Family Violence Statisticsbjs.gov/content/pub/pdf/fvs03.pdf · Federal prosecution of domestic...

Bureau of Justice Statistics

U.S. Department of Justice Office of Justice Programs

Family Violence Statistics Including Statistics on Strangers and Acquaintances

Reported and unreported family violence

Murder of family members

Family violence reported to police

Family violence recorded by police

State prosecution of family assault

Federal prosecution of domestic violence

Family violence offenders in prison

Family violence offenders in jail

BJS

Note

This file contains the Fatal family violence section of Family Violence Statistics. The full report is available at http://www.ojp.usdoj.gov/bjs/abstract/fvs.htm.

U.S. Department of JusticeOffice of Justice Programs810 Seventh Street, N.W.Washington, D.C. 20531

Alberto R. GonzalesAttorney General

Office of Justice ProgramsPartnerships for Safer Communities

Tracy HenkeActing Assistant Attorney General

World Wide Web site:http//www.ojp.usdoj.gov

Bureau of Justice Statistics

Lawrence A. GreenfeldDirector

World Wide Web site:http://www.ojp.usdoj.gov/bjs

For information contactNational Criminal Justice Reference Service

1-800-851-3420

U.S. Department of JusticeOffice of Justice ProgramsBureau of Justice Statistics

Family Violence StatisticsIncluding Statistics on Strangers and Acquaintances

By Matthew R. DuroseCaroline Wolf Harlow, Ph.D.Patrick A. Langan, Ph.D.Mark Motivans, Ph.D.Ramona R. RantalaErica L. Smith

BJS Statisticians

Assisted byElizabeth Constantin

BJS Statistical Assistant

June 2005, NCJ 207846

Contents

ii Family Violence Statistics

Highlights, definitions, and counting rules 1DefinitionsSources of data on family violenceCounting rules

Reported and unreported family violence 8About the data in this sectionOffense type (table 2.1)Offense location (table 2.2)Victim demographics (table 2.3)Victim injury (table 2.4)Victim medical care (table 2.5)Offender demographics (table 2.6)Offender weapon use (table 2.7)Offender drug or alcohol use (table 2.8)Number of victims and offenders (table 2.9)

Fatal family violence 17About the data in this sectionVictim demographics (table 3.1)Offender demographics (table 3.2)Offender weapon use (table 3.3)Number of victims and offenders (table 3.4)

Family violence reported to police 22About the data in this sectionPercent of family violence reported to police (table 4.1)Reporting rates by offense (table 4.2)Reporting rates by victim demographics (table 4.3)Reporting rates by whether offender had a weapon (table 4.4)Reasons for not reporting family violence to police (table 4.5)Person who reported family violence to police (table 4.6)Arrest rate by offense (table 4.7)

Family violence recorded by police 29

About the data in this sectionOffense type (table 5.1)Offense location (table 5.2)Victim demographics (table 5.3)Victim injury (table 5.4)Offender demographics (table 5.5)Offender weapon use (table 5.6)Number of victims and offenders (table 5.7)Arrests by offense (table 5.8)Arrestee demographics (table 5.9)Victim demographics in crimes leading to arrest (table 5.10)Arrest rate by offense (table 5.11)Arrest rate by arrestee demographics (table 5.12)Arrest rate by victim demographics (table 5.13)Weapon possession at time of arrest (table 5.14)Disposition of juvenile arrestees (table 5.15)

State prosecution of family assault 45About the data in this sectionDefendant demographics (table 6.1)Prior arrests (table 6.2)

Prior felony arrests (table 6.3)Prior convictions (table 6.4)Prior felony convictions (table 6.5)Criminal justice status at time of arrest (table 6.6)Pretrial release (table 6.7)Bail amount (table 6.8)Time from arrest to pretrial release (table 6.9)Time from arrest to adjudication (table 6.10)Court outcome (table 6.11)Time from conviction to sentencing (table 6.12)Sentence imposed (table 6.13)Length of prison and jail sentences imposed (table 6.14)

Federal prosecution of domestic violence 51About the data in this sectionOffense type (table 7.1)Investigating agency (table 7.2)Profile of convicted offenders (table 7.3)

Family violence offenders in prison 53About the data in this sectionOffense type (table 8.1)Offense location (table 8.2)Victim demographics (table 8.3)Victim injury (table 8.4)Offender demographics (table 8.5)Offender weapon use (table 8.6)Offender drug or alcohol use (table 8.7)Number of victims and offenders (table 8.8)Federal, State, and local correctional facilitiesPrior sentences (table 8.9)

Family violence offenders in jail 61About the data in this sectionOffense type (table 9.1)Offense location (table 9.2)Victim demographics (table 9.3)Victim injury (table 9.4)Restraining orders among jail inmates convicted of family violenceOffender demographics (table 9.5)Offender weapon use (table 9.6)Offender drug or alcohol use (table 9.7)Number of victims and offenders (table 9.8)

Methodology 68Reported and unreported family violenceFatal family violenceFamily violence recorded by policeState prosecution of family assault (tables 10.1 and 10.2)Federal prosecution of domestic violenceFamily violence offenders in prisonFamily violence offenders in jail

Appendix 71Discrepant findings from two different ways of measuring family violence (tables 11.1 and 11.2)

Introduction

This compendium contains the mostrecent family violence statistics fromthese sources: surveys conducted bythe Bureau of Justice Statistics (BJS),the BJS database of Federal statistics,and two statistical databasesmaintained by the FBI.

The sources provide statisticalsnapshots of family violence at differentstages in the administration of justice. First are statistics on the nature andextent of family violence. Next arestatistics on family violence that isreported to police, followed by statisticson the prosecution of persons chargedwith family violence. Lastly are statisticson persons sent to prison or jail forfamily violence.

The report is divided into eight sectionsgiving statistics on —

• Reported and unreported familyviolence

• Fatal family violence

• Family violence reported to police

• Family violence recorded by police

• State prosecution of family assault

• Federal prosecution of domesticviolence

• Family violence offenders in prison

• Family violence offenders in jail.

Highlights

Trends in family violence

The rate of family violence fell between1993 and 2002 from an estimated 5.4victims to 2.1 victims per 1,000 U.S.residents age 12 or older. Throughoutthe period family violence accounted forabout 1 in 10 violent victimizations.

Reported and unreported familyviolence

Family violence accounted for 11% ofall reported and unreported violencebetween 1998 and 2002. Of theseroughly 3.5 million violent crimescommitted against family members,49% were crimes against spouses, 11%were sons or daughters victimized by aparent, and 41% were crimes againstother family members.

The most frequent type of familyviolence offense was simple assault.Murder was less than half of 1% of allfamily violence between 1998 and 2002.

About three-fourths of all family violenceoccurred in or near the victim'sresidence.

Forty percent of family violence victimswere injured during the incident. Of the3.5 million victims of family violencebetween 1998 and 2002, less than 1%died as a result of the incident.

The majority (73%) of family violencevictims were female. Females were84% of spouse abuse victims and 86%of victims of abuse at the hands of aboyfriend or girlfriend.

While about three-fourths of the victimsof family violence were female, aboutthree-fourths of the persons whocommitted family violence were male.

Most family violence victims were white(74%), and the majority were betweenages 25 and 54 (65.7%). Most familyviolence offenders were white (79%),and most were age 30 or older (62%).

Fatal family violence

About 22% of murders in 2002 werefamily murders. Nearly 9% weremurders of a spouse, 6% were murdersof sons or daughters by a parent, and7% were murders by other familymembers.

Females were 58% of family murdervictims. Of all the murders of females in2002, family members were responsiblefor 43%.

Children under age 13 were 23% ofmurder victims killed by a familymember, and just over 3% of nonfamilymurder victims.

The average age among sons ordaughters killed by a parent was 7years, and 4 out of 5 victims killed by aparent were under age 13.

Eight in ten murderers who killed afamily member were male. Males were83% of spouse murderers and 75% ofmurderers who killed a boyfriend orgirlfriend.

In 2002 family murders were less likelythan nonfamily murders to involve afirearm (50% versus 68%). Parentswere the least likely family murderers to use a firearm (28%), compared tospouses (63%) or other family members(51%).

Among incidents of parents killing theirchildren, 19% involved one parent killingmultiple victims.

Highlights, definitions, and counting rules

Family Violence Statistics 1

Family violence reported to police

Approximately 60% of family violencevictimizations were reported to policebetween 1998 and 2002. The reportingrate among female victims was notsignificantly greater than the reportingrate among male victims.

The most common reason victims offamily violence cited for not reportingthe crime to police was that the incidentwas a "private/personal matter" (34%).Another 12% of non-reporting familyviolence victims did not report the crimein order to "protect the offender."

Among the 2.1 million incidents offamily violence reported to policebetween 1998 and 2002, 36% resultedin an arrest.


Family violence accounted for 33% ofall violent crimes recorded by police in18 States and the District of Columbiain 2000. Of these more than 207,000family violence crimes, about half (53%,or 110,000) were crimes betweenspouses.

Among crimes recorded by police, 2%of family violence involved a firearm,compared to 6% of nonfamily violence.A weapon was used in 16% of familyand 21% of nonfamily violence.

About 6% of all violent crime recordedby police in 2000 involved more thanone offender victimizing a lone victim.The exception was stranger crime, inwhich 14% of incidents involved multi-ple offenders victimizing a lone victim.

About 49% of family violence crimesrecorded by police resulted in an arrest.Males comprised 77% of suspectedfamily violence offenders arrested in2000.

State prosecution of family assault

Of the approximately 1,500 defendantscharged with felony assault during May2000 in the State courts of 11 largecounties, about a third were chargedwith family violence.

Among felony assault defendantscharged with family violence in Statecourts, 84% had at least one prior arrestfor either a felony or a misdemeanor(not necessarily for family violence),and 73% had been previously convictedof some type of felony or misdemeanor(not necessarily family violence).

Nearly half of felony assault defendantscharged with family violence werereleased pending case disposition.

Among the 1,500 felony assault cases,the probability of the case leading toconviction (felony or misdemeanor) wasgreater for family assault defendants(71%) than nonfamily assault defen-dants (61%).

State courts sentenced 83% of personsconvicted of assault (both family andnonfamily) to either prison or jail.Among felony assault defendantsconvicted in State courts —

• 68% of incarceration sentences for family assault were to jail

• 62% of incarceration sentences for nonfamily assault were to prison

• 45% of persons sent to prison forfamily assault received a sentence ofmore than 2 years, compared to 77%of nonfamily assault offenders sent toprison.

Federal prosecution of domesticviolence

Persons suspected of domesticviolence made up 4% of the total18,653 Federal suspects referred toU.S. attorneys for alleged violent crimesfrom 2000 to 2002.

Of the 757 suspects referred to U.S.attorneys for domestic violenceoffenses between 2000 and 2002, mostwere firearm-related domestic violenceoffenses rather than interstate domesticviolence offenses.

• The Bureau of Alcohol, Tobacco,Firearms and Explosives accountedfor 80% of all referrals for firearm-related domestic violence.

• The FBI accounted for 72% of allinterstate domestic violence referrals.

Federal courts convicted 90% of defen-dants adjudicated for an interstatedomestic violence offense.

Among defendants convicted in federalcourts —

• 79% of convictions were the productof a guilty plea, and the remaining21% were the product of convictionfollowing a trial

• most were male (96%), under age 40(67%), white (72%), and non-Hispanic(95%)

• 4 in 5 defendants had a prior adultconviction.

Of 47 Federal defendants sentenced foran interstate domestic violence offensebetween 2000 and 2002, 91% receiveda prison term with a median length of 60months.

2 Family Violence Statistics


Of the nearly 500,000 men and womenin State prisons for a violent crime in1997, 15% were there for a violentcrime against a family member.

Nearly half of all the family violenceoffenders in State prisons were servinga sentence for a sex offense against afamily member. More than three-quarters of parents convicted of aviolent crime against their son or daugh-ter were in prison for a sex offense.

Of the crimes for which family violenceoffenders were in prison —

• most were against a female (78%)

• more than half were against a childunder age 18

• more than a third were against achild under age 13.

About 90% of offenders in State prisonsfor family violence had injured theirvictim:

• 50% of family violence victims wereraped or sexually assaulted

• 28% of the victims of family violencewere killed

• 50% of offenders in State prisons forspousal abuse had killed their victims

• Of State prison inmates imprisonedfor a crime against their son or daugh-ter, 79% had raped or sexuallyassaulted the child, and another 10%had killed the child.

Among family violence offenders inState prisons in 1997 —

• most were male (93%)

• 6 out of 10 were white, while about aquarter were black

• about 80% were between ages 25and 54.

Among offenders whose incarcerationin State prisons was for family violence,23% had used a weapon to committheir crime. The comparable percentageamong State prisoners incarcerated fornonfamily violence was higher— 46%.


Convicted family violence offendersmade up about 22% of the nearly86,500 convicted violent offenders inlocal jails in 2002. Most (60%) of theseapproximately 18,700 jail inmates incar-cerated for family violence were in jailfor an aggravated assault.

Local jail inmates convicted of familyviolence reported that —

• their victims were predominantlyfemale (79%)

• nearly 30% of their victims wereunder age 18.

Among local jail inmates convicted offamily violence, 55% injured their victim.

Most convicted jail inmates serving timefor violence against a family member(88%) did not use a weapon during thecrime.

Among jail inmates convicted of familyviolence, 45% had been subject to arestraining order at some point in theirlife. About 18% were under an activerestraining order at the time of admis-sion to jail.


Definitions

Violent crime Unless indicated other-wise, statistics on violent crime in thisreport pertain to all forms of the follow-ing crimes: criminal homicide,completed and attempted rape, sexualassault (including threats), robbery,assault (including threats), kidnaping,intimidation, illegal abortion, extortion,cruelty towards child or wife, hit-and-rundriving with bodily injury, and miscella-neous crimes against persons (asopposed to crimes against property).

Family violence Unless indicated other-wise, family violence includes all typesof violent crime committed by anoffender who is related to the victimeither biologically or legally throughmarriage or adoption. A crime is consid-ered family violence if the victim wasthe offender’s current or former spouse;parent or adoptive parent; current orformer stepparent; legal guardian;biological or adoptive child; current orformer stepchild; sibling; current orformer step sibling; grandchild; currentor former step- or adoptive-grandchild;grandparent; current or former step- or

adoptive-grandparent; in-law; or otherrelative (aunt, uncle, nephew).

Nonfamily violence Unless indicatedotherwise, nonfamily violence includesall types of violent crime betweencurrent or former boyfriends andgirlfriends; between current or formerfriends and acquaintances; andbetween strangers.

Relationship of victim to offender Thedatabases used in this report all containsufficient information to permit identifi-cation of family violence cases.However, the types of information thatmake identification possible are notuniform. Some provide more relation-ship categories than others. Forexample, the Supplemental HomicideReports provides 28 different catego-ries of victim-offender relationship, whilethe National Crime Victimization Surveyhas 15 categories. Also, the databasesuse different terms to describe specificvictim-offender relationships. Forexample, one uses the category“employee/employer,” while anotheruses “colleague at work.”

In general, enough information wasavailable in each of the databases todistinguish six categories of victim-offender relationship: three familycategories (spouse, son or daughter,and other family), and three nonfamilycategories (boyfriend/girlfriend, friend/acquaintance, and stranger). Mostsections of the report present statisticson all six categories.

Sources of data on family violence

National Crime Victimization Survey (a BJS survey)

One data source used to document thenature and extent of family violence inthe United States is the BJS NationalCrime Victimization Survey, or NCVS,for 1998 to 2002. The NCVS compilesdata on family and nonfamily violencethrough biannual interviews with nation-ally representative samples of U.S.residents age 12 or older. In these inter-views, residents are asked if they werea recent victim of crime. Those whowere victims are then asked numerousquestions about the incident, such aswhere it occurred and whether theyknew the offender. Residents areencouraged to tell interviewers bothabout crimes that were reported topolice and about unreported crimes.

Because the survey data come frominterviews with victims, the NCVS hasno information on homicide. Thesurvey’s scope is limited to certainforms of nonfatal violence: rape andsexual assault, robbery, aggravatedassault, and simple assault. Based oninterviews with the Nation’s crimevictims, estimates are formed not onlyof how many of these crimes occureach year but also of the characteristicspertaining to the criminal incidents.Such characteristics include the numberof victims that obtained medical care fortheir injuries, the number that reportedthe crime to police, and the numberwhose assailant was a relative.


Data documenting the flow of family violence cases through the justice system

CrimeCrimereportedto police

Crimerecordedby police

ArrestProsecutionandadjudication

SentencingCorrections

National CrimeVictimization Survey (BJS)

SupplementaryHomicide Reports (FBI)

National Incident-Based ReportingSystem (FBI)

State Court ProcessingStatistics (BJS)

Federal JusticeStatistics Program (BJS)

Survey of Inmatesin State and FederalCorrectionalFacilities (BJS)

Survey of Inmatesin Local Jails (BJS)

Figure 1

Supplementary Homicide Reports (an FBI database)

For national statistics on familyhomicide for the year 2002, the sourceused in this report was the FBI'sSupplementary Homicide Reports, orSHR. The FBI compiles detailednational data on family and nonfamilyhomicide from the thousands of lawenforcement agencies across theNation. For each criminal homicide, theSHR record such information as theage and race of the offender, therelationship of the victim to the offender,and the type of weapon used in thekilling.

National Incident-Based ReportingSystem (an FBI database)

For statistics on family violence thatcomes to police attention and for statis-tics on arrests for family violence, thesource used in this report is thedatabase for the year 2000 from theNational Incident-Based ReportingSystem, or NIBRS. The database,compiled by the FBI, contains NIBRSdata from at least 1 police agency ineach of 18 States and the District ofColumbia. These jurisdictions coverabout 16% of the U.S. population anddo not include any areas with a popula-tion of one million or more.

The NIBRS data have information onvictims, offenders, persons arrested,and incidents of family and nonfamilyviolence. For example, information onvictims includes the type of injurysustained and the victim’s relationshipto the offender. Among the variousdetails available on offenders and arres-tees are their age, race, and gender.Offense characteristics available in theNIBRS data include the type of weaponused and the type of location where thecrime occurred.

State Court Processing Statistics (a BJS data collection)

No national data in the United Statesdescribe the processing of familyviolence cases from arrest through finaldisposition by a court. The availablealternative used in this report is anextract of the BJS data collection StateCourt Processing Statistics (SCPS).The SCPS data used are entirely frompolice and court records that trackedfamily and nonfamily assault cases in11 counties, from the filing of Statecourt charges in May of 2000 to theirfinal court disposition. These datainclude information about personsarrested and charged with family andnonfamily assault: their demographiccharacteristics, their prior arrest andconviction record, and their criminaljustice status at time of arrest. TheSCPS data also contain informationabout the processing of the assaultcases, such as type of pretrial release,adjudication outcome, and type ofsentence imposed.

Federal Justice Statistics Program (a BJS database)

Information regarding violations offamily violence-related Federal statutesthat were subsequently referred toFederal court come from the FederalJustice Statistics Program (FJSP), adatabase maintained by the Bureau ofJustice Statistics. FJSP provides annualdata on workload, activities, andoutcomes associated with Federalcriminal cases. Data for 2000 to 2002were acquired on all aspects ofprocessing in the Federal justicesystem, including the number ofpersons investigated, prosecuted,convicted, and incarcerated. The FJSPdatabase is a BJS database,constructed from files provided by theU.S. Marshals Service, the ExecutiveOffice for United States Attorneys, the

Administrative Office of the United States Courts, the United StatesSentencing Commission, and theFederal Bureau of Prisons.

Survey of Inmates in State and FederalCorrectional Facilities (a BJS survey)

For national statistics on persons inprison for family violence, this reportused the BJS Survey of Inmates inState and Federal CorrectionalFacilities, conducted in 1997. Thesurvey involves face-to-face interviewswith a nationally representative sampleof State prisoners. Persons in prison foreither family or nonfamily violence areincluded in the sample. Through inter-views with them, information is obtainedon their victims and on numerous othercharacteristics of the crime that broughtthem into prison, such as whether aweapon was used, and the location ofthe offense.

Survey of Inmates in Local Jails (a BJS survey)

For statistics on inmates who wereconvicted of family violence andsentenced to a period of incarcerationof less than 1 year, the source used inthis report is the BJS Survey of Inmatesin Local Jails. This is a periodic surveywhich describes the current offensesand offense characteristics of local jailinmates. Most recently conducted in2002, face-to-face interviews wereconducted with a nationally representa-tive sample of local jail inmates tocollect systematic information on thisspecial correctional population.Questions were asked regarding therelationship of the victim to the incarcer-ated offender, the gender, race/Hispanic origin, and age of victims andoffenders, injury to the victim, offenderuse of a weapon, offender substanceuse at the time of the crime, and theplace at which the crime occurred.


Uniform counting rules

Incidents of crime vary depending onthe number of offenders, number ofoffenses committed, and number ofvictims. For instance, in a singleincident where two men rape and rob awoman, there are two offenders, fouroffenses (a rape and a robbery commit-ted by each man), and a single victim.Such an incident differs from one inwhich a lone man assaults anotherindividual. That incident involves oneoffender, one offense, and one victim.

The various databases used in thisreport are not uniform in the amount ofinformation they contain about eachincident of crime. One of the databases(NIBRS) contains information on virtu-ally every offender, every victim, andevery offense in an incident. The otherdatabases contain less information. Forexample, offender characteristics avail-able in the prisoner database pertainonly to each individual incarceratedoffender. The incident that resulted inthat person’s incarceration may haveinvolved multiple offenders, but informa-tion on co-offenders was unavailable.

To improve the comparability of statis-tics across the databases, rules wereadopted to guide tabulations. All statis-tics in this report — that is, statistics onoffenders, on offenses, and on victims— follow the rule that each incident betreated as though it involved oneoffender, one offense, and one victim.

If an incident involved more than oneoffender and information (such as age,race, and gender) was available onmore than one offender, the availabledata on just one of the offenders weretabulated. Similarly, if an incidentinvolved more than one offense andinformation on the different offenses(such as the place where eachoccurred) was available, only the dataon one of the offenses were counted.

Likewise, if an incident involved morethan one victim and the databasecontained information (such as thevictim’s age, race, and gender) on morethan one victim, victim statistics used inthe analysis were based on just one ofthe victims.

Certain databases used in the report —NCVS, SHR, and NIBRS — werevictim-based. Choosing a particularvictim to characterize an incident wasunnecessary with these databases.Other data collections used in the report— Survey of Inmates in State andFederal Correctional Facilities andSurvey of Inmates in Local Jails — wereoffender-based. Choosing a particularoffender to characterize an incident wasunnecessary with these sources.

Where choices had to be made aboutwhich offender, which offense, or whichvictim characterized an incident, thechoices were guided by various hierar-chies. Information about the hierarchiesis summarized below.

In choosing a particular offender tocharacterize an incident, the choice wasguided by the victim’s relationship to theoffender. For example, in a singleincident in which a woman wasassaulted by her husband and astranger, the incident was treated as aspouse-on-spouse assault. Offenderstatistics for such an incident (such asage, race, and gender) thereforepertained solely to the characteristics ofthe husband; characteristics of thestranger were not tabulated.

Selecting the husband over the strangerto characterize the incident conforms toa rule adopted for this report that saysto select whichever offender is highestin the following victim-to-offenderrelationship hierarchy (shown in orderfrom highest to lowest, with columnheadings in italics):

Spousespouse and common-law spouseex-spouse

Son or daughterchildstepchild

Other familyparent and stepparentsibling and step-siblinggrandchildgrandparentin-lawsother relative

Boyfriend or girlfriendboy/girlfriendex-boy/girlfriendhomosexual partner

Friend or acquaintancechild of girl or boyfriendfriendneighboremployeremployeeacquaintancebabysitteeotherwise known

Strangerstranger

Not included in analysisrelationship unknown


When it was necessary to choose asingle victim to characterize an incident,the victim-offender relationship hierar-chy was also used. Again, the chosenvictim was the one highest in thehierarchy.

In choosing a particular offense tocharacterize an incident, the choice wasguided by the offense’s position in ahierarchy of offense seriousness. Theoffense highest in the hierarchy wasselected to characterize the incident.For example, in a single incident wherethe victim was raped and murdered, theoffense selected to characterize theincident was murder because murder ishigher in the offense seriousnesshierarchy than rape. The offenseseriousness hierarchy (from most toleast serious) generally corresponds tothe offense list shown in table 5.1.

As noted earlier, each incidenttabulated in this report was treated ashaving one offender, one offense, andone victim. Each incident was alsotreated as having one offense location,one victim injury (if any), one weaponinvolved (if any), one arrestee (if any),and one arrest offense (if any). Thecharacteristic’s position in a hierarchydetermined which characteristic (forexample, which injury) was tabulated.The characteristic with the highestposition was chosen.

Listed in order from highest to lowest,the offense location hierarchy is in table5.2; victim injury hierarchy, table 8.4;offender weapon use hierarchy, table5.6; and arrest offense hierarchy, table5.8. The arrestee hierarchy is the sameas the victim-to-offender relationshiphierarchy described above.


About the data in this section

This section compares family murdersto nonfamily murders. Statistics shownin tables 3.1 through 3.4, as well asthose summarized in the text, are allbased on police murder statisticscompiled by the FBI in the 2002 Supple-mentary Homicide Reports (SHR). The2002 SHR contains information on16,204 murders and nonnegligentmanslaughters that occurred throughoutthe United States in 2002. The victim’srelationship to the offender wasrecorded in 9,102 out of the 16,204murders. All data summarized in thissection pertain to these 9,102 murdervictims whose relationship to theoffender was recorded.

One reason the relationship of thevictim to the offender was not recordedfor all 16,204 murders is that manymurder investigations fail to uncover

the identity of the perpetrator. In 2002the murder clearance rate (a measureof how often police solve a murder) was 64%.

Family murder as a proportion of all murders

Of the 9,102 murder victims (withcomplete data on victim-offenderrelationship) in 2002, 21.5% (1,958victims) were killed by members of theirfamily (table 3.1). The 9,102 murdervictims consisted of 8.6% who werekilled by their spouse, 5.5% who weresons and daughters killed by a parent,7.4% who were killed by a familymember other than their spouse orparent, 7.3% who were killed by theirboyfriend or girlfriend, 45.2% who werekilled by a friend or acquaintance, and26% who were killed by a stranger.

Demographic characteristics ofmurder victims

Gender

Females were more likely than males to be victims of family murder. Femaleswere 50.9% of the U.S. population in2002 but 57.5% of the Nation’s victimsof family murder that year (table 3.1).Family members were responsible for 43.1% of all the murders of females in 2002.



aIncludes nonnegligent manslaughter.

56.943.1100Female87.2%12.8%100%Male

78.5%21.5%100%All murdersa

Nonfamilymember

FamilymemberTotal

Victim gender

Victim was offender’s —

Note: Murder includes nonnegligent manslaughter. Data identifying the victim's relationship to the offender were reported for 56.2% of 16,204 murders. Of these 9,102 murders with recorded relationships, victim gender was reported for virtually 100%; race for 98.8%; and age for 97.6%. Detail may not add to total because of rounding.Source: FBI, 2002 Supplementary Homicide Reports.

26.0%45.2%7.3%78.5%7.4%5.5%8.6%21.5%100%Percent of all murders

2,3624,1136687,1446715007871,9589,102Total murders

7.97.56.07.532.20.816.017.79.755 or older27.128.638.529.035.65.247.532.529.835 - 5426.426.330.226.79.94.227.515.424.325 - 3429.927.421.427.711.14.47.98.123.518 - 246.45.83.45.83.25.21.02.85.113 - 172.34.40.43.38.080.20.023.47.6Under 138.7%10.2%3.9%9.1%11.1%85.4%1.0%26.2%12.8%Under 18

Age

2.41.82.62.11.43.55.63.62.4Asian/Pacific Islander1.11.30.81.21.70.60.10.81.1

American Indian/Alaskan Native

3.53.13.33.33.04.15.84.43.5Other40.243.844.642.733.134.723.929.739.9Black56.3%53.1%52.0%54.1%63.9%61.2%70.3%65.9%56.6%White

Race

13.916.671.020.836.249.281.057.528.7Female86.1%83.4%29.0%79.2%63.8%50.8%19.0%42.5%71.3%Male

Gender

100%100%100%100%100%100%100%100%100%All murders

StrangerFriend oracquaintance

Boyfriendor girlfriendTotal

Otherfamily

Son ordaughter SpouseTotal

AllmurdersVictim characteristic

Nonfamily memberFamily memberPercent of murders in which the victim was the offender’s —

Table 3.1. Demographic characteristics of family murder victims compared to nonfamily murder victims in 2002,by relationship

Wives were more likely than husbandsto be killed by their spouse. Wives wereabout half of all spouses in the popula-tion in 2002 but 81% of all personskilled by their spouse.

Girlfriends were more likely thanboyfriends to be victims of murder.Girlfriends were about 50% of allboyfriend-girlfriend relationships but71% of all victims who were a boyfriendor girlfriend of the murderer.

Sons and daughters were about equallylikely to be killed by their parents. Sonsmade up about 50% of all offspring andabout 50% of all sons and daughterskilled by their parents.

In 2002 murders committed by friendsor acquaintances accounted for 4,113of the 9,102 murders (with completedata on victim-offender relationship)(table 3.1). Males were more likely thanfemales to be victims of this type ofnonfamily murder. While males were49.1% of the U.S. population in 2002,they were 83.4% of murder victimskilled by a friend or acquaintance. Thevictims of stranger murder were alsomore likely to be male (86.1%).

Race

Overall, whites were less likely to bevictims of murder than blacks. In 2002,whites (defined here as both Hispanicand non-Hispanic) were 80.7% of theU.S. population and 56.6% of murdervictims, while blacks (also defined asboth Hispanic and non-Hispanic) were12.7% of the U.S. population but 39.9%of those killed.

Of the 787 identifiable incidents ofspousal murder in 2002, whites were70% of those killed, blacks were 24%,and Asians (including Pacific Islanders)made up 5.6% of victims.

In the 500 identifiable incidents ofparents’ killing their children in 2002,whites were 61.2% of those killed,blacks were 34.7%, and Asians were3.5%.

Age

The percentage of children under age13 killed by a family member was largerthan that group's percentage of the totalU.S. population. In 2002 children underage 13 were 18.1% of the U.S. popula-tion and 23.4% of murder victims killedby a family member (table 3.1).

By contrast, children under 13 were asmaller percentage — 3.3% — ofnonfamily murder victims. Looking at itanother way, among all victims ofmurder who were under age 13, nearlytwo-thirds were killed by a familymember. Among victims age 13 orolder, fewer than 1 in 5 were killed by a member of their family.

An estimated 29.5% of the U.S. popula-tion in 2002 was between ages 35 and54, while 46.7% of the victims killed bya spouse were within this age range.The average age of spouse murdervictims was 41.

The average age among sons ordaughters killed by a parent was 7years, and 4 in 5 victims killed by aparent were under age 13.


Murder by siblings

Family members other than a spouse,son, or daughter accounted for 7.4%of the 9,102 murder victims in 2002. Among these 671 murders of otherfamily members, 18% — 119murders— involved a sibling victim.The remaining 82% of these murderswere against parents, in-laws, or otherfamily members.

In 2002, 72% of murders by siblingsinvolved a brother killing a brother and14% involved a brother killing a sister.An additional 14% of sibling murdersinvolved a sister killing a brother orsister.

aIncludes nonnegligent manslaughter.82.317.710013 or older34.3%65.7%100%Under 13

78.5%21.5%100%All murdersa

Nonfamilymember

FamilymemberTotalVictim age

Victim was offender’s —

Demographic characteristics of murderers

Gender

In 2002 nearly 9 in 10 murderers weremale (table 3.2). Males accounted for79.2% of family murderers and 92.7%of nonfamily murderers. Among offend-ers who committed murder against aspouse, 83.1% of the offenders weremale; against a boyfriend or girlfriend,74.6%; and against a stranger, 96.7%of the offenders were male.

Race/Hispanic origin

In about half of all murders in 2002, theoffenders were white. Sixty-five percentof family murderers were white; anadditional 30.7% were black; and 3.9%were some other race. Amongnonfamily murderers, 49.2% werewhite, 48% were black, and 2.8% weresome other race.

Age

Fewer than 1 out of 10 murderers wereunder age 18. About the same percent-age of family murderers (5.5%) asnonfamily murderers (6.8%) wereyounger than 18.

A larger percentage of spouse murder-ers (71.8%) were age 35 or older,compared to 45.2% of murderers whokilled a boyfriend or girlfriend. Inaddition, 28.1% of spouse murdererswere between 18 and 34, while 53.6%of offenders who killed a boyfriend orgirlfriend were in this age range.


Note: Murder includes nonnegligent manslaughter. Data identifying the victim's relationship to the offender were reported for 56.2% of 16,204 murders. Of these 9,102 murders with recorded relationships, offender gender was reported for 98.9%; race for 97.6%; and age for 94%. Detail may not add to total because of rounding.Source: FBI, 2002 Supplementary Homicide Reports.

26.0%45.2%7.3%78.5%7.4%5.5%8.6%21.5%100%Percent of all murders

2,3624,1136687,1446715007871,9589,102 Total murders

1.24.04.13.17.65.718.311.45.055 or older15.723.841.123.032.025.053.538.926.635 - 5426.727.532.227.725.233.924.026.927.525 - 3447.638.021.439.423.230.34.117.334.418 - 248.86.61.26.810.04.904.76.313 - 17

00.100.12.00.200.80.2Under 138.8%6.7%1.2%6.8%12.0%5.1%0%5.5%6.5%Under 18

Age

1.41.71.71.61.12.74.83.01.9Asian/Pacific Islander0.71.70.61.32.00.80.11.01.2

American Indian/Alaska Native

2.03.42.32.83.03.54.93.93.1Other50.846.348.848.034.134.225.730.744.3Black47.2%50.3%48.9%49.2%62.9%62.4%69.4%65.3%52.7%White

Race

3.36.625.47.312.538.116.920.810.2Female96.7%93.4%74.6%92.7%87.5%61.9%83.1%79.2%89.8%Male

Gender

100%100%100%100%100%100%100%100%100%All murders



Otherfamily

Son ordaughterSpouseTotal

AllmurdersOffender characteristic


Table 3.2. Demographic characteristics of family murderers compared to nonfamily murderersin 2002, by relationship

Weapon use and family murder

When an offender kills the victim, theweapon is usually a firearm, knife, orblunt object such as a club. Lessfrequent are murders that result fromthe offender’s use of hands, fists, orfeet. Murders can also involve the useof items that are not typically consideredweapons, such as poisons, narcotics, orincendiary devices.

In 2002 family murders were less likelythan nonfamily murders to involve afirearm (50.1% versus 67.7%) (table3.3). Parents were the least likely typeof family murderers to use a firearm(27.8%), compared to spouses (62.6%)or other family members (51%). Themost common way a parent killed a son

or daughter was by the use of hands,fists, or feet (43.7%). The use of afirearm in nonfamily murders washighest among stranger offenders(75.7%), compared to murders commit-ted by friends or acquaintances (66.4%)and by boyfriends and girlfriends(47.2%).

Of the 9,102 murders (with completedata on victim-offender relationship) in2002, 207 were offenders who killedtheir parents (not shown in table).Firearms, knives, or blunt objects wereused in about three-fourths of murdersinvolving parent victims. Hands, fists, orfeet were used to kill 16% of parentmurder victims. The remaining 8% ofparent murder victims, including 4%who were asphyxiated, were killed withother weapons.

Number of victims and offenders infamily murders

About 3 out of 4 murders involved oneoffender and one victim in 2002 (table3.4). Murders in which one offenderkilled one victim described 84.3% offamily murders and 73.7% of nonfamilymurders.

A single offender against a single victimwas most characteristic of spousemurder and boyfriend/girlfriend murder.Of the 787 identifiable incidents ofspouse murder, 90.6% involved aspouse killing his/her spouse and noone else. The comparable figure for the668 identifiable incidents of boyfriend/girlfriend murder was 92.7%.

Multiple victims of a lone offenderoccurred most often in murders of sonsor daughters by a parent. In the 500identifiable incidents of parents killingtheir children, 18.8% involved oneparent killing multiple victims (forexample, his children, or his wife andone of his children).



Note: Murder includes nonnegligent manslaughter. Data identifying the victim's relationship to the offender were reported for 56.2% of 16,204 murders. Of these 9,102 murders with recorded relationships, offender weapon use was reported for 96.3%. Detail may not add to total because of rounding.Source: FBI, 2002 Supplementary Homicide Reports.

2,3624,1136687,1446715007871,9589,102 Total murders

5.9%8.5%9.0%7.7%11.8%43.7%5.8%17.0%9.6%Hands, fists, or feet

0.20.50.30.40.84.90.71.70.7Other0.72.03.31.73.78.03.34.62.3Asphyxiation

00.40.50.30.81.10.50.80.4Narcotics0.50.61.90.71.21.10.71.00.8Fire/incendiary1.5%3.5%6.0%3.1%6.5%15.1%5.1%8.1%4.2%Other

4.6%6.2%5.0%5.6%8.1%5.3%5.0%6.1%5.7%Blunt object

12.3%15.3%32.8%15.9%22.6%8.0%21.5%18.6%16.5%Knife

3.14.33.13.83.11.83.52.93.6Other/not specified3.04.53.13.95.63.68.06.14.3Shotgun4.14.63.14.35.31.65.24.44.3Rifle

10.213.39.411.914.06.916.813.312.2Other65.553.137.855.837.120.945.936.851.7Handgun75.7%66.4%47.2%67.7%51.0%27.8%62.6%50.1%64.0%Firearm

100%100%100%100%100%100%100%100%100%All murders



Otherfamily

Son ordaughterSpouseTotal

Allmurders

Weapon useby offender


Table 3.3. Weapon use in family murder compared to nonfamily murder in 2002, by relationship

Note: Murder includes nonnegligent manslaughter. Data identifying the victim's relationship to the offender were reported for 56.2% of 16,204 murders. The number of victims and offenders was reported for 100% of the murders with recorded relationships. Detail maynot add to total because of rounding.Source: FBI, 2002 Supplementary Homicide Reports.

2,3624,1136687,1446715007871,9589,102Total murders

3.72.90.73.000.800.22.4Multiple offenders5.76.54.66.011.518.88.912.37.4One offender9.4%9.4%5.4%9.0%11.5%19.6%8.9%12.5%9.8%Multiple victims

23.716.11.917.34.06.40.53.214.3Multiple offenders66.974.592.773.784.574.090.684.376.0One offender90.6%90.6%94.6%91.0%88.5%80.4%91.1%87.5%90.2%One victim

100%100%100%100%100%100%100%100%100%All murders


Boyfriend or girlfriendTotal

Otherfamily

Son ordaughterSpouseTotalAll murders

Number of victimsand offenders


Table 3.4. Number of victims and offenders in family murder compared to nonfamily murder in 2002, by relationship

Reported and unreported familyviolence

Statistics comparing reported andunreported nonfatal family violence toreported and unreported nonfatalnonfamily violence come from theBureau of Justice Statistics (BJS)National Crime Victimization Survey(NCVS) for 1998 to 2002. The statisticsare based on data gathered fromresidents living throughout the UnitedStates, including persons living in groupquarters, such as dormitories, roominghouses, and religious group dwellings.Crew members of merchant vessels,Armed Forces personnel living inmilitary barracks, and institutionalizedpersons, such as correctional facilityinmates, were not included in the scopeof this survey. Similarly, U.S. citizensresiding abroad and foreign visitors tothis country were excluded. With theseexceptions, individuals age 12 or olderliving in units selected for the samplewere eligible to be interviewed.

In addition to single incidents of criminalvictimization, the NCVS records data onseries victimizations. A series victimiza-tion is defined as six or more similar butseparate crimes which the victim isunable to recall individually or describein detail to an interviewer. Surveyrespondents who indicated they wereseries victims were asked to providedetails for the most recent incident.In annual BJS reports on the NCVS,series victimizations are excluded fromthe published victimization estimates.However, because repeat victimizationis an important aspect of familyviolence, series crimes were included inthese analyses, counted as "1" victimi-zation to represent all the incidents inthe series.

NCVS data in this section are based ona nationally representative sample ofthe general population age 12 or olderwho experienced a violent victimizationthat occurred between January 1998

and December 2002. Data were aggre-gated over this 5-year period to obtainsample sizes large enough to detectdifferences between family andnonfamily violence victims, as well as todetect differences between family andnonfamily violent crimes based on thecharacteristics of the incident.

Additional information on weighting,along with other methodological details,can be found in the BJS Internet-onlyreport Criminal Victimization in theUnited States — Statistical Tables,2002 (NCJ 200561).


To compare family murder to nonfamilymurder, this report uses data from theFBI's 2002 Supplementary HomicideReports (SHR). Murder as defined hereincludes murder and nonnegligentmanslaughter which is the willful killingof one human being by another. Notincluded in the analyses are deathscaused by negligence, suicide, oraccident; justifiable homicides; andattempts to murder.

SHR data are based solely on policeinvestigation. Not all law enforcementagencies which report offense informa-tion to the FBI also submit supplementaldata on murder. At the time of analysis,the most recent SHR data containedinformation on 12,940 of the 16,204murders reported in the FBI's UniformCrime Report. To account for the totalnumber of homicides, the SHR datawere weighted to match national andState murder estimates prepared by theFBI. All victim-based analyses areadjusted in this manner.

While many law enforcement agenciesreport supplemental data on murder,one reason why much of the dataconcerning victim-offender relationshipmay not be reported is because nosuspects were identified. In 2002, SHRdata recorded information on the

relationship of the victim to the offenderfor 56.2% (9,102) of the 16,204murders committed that year.

The dataset utilized in this report wascompiled by James Alan Fox, theLipman Family Professor of CriminalJustice at Northeastern University. Thedataset, along with additional detailsabout imputation and estimation proce-dures, is available from the NationalArchive of Criminal Justice Data<http://www.icpsr.umich.edu/NACJD>,please refer to ICPSR 4179. After theanalysis for this report was completed,the SHR database was updated toinclude information on an additional2,117 murders, bringing the totalnumber of murders included in thedatabase to 15,057 of the 16,204murders in the United States in 2002.


To compare family violence recorded bypolice to nonfamily violence recorded bypolice, this report utilizes official policestatistics reported to the Federal Bureauof Investigation (FBI) in 2000. Lawenforcement agencies can report officialpolice statistics to the FBI's UniformCrime Reporting System in one of twoways:

1. the Summary reporting system

2. the Incident-based reporting system(NIBRS).

Under the Summary system, lawenforcement authorities aggregate thenumber of incidents by offense type andreport these totals to the FBI on amonthly basis. Under the NIBRSsystem, agencies provide an individualrecord for each reported crime, includ-ing detailed information regarding thevictim(s) and offender(s) involved,characteristics of the incident, and infor-mation on any arrests made in connec-tion with the incident. In this regard,NIBRS represents a more comprehen-sive and detailed crime reporting

Methodology


system, with the ability to capture awide range of information on specificincidents.

To compare the characteristics of familyviolence to nonfamily violence, onlycases with incident dates during calen-dar year 2000 were included in theanalysis. In addition, only incidents fromlaw enforcement agencies officiallycertified by the FBI to report NIBRSdata were included. Currently, noagencies serving jurisdictions withpopulations of one million or more arecertified to report NIBRS data. In 2000,agencies from 18 States and the Districtof Columbia submitted NIBRS data,representing 16% of the U.S. populationand 13% of crime.

Additional methodological details canbe found on the BJS website<http://www.ojp.usdoj.gov/bjs/nibrs.htm>, as well as at the Incident-Based Reporting Resource Center<http://www.jrsa.org/ibrrc>.

State prosecution of family assault To compare the prosecution of familyand nonfamily violence defendants, thisreport uses data from the BJS datacollection State Court ProcessingStatistics (SCPS). The most recentSCPS study collected data for 14,877felony cases filed during May 2000 in 40large counties selected to represent theestimated 54,590 felonies filed in theNation’s 75 most populous countiesduring that month. Results of that studyand information regarding sampling andother methodological details are avail-able in the BJS publication FelonyDefendants in Large Urban Counties,2000, NCJ 202021, December 2003.

Data from all 40 counties were not usedto compare the case processing offamily violence to nonfamily violence.Since assault was the only crime in the

2000 SCPS data to have specificstatutes that identified family violence,other violent crimes (such as murder,rape, or robbery) were excluded fromthe analysis. To be in this study, acounty also had to meet two criteria:

1. The county had an estimated 10 ormore assault defendants charged undera family or domestic violence statute(table 10.1).

2. The county had at least 17% of itsassault defendants charged withcommitting family or domestic violence.

A total of 11 counties met the selectioncriteria. For instance, Travis County,Texas, had an estimated 16 familyviolence cases which accounted for25% of the 64 assault cases filed in thatcounty during May 2000. Within theselected counties, a sample of 637cases was statistically weighted torepresent 1,517 family and nonfamilyassault defendants. Although theestimated 1,517 cases were less than3% of the 54,428 felony cases filed inthe Nation’s 75 largest counties in May2000, demographic characteristics ofboth cohorts were similar (table 10.2).

Where family assault defendants werecompared to nonfamily assault defen-dants in this section, sampling error was

taken into account. All differencesdiscussed were statistically significantat the .05 level.


Source: BJS, 2000 State Court Processing Statistics.

1,024493420217Total defendants

4816248Travis (TX)7616388Tarrant (TX)

15636399Harris (TX)181497Franklin (OH)

144307215Essex (NJ)106305315Pinellas (FL)48382419Santa Clara (CA)66543327San Bernardino (CA)721203660Riverside (CA)

2641206630Los Angeles (CA)26192619Contra Costa (CA)

Nonfamily Family Nonfamily FamilyCounty (State)

Weighted Unweighted

Table 10.1. Felony assault cases in the State courts of 11 countiesduring May 2000, by county and type of assault

*Excludes 162 weighted cases for which thearrest charge could not be classified.Source: BJS, 2000 State Court ProcessingStatistics.

54,428*1,517Totaldefendants

2.13.455 or older33.232.835 - 5430.931.525 - 3430.830.718 - 242.9%1.6%Under 18

Age at arrest

2.42.0Other22.732.6Hispanic45.439.2Black29.6%26.2%White

Race/Hispanic origin

19.015.3Female81.0%84.7%Male

Gender

100%100%Total

All cases inthe 75 largestcounties

Assaultcases in 11counties

Felony defendants in Statecourts during May 2000

Defendantcharacteristic

Table 10.2. Demographic distributionof felony defendants in the 11-countystudy compared to the Nation’s 75largest counties, 2000

Federal prosecution of domesticviolence

The source of data for statistics onFederal prosecution of domesticviolence is the BJS Federal JusticeStatistics Program database coveringthe years 2000 to 2002. The databaseis presently constructed from sourcefiles provided by the U.S. MarshalsService, the Drug Enforcement Admin-istration, the Executive Office for U.S.Attorneys, the Administrative Office ofthe U.S. Courts, the United StatesSentencing Commission, and the U.S.Bureau of Prisons. A description of thesource agency data files, along withother methodological details, can befound in the methodology section in theBJS Internet-only document Compen-dium of Federal Justice Statistics, 2002(NCJ 205368).


Information about offenders in prison forfamily violence were obtained fromBJS's 1997 Survey of Inmates in Stateand Federal Correctional Facilities.From June through October, 1997, anationally representative sample ofinmates in State prisons were inter-viewed about their current offense andsentences, criminal histories, family andpersonal backgrounds, gun possessionand use, prior drug and alcohol use andtreatment, educational programs, andother services provided while in prison.

Data and documentation with additionaldetails are available from the NationalArchive of Criminal Justice Data<http://www.icpsr.umich.edu/NACJD>.Please refer to ICPSR 2598.


The 2002 Survey of Inmates in LocalJails provides nationally representativedata on persons held in local jailsincluding those held prior to trial andconvicted offenders serving sentencesin local jails or awaiting transfer toprison. Data in the survey include theindividual characteristics of jail inmates;current offenses, sentences and timeserved; criminal histories; jail activities,conditions and programs; prior drug andalcohol use and treatment; and healthcare services provided while in jail. Inthe last Survey, completed in 1996, arepresentative sample of 6,000 inmateswere interviewed. The sample for the2002 survey was selected from 3,365jails housing 558017 males and 73,224females. The dataset and documenta-tion with additional details about thesurvey will be available from theNational Archive of Criminal JusticeData <http://www.icpsr.umich.edu/NACJD>.


Discrepant findings from twodifferent ways of measuringfamily violence

The extent of family violence in theUnited States is measured two differentways in this report. One way C throughthe National Crime Victimization Survey(NCVS) C is based on survey interviewswith samples of the U.S. population.The other way C through the FBI'sNational Incident-Based ReportingSystem (NIBRS) C is based on statis-tics compiled by police.

NCVS data summarized in this reportindicated that family violence makes up11% of all violence (both reported topolice and unreported) (table 2.1), butpolice statistics summarized in thisreport indicated that family violencemakes up nearly 33% of all police-recorded violence (table 5.1). No firmexplanation for this apparent discrep-ancy is offered in the report. Presuma-bly the discrepancy is caused by a hostof factors. The discussion that followsidentifies some of them, not necessarilythe most important ones.

One factor pertains to the willingness ofvictims (and others) to report crime topolice. If family violence is more likelythan nonfamily violence to be reportedto law enforcement authorities, thefamily violence percentage in policestatistics will tend to be larger than thepercentage in the statistics for allviolence (reported and unreported). Tocheck that possibility, reporting rates forfamily and nonfamily violence werecompared. Results indicated that familyviolence (59%) is more likely thannonfamily violence (46%) to be reportedto police (table 4.1). While the differ-ence is not large enough to fullyaccount for the 11%-versus-33%discrepancy between NCVS and policestatistics, results do suggest that adifference in reporting between familyand nonfamily violence is a contributingfactor.

A second factor pertains to the fact thatthe police statistics on family violenceused here are not directly comparableto the NCVS statistics in terms ofgeographical coverage. That is, thepolice statistics are based on data fromagencies reporting in 18 States and theDistrict of Columbia, while the NCVSstatistics are for the entire UnitedStates. If the family violence percentagein the 18 States and the District ofColumbia happens to be much largerthan other States, using these jurisdic-tions to compare to the NCVS willcreate the appearance of a greaterdiscrepancy than may actually exist.

A third factor pertains to the handling of"series" victimizations in the analysis ofthe NCVS data. The 11% familyviolence percentage from NCVS datawas based on an analysis that treatedeach "series" victimization (6 or morecrimes of a similar nature that the victimis unable to recall individually) as asingle victimization. Since familyviolence series victimizations comprised17% (rather than 11%) of all violentseries victimizations, treating eachseries victimization as 6 or more victimi-zations (rather than 1) would tend toraise the family violence percentageabove 11%. To illustrate, when thefamily violence percentage wasre-computed, this time treating eachseries using the actual number ofincidents (up to 20) indicated by thevictim, the percentage rose to 12.5%.

A fourth factor pertains to the possibilitythat the NCVS undercounts familyviolence to a greater extent thannonfamily violence. If so, that wouldmake the family violence percentagelook lower than it actually is. The littleresearch on the undercount that hasbeen done C notably, a small study ofassault victims whose victimization wasreported to police in San Jose, Califor-nia, and a national study of victims whowent to hospital emergency rooms fortheir injuries — compared undercounts

between family and nonfamily violence.Findings from both studies suggest thatthe NCVS undercounts family violenceto a greater extent than nonfamilyviolence. The two studies are summa-rized below.

San Jose study of known assaultvictims

In January 1971, a sample of victims ofassault was drawn from the records ofthe San Jose Police Department. Theseknown victims were interviewed andasked whether they had been a victimof crime at any time in 1970. Thevictims and their interviewers were nottold that the researchers wanted todetermine to what extent known assaultvictims tell interviewers about recentvictimizations.

Results suggested that family assaultvictims were less likely than nonfamilyassault victims to tell interviewers abouttheir victimization (table 11.1). While78% of 18 family assault victims did notmention being a victim of familyviolence, the comparable percentagefor 62 nonfamily assault victims waslower: 29%.

Appendix


Source: Adapted from U.S. Department of Justice, Law Enforcement AssistanceAdministration (1972). San Jose MethodsTest of Known Crime Victims, StatisticsTechnical Report No. 1. Washington, DC:USGPO, June.

29Nonfamily violence78%Family violence

Known assault victimswho did not tell NCVSinterviewers about theirvictimization

Relationship ofvictim to offender

Table 11.1. Among known assaultvictims, percent who told NCVSinterviewers about the victimization

The original purpose of the San Josestudy was to determine the effects ofthe passage of time on the recall ofcriminal victimizations. The surveydesign emphasized overall reportingissues, not the specific reporting offamily violence. Hence, one limitation ofthe study is the small sample size. Of126 sampled victims of assault, 18family assault victims and 62 nonfamilyassault victims were located and agreedto participate, a 63.5% response rate.

Another limitation pertains to the factthat the questionnaire used in the studywas more similar to the pre-1993 NCVSquestionnaire than the one in use today.Consequently, the San Jose studyresults may be more relevant to thepre-1993 version than today’s question-naire.

The pre-1993 version was replaced withone that was specifically designed toreduce undercounting of family (anddomestic) violence. There is someevidence that the redesign may havesucceeded in reducing the undercount.In 1991 (the last full year of interviewsusing the old questionnaire), familyviolence victims made up 7.3% of allviolence victims (8% if series victimiza-tions totaling 3 are counted as 3 crimesrather than 1; 4 counted as 4; and 5counted as 5). In 1993 (the first full yearusing the redesigned questionnaire),the percentage was 10.2%.

Hospital emergency room study

While results of the San Jose study onlyshed light on the extent to which police-reported family violence (family violencethat was brought to police attention) ismentioned in NCVS interviews, resultsof another study may shed light on theextent to which victims tell interviewersabout both reported and unreported

family violence of a particularly seriousnature. In this other study, two nationalnumbers were compared:

1. from the NCVS, the total (reportedand unreported) estimated number offamily violence victims who said theywent to a hospital emergency room fortreatment for their injuries in 1994

2. from a national survey — the Study ofInjured Victims of Violence (SIVV) —the total (reported and unreported,presumably) estimated number offamily violence victims whose injuriesbrought them to hospital emergencyrooms for treatment in 1994.

All other things being equal, the twonumbers should be the same if injuredfamily violence victims tell NCVS inter-viewers about any incidents that lead toemergency room treatment.

In fact, SIVV recorded far moreemergency room admissions in 1994than the NCVS, suggesting that theNCVS undercounts victims of seriousfamily violence. The SIVV number ofemergency room admissions for familyviolence (179,000) was found to be over5 times the number recorded by theNCVS (33,300) (table 11.2). Bycomparison, the SIVV number ofemergency room admissions fornonfamily violence (775,000) was 1.6times greater than the NCVS-estimate(471,400), suggesting that the NCVS

undercounts particularly serious typesof family violence to a greater extentthan nonfamily violence.

The relevance of the SIVV study islimited by the fact that its findingspertain to victims who were injured andwent to an emergency room. Suchvictims are a small percentage of allfamily violence victims. Furthermore,the statistics for the NCVS-documentedvictims of family violence (who went toemergency rooms) may be unreliablebecause they are based on a smallsample. Also, the SIVV study did notdocument how many victims had, andhow many had not, reported the crimeto police. Presumably, the SIVV studyincluded some of both.


*Age 12 or olderSource: Adapted from BJS, Violence-Related Injuries Treated in Hospital Emergency Departments, NCJ 156921, August 1997.

1.693.4471,40081.2775,000Nonfamily violence5.46.633,30018.8179,000Family violence

1.9100%504,700100%954,000Total

Ratio ofSIVV/NCVSPercentNumberPercentNumberType of offense

NCVS-estimatedadmissions tohospital emergencyrooms in 1994

SIVV-recordedadmissions tohospital emergencyrooms in 1994*

Table 11.2. Comparison of the Study of Injured Victims of Violence (SIVV) and the National Crime Victimization Survey (NCVS), by offense and by relationship, 1994

Family Violence Statisticsbjs.gov/content/pub/pdf/fvs03.pdf · Federal prosecution of domestic...

Documents

Transcript of Family Violence Statisticsbjs.gov/content/pub/pdf/fvs03.pdf · Federal prosecution of domestic...