number of possible duplicates. In addition, we confirmed that we have no duplicates in our count of homicides.
Analytical Technique for the Spatial Analysis
We used geographic mapping and cluster analysis to identify patterns of violence in the District for 1999. We examined demographic information on each victim and offender, as well as the time of day of the incident, and created a geographic information system (GIS) to analyze the
distribution of crime across Police Service Areas (PSA) and map concentrations of crime.
Two types of spatial analysis were conducted that related offense reports filed by the police to specific locations. First, the PSA maps show the total number of reported victimizations that occurred within PSA boundaries. The cluster analysis, on the other hand, analyzes entire incidents as a whole as opposed to the number of victims involved in each incident. For the cluster analysis, incidents with multiple victims were counted as one event.
High concentration areas or clusters were identified using the "nearest neighbor clustering" technique of the CrimeStat program.5 Nearest neighbor clustering is a hierarchical technique that first groups points based on the next closest point (nearest neighbor). This technique then generates ellipses around clusters of incidents based on the following user-specified criteria: the probability level for the random expected nearest neighbor to isolate clusters; the minimum number of incidents (events) to form an ellipse; and the number of standard deviations for determining the size of the ellipse.
This study uses one-standard-deviation ellipses with 95 percent confidence that the clustering is spatially closer than would have happened by chance if the crimes were spread evenly over the
District. Criteria for choosing the minimum number of events were selected after preliminary examination of the data. Due to variation in the frequency of the different types of violent crimes, the number of minimum events used to form the clusters varied for each crime type.
Notes from this section
1. MPDC does not generally include simple assaults in their violent crime analyses. Also, here we examine the broad category of "sexual assault" rather than the narrower "rape" category. Therefore, the aggregate data presented in this report is not comparable to MPDC reports of Part I crime.
2. This report only includes analysis of the major violent crime categories as described. Attempts to locate
descriptions of all numeric codes used by MPDC for offense reporting were unsuccessful.
3. The term "events" is synonymous with incidents and is used in the cluster analysis. Any reference to
"victimization" denotes the number of victims reported by the police through incident reports.
4. These data are based on District populations since it is unknown how many crimes are committed against non-District residents.
5. Levine, Ned. 2000. CrimeStat1.1: A Spatial Statistics Program for the Analysis of Crime Incident Locations. Ned Levine and Associates, Annandale, VA, and the National Institute of Justice, Washington, D.C.
This report is available in its entirety in the Portable Document Format (PDF), which many find convenient when printing.
Latest Research
Browse by Topic:
PDF
Printer-friendly summary
