School Violence

CRIME IN SCHOOLS AND COLLEGES:
A Study of Offenders and Arrestees Reported via
National Incident-Based Reporting System Data

James H. Noonan and Malissa C. Vavra
Crime Analysis, Research and Development Unit
Criminal Justice Information Services Division
Federal Bureau of Investigation
October 2007

Methodology

The NIBRS was designed in the 1980s to enhance the Summary Reporting System by capturing detailed information at the incident level.  Once the system was developed, the FBI began collecting NIBRS data in 1991 from a small group of law enforcement agencies.  By the end of 2004, approximately 33 percent of the Nation’s state and local law enforcement agencies covering 22 percent of the US population reported UCR data to the FBI in the NIBRS format.  (See Table 1.)   In addition, the percentage of crime reported to the UCR Program via the NIBRS had risen from 13 percent in 2000 to 20 percent in 2004.  However, growing increases in the amount of crime reported via NIBRS do not necessarily indicate increases in crime in general or the actual occurrences of crime in schools.  However, increases in the number of NIBRS offenses may be largely the result of more law enforcement agencies using the NIBRS data collection format.  

Table 1: UCR Participation via the NIBRS, by Year
	Year of Incident
	2000	2001	2002	2003	2004
United States Population	281,421,906	285,317,559	287,973,924	290,788,976	293,655,404
Percent of US Population Covered by Agencies Reporting via the NIBRS	16%	17%	17%	20%	22%
Number of Agencies Participating in the UCR Program via the NIBRS1	3,801	4,259	4,302	5,271	5,735
Percent of Agencies Reporting via the NIBRS	22%	25%	25%	31%	33%
Percent of Crime Reported to the UCR Program via the NIBRS	13%	15%	18%	17%	20%

1 Based on law enforcement agencies that submitted their UCR data to the FBI in accordance with NIBRS reporting requirements for inclusion in the annual NIBRS database.

Note: See the study text for specific data definitions, uses, and limitations.

Using a combination of six possible data segments (administrative, offense, victim, property, offender, and arrestee), the NIBRS captures information on criminal incidents involving any of 22 offense categories made up of 46 specific crimes.  To date, the NIBRS offers 56 data elements, i.e., data fields, that law enforcement may use to capture descriptive data about the victims, offenders, and circumstances of criminal incidents and arrests.  Examples of NIBRS data elements include UCR Offense Code, Type of Victim, and Age of Offender (see Appendix E for a complete list of data elements). ⁹  Furthermore, each of the 56 data elements is translated into a series of codes that specify the information being collected.

Of particular importance to the present study is the NIBRS data element Location Type, specifically Code 22,¹⁰ which identifies offenses occurring at schools and colleges.  All the crime data used in the tables and discussions throughout this study were reported by law enforcement as occurring at NIBRS Location Type, Code 22, which hereafter is referred to as school(s), unless otherwise noted.  

As illustrated in Figure 1, an incident report contains various types of data collection segments in addition to the administrative segment.  The report may also include multiple segment records within one or more segments if the incident should warrant them.  For example, an incident occurred during which three victims were held up on school property by two offenders using guns.  The offenders shot and killed the victims and were subsequently arrested.  This incident will have one administrative record, two offense records, various property records for the stolen or recovered property, three victim records, two offender records, and two arrestee records separated into the appropriate segments in the NIBRS database structure.
                            
This study focuses primarily on the offender and arrestee data records; it looks at other records only as they pertain to offenders.  Using the narrowly-defined set of data records, the study specifically addresses incident characteristics (Tables 2 and 7), offender characteristics (Tables 3-5), victim-to-offender relationships (Table 6), offense characteristics (Tables 8 and 9), and arrestee characteristics (Tables 10-15).  Expanding Tables 2 and 8, Appendices A and B show the number of offenses by offense type by year and the weapon type by offense type, respectively. 

This graphic sample illustrates the extraction of the Victim data elements from the Incident Report to the NIBRS Flat File and how they are segmented in the NIBRS Database Structure.

Throughout this study, age groups are aggregated and cross tabulated to help readers view the traits of offenders and arrestees.  These age groups are formulated based upon the following age divisions:  Birth to 4 years old, 5 to 9 years old, 10 to 12 years old, 13 to 15 years old, 16 to 18 years old, and 19 years or older.

Additional considerations for this study follow:

• The term gun refers collectively to all firearm codes found in the NIBRS format, including:  firearm, handgun, rifle, shotgun, and other firearm types.¹¹  Furthermore, information for type weapon/force involved is only collected for Murder and Nonnegligent Manslaughter, Negligent Manslaughter, Justifiable Homicide, Kidnapping/Abduction, Forcible Rape, Forcible Sodomy, Sexual Assault with an Object, Forcible Fondling, Robbery, Aggravated Assault, Simple Assault, Extortion/Blackmail, and Weapon Law Violations.
• Victim-to-offender relationships are only collected for Murder and Nonnegligent Manslaughter, Negligent Manslaughter, Justifiable Homicide, Kidnapping/Abduction, Forcible Rape, Forcible Sodomy, Sexual Assault with an Object, Forcible Fondling, Robbery, Aggravated Assault, Simple Assault, Intimidation, Incest, and Statutory Rape.
• Offenders may be suspected of using alcohol, computers, and/or drugs in an offense.  Because more than one of these codes can be collected on any given offense record, there may be multiple counts of use in incidents.  Multiple counts can be generated in two ways.  An incident with one offense record may be associated with the use of alcohol and drugs.  Another incident may contain two offense records, one offense indicating the suspected use of alcohol and a second offense where the offender is suspected of using drugs.  Both types of incidents will indicate that the offender was suspected of using both alcohol and drugs.  Caution is needed when interpreting this information.
• Frequency tables and cross tabulations are used to examine the characteristics discussed for the 5-year study time frame.  See the Limitations section on cautions for comparing frequencies from year to year.
• The data used in this study reflect the NIBRS submissions originally made for each year within the 5-year period.  They do not include data that were subsequently reported via time-window submissions.¹² 

Special Offense Definitions
An important distinction must be made in the context of this study concerning the use of the term offense.  In the UCR Program, the term offense has a very particular definition that employs a series of rules defining the way offenses are counted.  Specifically, the number of offenses for crimes against persons is determined by the number of victims, while the number of offenses for crimes against property and society is based on each distinct operation. 

This study also introduces a nontraditional counting method that counts the number of records associated with the offense segment.  Within any incident reported in NIBRS, there is only one record reported for each unique offense code.  In NIBRS, there are 46 Group A offense codes in which full incident information is collected and 11 Group B offense codes where only arrestee information is reported.  This study uses the number of offense records in Table 2 and Appendix B which shows weapon type by offense type.  The effect of this counting method is that the number of victims is not considered when counting the number of offense records.  (See Figure 2.)

Figure 2, on the following page, illustrates how offenses and offense records are counted in NIBRS for a hypothetical incident with three victims of homicide who were also robbed.  The left of the figure shows the counts of offenses which include three counts of homicide, Code 09A, and one of robbery, Code 120, for a total of four offenses.  Because homicide is a crime against persons, one offense is tallied for each victim, while robbery, a crime against property, only counts one offense for each distinct operation, regardless of the number of victims.  However, on the right of the figure where offense records are counted, only one count of homicide and one of robbery are used to determine the sum of offense records.  Therefore, the number of offense records equals two, one for each unique offense type in this example.

 

Because this study focuses on offenders and arrestees, this approach is beneficial, particularly when examining weapon type by offense type (see Appendix B), since it eliminates the natural weighting that occurs when using traditional UCR offense counting rules based on the number of victims.  For example, because the weapon type is associated with the offense segment in NIBRS, an incident involving three murder victims has three offenses connected to the weapon type, overestimating the presence of that weapon type.  However, if the offense type is maintained as the unit of analysis for weapon type, the weighting of weapon types by the number of victims is avoided.  Of greater relevance to the objective of this study are the types of offenses that offenders commit in schools.  Certainly, studies involving NIBRS data that focus on the victim or offense segments require the examination of offenses based on the traditional UCR offense counting practices.

School Violence Index