Missing,Abducted,

Runaway, andThrownaway

Childrenin America

First Report:Numbers and CharacteristicsNational Incidence Studies

EXECUTIVE SUMMARY

May 1990

David Finkelhor, Ph.D.University of New Hampshire

Gerald Hotaling, Ph.D.University of Lowell

Andrea Sedlak, Ph.D.Westat, Inc.

This research was conducted for the U.S. Department of justice. Office ofJuvenile justice and Delinquency Prevention, Cooperative Agreement#87-MC-CX-K069 under the direction of Gerald T. Hotaling and DavidFinkelhor, Co-Principal Investigators, and Andrea j. Sedlak, Westat, Inc.,Project Director.

Points of view or opinions expressed in this publication are (hose of theauthors and do not necessarily represent (he official position or policies o!the U.S. Department of justice.

The Assistant Attorney General. Office of Justice Programs, coordinatesthe activities of the following program Offices and Bureaus: The Bureauof Justice Statistics. National Institute of Justice. Bureau of Justice Assis-tance, Office of Juvenile Justice and Delinquency Prevention, and theOffice for Victims of Crime.

T he release of the First Report from theNational Incidence Studies, Missing,Abducted, Runaway, and Thrownaway

Children in America, marks the beginning of anew era of better understanding of the extent andnature of these problems. For nearly a decade,the lack of accurate information on missingchildren in America has hampered the develop-ment of policies and allocation of resources.

The Office of Juvenile Justice and DelinquencyPrevention sponsored this seminal study of miss-ing children by the University of New Hampshirein response to a congressional mandate of theMissing Children's Act. Through this and otherresearch, the Office has come to recognize thatthe problem of missing children is not singular,nor is it wholly separated from the problems ofdelinquency with which the Office also deals. Asis true of the latter, the incidence of missingchildren is composed of different social problemsgreatly stemming from the weakening of theAmerican family.

Effectively preventing and dealing with themultifaceted problems of missing children

requires accurate, reliable information. There-fore, the incidence studies focused on identifyingrisk factors, on the children's experiences, andon the responses of parents and police.

The release of these first findings culminates a5-year effort. While these studies were carefullydesigned to answer as many questions as possible,we also recognize that the results may raise newquestions for which answers will be needed. TheOffice of Juvenile Justice and Delinquency Pre-vention is continuing to sponsor research to findthe facts and to develop useful programs that willprotect children and reduce delinquency.

The point to remember is that all of the numbersin this study represent real children, real lives, realneeds. The value of our research is in helpingsuch children to lead safe and normal lives.

Robert W. Sweet, Jr.Administrator

We are very grateful to the followingindividuals who assisted in a numberof phases of this project from initial con-

ceptualization and the development of definitionsto instrument development and review ofthe final.report.

Michael AgopianJoel BestJohn BlairJune Busey|an ChaikenJames CollinsHoward DavidsonCharles CowanGloria DeHartRoberf Figlioloan FisherBrian ForstJames GarafoloJames GarbarinoRichard CedesChris HatcherRobert HeckRuth HeronDavid HuizingaCharles KindermanMary Ann KiserJanet KosidKenneth LanningJohn Laub

Paul LavrakasDavid LloydGreg LokenArlene McCormackJohn McDermottMembers of the Family

Research LaboratorySeminar, 1988-1989

David MooreJohn RabunW. Ray RackleyLenore RadloffDarrell RegierCarl RogersSusan RosenbaumPeter RossiJudith SchretterMonroe SirkinMurray StrausSeymour SudmanDavid SugarmanRobert TinariJoseph WetsKirk Williams

Special thanks is due to Barbara Allen-Hagen,Social Science Analyst with the Office of JuvenileJustice and Delinquency Prevention, who workedclosely with the project since its inception andmade many useful comments and suggestions.We are also indebted to other OJJDP officials forsupport at various junctures, including formerActing Administrator Terrence Donahue; PamSwain, the Director of the State Relations andAssistance Division; Irving Slott, the Director of theResearch and Program Development Division;James C. Howell, the Director of the InformationDissemination Unit; and D. Elen Grigg, Publica-tions Coordinator.

We also want to thank and acknowledge the con-tributions of the following people at the Universi-ties of New Hampshire and Lowell for their assis-tance in data collection, data analysis, manuscriptpreparation, and editing.

Dianne CarmodyKathy ColeBruce DayJohn FitzgibbonsRobert FlewellingKaren GartnerMaggie Lowe

Joni MontcmagnoPeggy PlassSabrina RfggiLynn RuggeriNancy StrapkoJohn WhitehurstDonna Wilson

Jenny BonaRoger Campos

Lisa MacKenzieMary Masterson

We also want to thank the staff of the HouseholdSurvey and Police Records Study at Westat.

CATI System Development and OperationPhilip ("Tip") Dow, DirectorMichelle Hall, Linda Gowen, and Mary Usher, CATI

ProgrammersFelipe Lopeceron, CATI Pretesting

Dala Coding and EditingValerie Hudock, Coding Supervisor and Data ManagerJos Richards and Carol Reed, Coders and EditorsRay Gannon, Coder

Field Director, Police StudyJoanne McFarland

StatisticianLeyla Mohadjer, Ph.D.

Telephone Center OperationsDeborah Bitiner, Director of Training, Operations, and

PersonnelAnne Denbow, Project Coordinator

Codebook Development, Database Editing, andPublic Use Tape Preparation

Sandra Wodlinger, George Novosel, Beverly Lewis, andMarjorie Binzer

Computer Processing and Data Analyses

(ohn Brown, Systems Analyst

Monique Smith and Cynthia Diehm of AspenSystems Corporation provided help with editingand preparation of the report.

Finally, this research would not have been pos-sible without the support of a great many un-named individuals. Special thanks are expressedto all the highly skilled and dedicated telephonesupervisors and interviewers, to all the cooperativestaff at police agencies who contributed their time,to all the officials at juvenile facilities who com-pleted questionnaires about their agencies, to allthe technical and word-processing staff at UNHand Westat, and to all the survey respondents whocontributed their time and experience toward thesuccess of this effort.

ajor Conclusions

• What has in the past been called themissing children problem is in reality a setof at least five very different, distinctproblems. Each of these problems needsto be researched, analyzed, and treatedseparately.

• Many of the children in at least four ofthese categories were not literally missing.Caretakers did know where they were.The problem was in recovering them.

• Because of definitional controversiesand confusion about the concept of missingchildren, public policy still needs to clarifythe domain of this problem. Which childrenand which situations should be included,what do they have in common, and whatare they to be called?

• Family Abduction appeared to be a sub-stantially larger problem than previouslythought.

• The Runaway problem did not appear tobe larger in 1988 than at the time of thelast national survey in 1975.

• More than a fifth of the childrenwho have previously been termed Run-aways should actually be consideredThrownaways.

• There were a large group of literallymissing children who have not beenadequately recognized by previous re-search and policy concerning missingchildren. These were children who weremissing because they got lost, or injured,or because they miscommunicated withcaretakers about where they would be orwhen they would be home.

T he National Incidence Studies of Miss-ing, Abducted, Runaway, and Thrown-away Children (NISMART} was undertaken

in response to the mandate of the 1984 MissingChildren Act. Its objective was to estimate theincidence of five categories of children, thosewho were:

• Abducted by family members• Abducted by non-family members• Runaways• Thrownaways• Missing because they had gotten lost or

injured, or for some other reason.

NISMART collected data from six separatesources:

Household Survey. The centerpiece study was atelephone survey of 34,822 randomly selectedhouseholds, which yielded interviews with 10,544caretakers about the experiences of 20,505children. The response rate for eligible householdswas 89 percent. The modern sophistication ofsuch surveys allowed us to derive accurate na-tional estimates, while compensating for house-holds without phones and other nonparticipatinghouseholds.

Juvenile Facilities Survey. This was a survey ofresidential facilities, such as boarding schools andgroup homes, to find out how many children hadrun away from these facilities, in addition tochildren who ran from households.

Returned Runaway Study. This interview studywith children who had run away and returnedhome was primarily methodological. Its goal wasto find out if children's accounts of episodesmatched those of their parents.

Police Records Study. This was a study of policerecords in 83 law enforcement agencies in a

national random sample of 21 counties to find outhow many Non-Family Abductions were reported.

FBI Data Reanalysis. A reanalysis was conductedof 1 2 years of FBI homicide data to determinehow many children were murdered in conjunc-tion with possible abductions by strangers.

Community Professionals Study. This was a studyof 735 agencies haying contact with children in anational random sample of 29 counties to deter-mine how many children known to these agencieswere abandoned or thrown away.

Serious definitional controversies surround each ofthe problems studied, which made it necessary toestimate the incidence of each according to atleast two definitions. For example, in many Statesthe crime of abduction can entail the coercedmovement of a person as little as a few feet. Yetthe public thinks of abduction in terms of notori-ous crimes like the Lindbergh or Adam Walshkidnappings, where a child is taken a substantialdistance, for a substantial period of time, or withthe intent to keep or kill. Similarly, some Statelaws define parental abduction as an episode inwhich a parent takes a child or keeps a child forany length of time in violation of a custody de-cree. But the popular image of a parental abduc-tion is of a parent whc flees to anolher city oranother country with a child or who hides thechild incommunicado.

Thus, within each of the individual problems, wepresent incidence estimates according to at leasttwo definitions: what we call, first, a "BroadScope" and then a "Policy Focal" definition."Broad Scope" generally defines the problem theway the affected families might define it. It in-cludes both serious and also more minor episodesthat may nonetheless be alarming to the partici-pants. By contrast, "Policy Focal" generallydetines the problem from the point of view of the

police or other social agencies. It is restricted toepisodes of a more serious nature, where childrenare at risk and there is a need for immediate inter-vention. Policy Focal cases are a subset of BroadScope ones.

We have also created two definitions of non-family abduction: the Legal Definition Abductionsand Stereotypical Kidnappings. The Legal Defini-tion Abduction corresponds to the crime ofabduction as it is specified in the criminal law of

many States and includes the short-term, coercivemovement entailed in many rapes and assaults.Stereotypical Kidnappings, by contrast, reflectmore closely the popular stereotype of a kidnap-ping, as a long-term, long-distance, or fatalepisode.

These carefully crafted definitions were the resultof a three-stage process, involving a panel of 34experts, and a review of relevant legal statutes,law review articles, and prior studies.

F amily Abduction (Broad Scope) wasdefined (Figure FA-1) as situations wherea family member 1) took a child in violation

of a custody agreement or decree; or 2) in viola-tion of a custody agreement or decree failed toreturn a child at the end of a legal or agreed-uponvisit, with the child being away at least overnight.

A "family member," in addition to the usualmeaning, included anyone with a romantic orsexual involvement with a parent. Moreover,"abductions" could be perpetrated by custodial aswell as noncustodial caretakers. The incidenceestimates were based entirely on the HouseholdSurvey portion of NISMART.

There were an estimated 354,100 Broad ScopeFamily Abductions in 1988 {Figure FA-2). This isquite a bit higher than earlier guesstimates of25,000 to 100,000.

A Policy Focal Family Abduction was a moreserious episode, entailing one of three additionalaggravating conditions:

• An attempt was made to conceal the taking orthe whereabouts of the child or to prevent contactwith the child; or

• The child was transported out of State; or

• There was evidence that the abductor had theintent to keep the child indefinitely or to perma-nently alter custodial privileges.

There were an estimated 163,200 Policy FocalFamily Abductions in 1988, or 46 percent of theBroad Scope cases. Family Abduction had thelargest estimated incidence of any Policy Focalcategory in NISMART.

Most of the Broad Scope Family Abductions wereperpetrated by men, noncustodial fathers, andfather figures. Most victims were children fromages 2 to 11 with slightly more at younger ages,but relatively few infants and older teens. Halfinvolved unauthorized takings, mostly from thechildren's homes; half involved failures to returnthe child after an authorized visit or stay.

The most common times for Family Abductionswere in lanuary and August. These are the timeswhen school vacations end and visitations areexchanged. Most of the episodes lasted 2 days toa week, with very few, 1 0 percent, a month ormore. In only a tiny fraction, 1 percent or less,was the child still being held by the abductor.

The period immediately after a divorce was notwhen most Family Abductions occurred. Instead41 percent occurred before the relationship ended.Another 41 percent did not occur until 2 or moreyears after a divorce or separation. This was proba-bly because it took time for parents to develop newstable households, move to other communities,develop new relationships and become disen-chanted with the legal system—all factors thatcould precipitate abductions.

A number of figures give a sense of the scope ofthe most serious Broad Scope cases. In 1 out of10 cases the child was removed from the State.In 3 out of 10 cases the child experiencedserious or mild mental harm, according to thecaretaker. In about a third of the cases, there wasan attempt to conceal the child's whereabouts.In 4 out of 10 cases, the caretaker contacted the

police. In 5 out of 10 cases, the caretaker con-tacted an attorney. Although sexua! abuse is oneof the most feared components of family abduc-tion, its occurrence was unusual (less than 1 %).

Also of interest, in half the episodes, the caretakersdid know where the children were most of thetime. The problem was not discovering the where-abouts of the child, but getting the child returnedto proper custody.

There were interesting regional disparities in theoccurrence of Family Abduction, with the Southoverrepresented and the Midwest underrepre-sented. It is possible that the more traditionallegal system in the South makes noncustodialfathers pessimistic about getting a favorableoutcome, so that they take matters into theirown hands.

Much of the controversy about abduc-tions by non-family perpetrators hasreally been over definitions. To the

public, stranger abduction means notoriouscrimes like the Adam Walsh case. But in manyStates the legal definition of abduction is bothencompassing and broad, including the coercedmovement of many brief sexual assaults. Whenpeople hear high-incidence estimates based onthe legal definition, many are disbelievingbecause they are thinking of cases like the AdamWalsh kidnapping. Thus, NI SMART estimateswere made for both Legal Definition and Stereo-typical Kidnappings.

Legal Definition Non-Family Abduction meantthe coerced and unauthorized taking of a childinto a building, a vehicle, or a distance of morethan 20 feet; the detention of a child for a period

of more than an hour; or the luring of a child forthe purposes of committing another crime (FigureNFA-1). Many short-term abductions that tookplace in the course of other crimes like sexualassault were counted under this definition.

Stereotypical Kidnappings required that 1) thechild be gone overnight; 2) be killed; 3) be trans-ported a distance of 50 miles or more; 4) beransomed; or 5) that the perpetrator evidence anintent to keep the child permanently. The perpe-trator also needed to be a stranger. A contentanalysis of newspaper articles showed that 92percent of the crimes against children describedwith the words "abduction" or "kidnapping," infact, met one of the five criteria. This confirmedthat when reporters write and the public readsabout abduction, they are thinking primarily ofthe Stereotypical Kidnappings.

Because of small numbers, no reliable estimatescould be constructed from the HouseholdSurvey. So, the Non-Family Abduction estimatescame primarily from the Police Records Study. Inaddition, we developed an estimate of the numberof Stranger Abduction Homicides by reanalyzing12 years of FBI homicide data.

An estimated 3,200 to 4,600 Legal DefinitionNon-Family Abductions were known to lawenforcement in 1988 (Figure NFA-2).

Of these, an estimated 200 to 300 were Stereo-typical Kidnappings.

Based on FBI data, there were also an estimated43 to 147 Stranger Abduction Homicidesannually between 1976 and 1987.

Although we have a high degree of confidence inthe estimates for Abduction Homicides and Stere-otypical Kidnappings, we believe that the estimatefor Legal Definition Non-Family Abduction maybe low. A number of these Legal DefinitionAbductions may never be reported to the police,because the victims of these assaults or rapes areashamed or intimidated. Such episodes would notbe included in our estimate, which is based onpolice records.

Teenagers and girls were the most commonvictims of Non-Family Abduction. In LegalDefinition Abductions, half the victims were 12

or older and three-quarters of the victims weregirls. Blacks and Hispanics were heavily over-represented among victims compared to the U.S.population.

Two-thirds or more of the Legal Definition Abduc-tions involved sexual assault. A majority of victimswere abducted from the street. Over 85 percent ofthe Legal Definition Abductions involved force,and over 75 percent involved a weapon. Mostepisodes lasted less than a whole day; 1 2 to 21percent lasted less than an hour. In 2 percent thechild was still missing at the time of the last policeentry into the file. In 14 to 21 percent the childwas known to have been injured as a result of theabduction, but much information was missingfrom police records on this score.

The analysis of FBI homicide data from 1976 to1987 showed no discernible change in the rate forStranger Abduction Homicides over the 12-yearperiod.

In the Household Survey, caretakers did report alarge number of attempted abductions: an esti-mated 114,600, all involving strangers. Most ofthese consisted of an attempt by a passing motoristto lure a child into a car, and no actual harm oreven coercion against the child occurred. In amajority the police were not contacted. Yetchildren seem to have a fairly large number ofencounters with strangers where an abductionseems to have been threatened.

10

B road Scope Runaways were childrenwho left home without permission andstayed away overnight (Figure RA-1). In

addition, if children were already away andrefused to return home, they were also counted asRunaways, depending on their age and the amountof time away: 2 nights away If they were 1 5 orolder, and 1 night away if they were 14 or younger.The estimates for Runaways came from twosources: the Household Survey, and also from theJuvenile Facilities Survey, which counted childrenwho ran from institutions.

There were an estimated 446,700 Broad ScopeRunaways from households in 1988 (FigureRA-2). In addition, an estimated 12,800 childrenran from juvenile facilities (Figure RA-3),Because many children who ran from facilitiesalso ran from households during the same year,the joint number of Broad Scope Runawaysfrom households and facilities was estimated tobe 450,700 (the household estimate plus the4,000 who ran from juvenile facilities only)(Figure RA-4).

Policy Focal Runaways were Broad Scope childrenwho in the course of their episode were without asecure and familiar place to stay.

There were an estimated 129,500 Policy FocalRunaways from households (Figure RA-2). Sincewe considered all runaways from juvenilefacilities to be Policy Focal, the joint number ofPolicy Focal Runaways from households andjuvenile facilities was estimated to be 133,500(Figure RA-4).

There was no evidence of any higher level ofrunning away in 1988 than in 1975. A compari-son of NISMART results with results from the1975 National Statistical Survey on RunawayYouth, using very similar definitions and similarmethodology, showed almost exactly the same

rate of running away from households in bothyears.

Almost all Runaways were teenagers. They tendedto come disproportionately from step-parent-typehouseholds (where a parent was living with apartner who was not the child's other parent),compared to the occurrence of such households inthe general population. The runaways from facili-ties ran primarily from group foster homes, residen-tial treatment centers, and other mental healthfacilities.

Runaway episodes occurred more often in thesummer. Two-thirds of Broad Scope Runawaysfrom households ran to a friend's or relative'shome. Eighty-two percent were accompanied by-others during the episode. Half returned within 2days. For 39 percent, their caretakers knew theirwhereabouts most of the time.

Among the most serious Broad Scope householdcases, however, 1 child in 10 went a distance ofmore than 100 miles. One child in 14 went out ofState. And 1 in 10 was still gone from the home atthe time of the interview. Three percent had beensexually abused and 1 percent physically harmed.Thirty-six percent of the Broad Scope children hadrun away previously in the last 1 2 months. Thepolice were contacted in 40 percent of theepisodes.

The runaways from juvenile facilities tended tohave even more serious episodes. Almost one-halfleft the State. One-third were picked up by thepolice. One in 1 0 was placed in jail and 1 in 20 ina juvenile detention center. There were smallpercentages who got involved in prostitution (5%),drug dealing (3%), armed robbery (3%), and othercrimes. The staff of the facilities knew nothingabout the whereabouts of these children in half thecases. Police were contacted in 73 percent of theepisodes.

11

A child qualified as a Broad ScopeThrownaway if any one of four situationsoccurred: 1) the child had been directly

told to leave the household; 2} the child hadbeen away from home, and a caretaker refusedto allow the child back; 3) the child had runaway but the caretaker made no effort to recoverthe child or did not care whether or not the childreturned; or 4) the child had been abandoned ordeserted (Figure TA-1). In any case, the childhad to be out of the household for at least 1night.

The estimates for Thrownaways came from twosources: the Household Survey, and the Commu-nity Professionals Study, which was used toestimate the number of children who had beenabandoned.

There were an estimated 127,100 Broad ScopeThrownaways in 1988, including 112,600 fromthe Household Survey and 14,500 who wereabandoned based on the Community Profes-sionals Study (Figure TA-2).

A Policy Focal case was a Thrownaway whowas without a secure and familiar place to stayduring some portion of the episode. All theabandoned children were considered PolicyFocal.

There were an estimated 59,200 Policy FocalThrownaways in 1988.

Thrownaways constituted about 22 percent ofthe pooled group of Runaways and Thrown-aways, the group that in the past has simply beenlabeled Runaways.

The Broad Scope Thrownaways identified in theHousehold Survey were predominantly olderteenagers. By contrast, the abandoned childrentended to be young, one-half under the age of 4.Abandoned children, in contrast to the otherThrownaways, were also heavily concentrated inlow-income families. In either case, fewer childrenfrom households with both natural parents werethrown away or abandoned than would havebeen expected based on their proportion of theU.S. population.

Broad Scope Thrownaways from householdstended to occur in the spring. Fifty-nine percentwere preceded by an argument and 27 percent byviolence. Most Thrownaways went to the homesof friends and relatives. Most also stayed within a10-mile radius of their home. In 60 percent of thecases caretakers claimed to know the Thrown-away's whereabouts most of the time; but sincemany of these caretakers were not actually look-ing for the children, they may have expressed afalse degree of confidence.

Among the most serious episodes, a full fifth of theBroad Scope Thrownaways from households werestill gone from their households at the time of theinterview. This percentage of nonreturned chil-dren is higher than for any other category ofNISMART children. Thirteen percent spent somenight without a place to sleep. Fifteen percent hadbeen in a juvenile detention center.

Compared to Runaways, Thrownaways experi-enced more violence and conflict with theirfamilies and were more likely to still be gonefrom their home. Also, fewer caretakers inThrownaway cases contacted the police.

14

T his was a mixed group of children,missing from their caretakers for a varietyof reasons, who did not fit into the other

categories. Broad Scope cases were defined aschildren missing for varying periods of time(from a few minutes to overnight) depending onthe child's age, disability, and whether theabsence was due to an injury (Figure LOM-1).The estimate for Lost, Injured, or OtherwiseMissing was based entirely on responses fromcaretakers in the Household Survey.

There were an estimated 438,200 Broad ScopeLost, Injured, or Otherwise Missing children in1988 (Figure LOM-2).

Policy Focal cases were Broad Scope episodesserious enough that the police were called.

There were an estimated 139,100 Policy FocalLost, Injured, or Otherwise Missing children in1988, or 32 percent of the Broad Scopechildren.

Two groups of children experienced the largemajority of Broad Scope incidents: childrenunder 4 (47%), for whom even short absencescan be a source of alarm, and children 16-17(34%), who are the most independent, involved inrisky activities, and most likely to test, forget

about, or misunderstand the degree of responsi-bility they have to inform parents about wherethey are.

Children who truly lost their own way accountedfor only a small percentage (1 %) of this group.Injured children made up 6 percent. Thelargest subgroup consisted of children whoforgot the time, misunderstood expectations, orwhose caretakers misunderstood when thechildren would return.

Most of these episodes were short, a third lastingless than 6 hours. Only 2 percent of the childrenwere gone more than a day and only 1 percentwere still missing at the time of the interview.Half the episodes occurred in the summer.

Although many of the episodes in this categoryappeared relatively benign, a substantial fractionseemed quite serious. In 21 percent the childexperienced physical harm. In 14 percent thechifd was abused or assaulted in the course of theepisode. This is more harm than in any othercategory except Non-Family Abduction. More-over, \n 32 percent of Broad Scope cases, care-takers were alarmed enough that the police werecalled (these are the Policy Focal cases).

N ISMART drew two important conclusionsconcerning the overall "missing children"problem.

• Although the five problems studied here areoften grouped together as one—"missing chil-dren"—in fact, they are extremely dissimilar socialproblems. They affect different children and differ-ent families. They have very different causes,different dynamics, different remedies, differentpolicy advocates, and different types of institutionsand professionals who are concerned. They couldnot be lumped together for meaningful scientificanalysis.

• There was a second serious obstacle to group-ing these five categories of children under the

rubric "missing children'': not all these childrenwere literally missing. As the studies revealed, alarge proportion of the caretakers knew wheretheir children were most of the time during theepisodes. For example, in the case of familyabductions, only 17 percent of the children hadtheir whereabouts not at all known to caretakers(see figure). Many caretakers knew that the chil-dren were at the home of their ex-spouse, but theycould not get them back. In the case of runawaysfrom households, only 28 percent of the childrenwere entirely missing. Most runaways were knownto be at the homes of friends or relatives. Even inthe case of non-family abductions, most episodeswere so short-lived, as in the case of an abductionand rape, that the child may not have beenmissed by anyone.

Thus, we determined that it was not possible todevelop a meaningful and useful global figure forthe ''number of missing children." First, because ofthe profound differences among the problems, itdid not make sense from a scientific standpoint toadd together such disparate episodes as runaways,stranger abducted children, parentally abductedchildren and so forth, or even some portion ofeach of these problems, into a single number ofso-called missing children. Second, children inthese categories were "missing" in different senses,and In many cases, as we pointed out earlier, notmissing at all. Finally, when such numbers asthese have been lumped together in the past, ithas created a great deal of confusion. People haveassumed that missing children meant childrenwho had been abducted or who had permanentlydisappeared. Thus, all the statistical findings andconclusions of this study are made about the fivedistinct social problems, and there are no globalfigures. We specifically discourage anyone fromtrying to create or use such a global number onthe basis of N1SMART statistics.

We offer the following recommendations:

• Public policy around what has become knownas "missing children" needs to clarify its domain. Itneeds to be more specific about which childrenand which situations are included, why they areincluded, and what they are to be called. If thefive problems studied here need an overarchingframework, we propose the compound term"Missing and Displaced," rather than the simpleterm "Missing."

• Public policy needs to more clearly differenti-ate each of the separate social problems includedunder the so-called "missing children" umbrella.

• Increased attention needs to be given to theproblem of Family Abduction. The incidence ofthis problem proved larger than earlier estimates,and its 163,200 Policy Focal cases were the mostnumerous of all Policy Focal categories. FamilyAbductions may well be on the rise and yet couldbe readily amenable to prevention.

• All policy, publication, and research on theproblem of Runaways should take into accountthe difference between Runaways and Thrown-aways. Thrownaways are a large group withdifferent dynamics; they suffer from being lumpedtogether indiscriminately with Runaways.

• There needs to be special attention and anincreased policy focus on the problem of childrenwho run away from institutions. These childrenare among the most chronic runaways and theones at highest risk of becoming crime victimsand perpetrators; they need a specializedapproach.

• New attention should be given to the prob-lems of children who fell into our category of Lost,Injured, or Otherwise Missing. This group, asnumerous in total as Runaways, experiencedsubstantally more physical harm than any othercategory except those who were victims of Non-Family Abductions. The 139,000 childrenreported to police in this category are almost asnumerous as the Runaways reported to police.Some of the children in this category probablyexperienced quite minor episodes, but otherswere very serious cases. A policy about missingchildren needs especially to include the seriousgroup in this category.

• Another set of incidence studies should beundertaken 5 years from now, conducted largelyalong the lines of the present approach with a fewmodifications. These modifications would includea more comprehensive canvass of police records,a more direct sample of juvenile facilities, and aplanned coordination with future child abuse andneglect incidence studies. In addition, we urgethat interim methodological studies be undertakento improve the future incidence efforts.

• The Department of justice should consider thepossibility of ongoing data collection systems, forexample, using the National Crime Survey or apolice-based "sentinel" system that could provideyearly incidence statistics for some categories ofmissing and displaced children.

19