on't take candy from strangers." We all remember our parents passing on these words of wisdom with the hope that they would protect us from harm. Wouldn't it be wonderful if life were that simple? Unfortunately, children are at risk of abduction and sexual victimization, and most of the individuals who perpetrate these crimes are not perceived as strangers by their victims. Are traditional child safety messages effective, accurate, and complete? Do they adequately warn children about the threats to their safety? Do they unduly frighten children and parents? Are we giving children information that makes them more vulnerable to victimization rather than less? To answer these questions, the National Center for Missing and Exploited Children (NCMEC) reviewed existing research and its own data base of long-term abduction cases that do not involve family members. This review helped test long-standing child-protection messages while providing a basis for creating more effective messages. An Underreported Problem Child victimization is a large and underreported problem. Too many times, problems are not found because no one is looking for them. In recent years, we have finally begun to look: "Considerable evidence exists to show that at least 20% of American women and 5% to 10% of American men experienced some form of sexual abuse as children" (Finkelhor, 1994). "Most sexual abuse is committed by men (90%) and by persons known to the child (70% to 90%), with family members constituting one-third to one-half of the perpetrators against girls and 10% to 20% of the perpetrators against boys" (Finkelhor, 1994). According to the U.S. Department of Justice, teenagers and girls are among the most frequent victims of sexual attacks (Masquire and Pastore, 1997). The U.S. Department of Justice also estimates that the victims of two-thirds of imprisoned sexual assault offenders are younger than the age of 18 (Greenfield, 1997). According to the National Victim Center, 29 percent of rape victims are younger than 11, and 32 percent are between 11 and 18 years of age (National Victim Center, 1992). According to the Washington State Attorney General's Office, the average victim of abduction and murder is an 11-year-old girl who is described as a low-risk, "normal" child from a middle-class neighborhood who has a stable family relationship and whose initial contact with an abductor occurs within a quarter of a mile of her home (Hanfland, Keppel, and Weis, 1997). The Office of Juvenile Justice and Delinquency Prevention's (OJJDP's) National Incidence Studies of Missing, Abducted, Runaway, and Thrownaway Children (NISMART) found that as many as 114,600 children reported attempted abductions by nonfamily members in 1988. An additional 3,200 to 4,600 successful abductions were reported to police (Finkelhor, Hotaling, and Sedlak, 1990). NISMART also found that two-thirds of the cases of nonfamily abductions reported to police, most of which were for relatively short periods, involved sexual assault (Finkelhor, Hotaling, and Sedlak, 1990). "Stranger Danger" Dr. David Warden, psychologist at the University of Strathclyde in the United Kingdom, evaluated the efficacy of child safety programs. He emphasized that the responsibility to identify a potential assailant cannot be left to the child alone (Kent, 1990): No matter how intelligent the child, he or she does not see the world through skeptical adult eyes . . . Children live very much in the present. They can't foresee someone's actions or judge their intentions, certainly not at primary school age. They have a very weak understanding of motives, they simply take someone at face value. The concept of stranger danger is difficult, because it clashes with the social constraints on children to be polite to adults. Research suggests that children don't really know what a stranger is. They feel that once someone tells his name, he ceases to be a stranger. Dr. Ray Wyre, a noted authority on the treatment of sex offenders and director of therapy at the Gracewell Institute in Birmingham, England, cautioned that "the first step in advising 'never talk to strangers' is to make sure that the child understands what a stranger is. Children might believe it means a person who looks odd, rather than someone they do not know." Dr. Wyre further observed that a child's image "of a stranger is different from an adult's. The person trying to ensnare them could seem caring and persuasive and not at all threatening. After ten seconds' chat, they are no longer a stranger to a child" (Rayment, 1991). On the HBO special "How to Raise a Street Smart Child" (1987), host Daniel J. Travanti asked, "Does your child know what a stranger is? The fact is most children just do not know. They think a stranger is someone threatening and evil. The problem with telling your children, 'don't talk to strangers' is that the bad guys don't always look bad." On the same cable program, young, elementary schoolchildren provided their definitions of a stranger: "A stranger sometimes wears a hat . . . sometimes a black or brown jacket and is a guy with a beard . . . some hair and a moustache and some glasses." "I think a stranger is like . . . a punk rocker that drinks beer all day and sits around in a vacant lot." "A stranger looks mean and ugly . . . a creep." "Big . . . bigger than you, bigger than most people." The concept is clearly a difficult one for a child to grasp. A neighbor, a familiar face in a child's daily routine, or someone the child's parents know well enough to speak to or whose name the child knows is probably not regarded as a stranger. The Myth of the Stranger Research on the victim/offender relationship in child abduction/molestation cases is not new. Using a sample of 148 offenders who sexually assaulted youth and were sent for observation to a Massachusetts treatment facility, Groth and colleagues (1978) concluded that only 29 percent of the offenders studied were complete strangers to their child victims. In 71 percent of the cases, the offender and victim knew each other at least casually, and in 14 percent of the cases, the offender was a member of the child's immediate family. In 1985, research conducted by Dr. Gene Abel of Emory University in Atlanta examined a group of sex offenders. Funded by the National Institute of Mental Health, Dr. Abel's research concluded that the typical sexual offender against children is male, begins molesting by age 15, engages in a variety of deviant behavior, and "molests an average of 117 youngsters, most of whom do not report the offense." Dr. Abel emphasized that offenders seek legitimate access to children, noting that "child molesters seek out jobs to access kids" (Abel, 1985). Case in Point OJJDP, in conjunction with NCMEC, the Federal Bureau of Investigation (FBI), and the University of Pennsylvania School of Nursing (1990), sponsored the Case in Point series. The series used case studies to profile and analyze the methods of offenders who victimized children. For example, the series reported on a study of 157 abducting and nonabducting child molesters at the Massachusetts Treatment Center (MTC) in Bridgewater, MA, which was established in 1959 for the evaluation and treatment of sexually dangerous persons. State law provides that a person found guilty of a sexual offense can be committed to MTC for terms of 1 day to life if judged sexually dangerous. In the MTC study, the term "child molester" was defined as someone whose sexual offenses were against victims under the age of 16. "Sexual offense" was defined as any sexually motivated assault involving physical contact with the victim. When the victim-age criterion was not sufficient (for example, because the offender was young or because of multiple victims of varying ages), several additional guidelines were used. Age discrepancy between offender and victim was considered, as was the predominant age of victims and any other victim age trends. The study determined whether an offender was an abductor or a nonabductor based on detailed information in the offender's research file. Each offense was coded in terms of the place in which the offender initially encountered the victim and the place in which the offense actually took place. If the place of encounter and offense differed on 50 percent or more occasions for all known offenses, the offender was coded as an abductor. Offenders were coded as nonabductors when the place of encounter and offense differed on less than 50 percent of all known offenses. The MTC study examined a number of characteristics of child molesters (table 1). It is particularly significant to note the similarity of the MTC data to Groth and colleagues' results more than a decade earlier. The MTC study found that 66 percent of abducting child molesters and 80 percent of nonabducting child molesters were known to their victims. Table 1: Characteristics of Child Molesters Who Abduct Compared With Those Who Do Not Abduct   Abductors Nonabductors Number 97 60 Relationship to victims     Family member 8% 26%     Acquaintance 58% 54%     Not known by offender 33% 20% Average age of victims     Average 10.28 9.69     Standard deviation 3.30 2.91 Average number of victims while offender was still a juvenile     Average 0.52 0.20     Standard deviation 1.19 0.55     Range 0-6 0-3 Average number of victims in offender's adult years     Average 3.63 2.92     Standard deviation 3.22 2.41     Range 1-16 1-16 Total number of offenses committed with weapon present, but not used     Average 0.59 0.19     Standard deviation 0.94 0.44 Total number of offenses committed with weapon     Average 0.55 0.19     Standard deviation 0.91 0.44 Nonfamily Abductions In another study, NCMEC reviewed cases of nonfamily abduction (NFA). Because NCMEC serves as a resource center for law enforcement agencies and families in cases of missing and exploited children, its caseload is weighted toward long-term cases. If an abducted child is recovered quickly, the local agency is less likely to ask NCMEC for assistance. Therefore, the sample of cases chosen included the most serious child abduction cases in which there was the greatest threat to the child, including risk of loss of life, and in which the child was missing for a substantial period of time. The composition of this sample was such that it more likely included a higher share of abductions by "strangers." A sample of 260 NFA cases was selected. The following definitions were used: Stranger: An individual completely unknown to the child; someone with whom the child has had no prior contact of any kind. Acquaintance: An individual whom the child has seen on a regular basis or with whom the child may have had some contact, but does not necessarily know by name. Examples include babysitters, neighbors, custodians, workers at a school or apartment complex, children's group leaders or volunteers, teachers, coaches, cashiers at a grocery or drug store, friends of a parent, or other authority figures. All cases involving infants were eliminated due to the inability of the child to make a judgment about the abductor. Several other cases were eliminated due to subsequent information indicating that the cause of death was an accident or a suicide. The 260 NFA cases broke down as follows: The child was deceased and the abductor unidentified in 72 cases. The child was deceased and the abductor identified in 54 cases. The child was alive and the abductor unidentified in 34 cases. The child was alive and the abductor identified in 100 cases. For purposes of further analysis, the 72 cases in which the child was recovered deceased and the abductor remained unknown were eliminated. In another 33 cases, NCMEC was unable to develop information sufficient to make a confident categorization. Of the 155 NFA cases chosen for final analysis, the child was recovered alive in 106 cases (68.4 percent). In 89 of these cases (84 percent), the child knew or was acquainted with the abductor to some extent. In 33 of the 49 cases in which the child was recovered deceased (67.3 percent), the child knew or was acquainted with the abductor; in 16 of these cases (32.7 percent), the child did not know the abductor (table 2). Table 2: Acquaintance Abductions Child Recovered Alive Abductors Number of Cases Neighbors 4 Babysitters 12 Friend of family 29 Child's acquaintance 18 Authority figure 7 Parents/grandparents of child's playmate 2 Former assailant/ revictimizing child 7 Child Recovered Deceased Abductors Number of Cases Neighbors 6 Babysitters 2 Friend of family 12 Child's acquaintance 9 Authority figure 1 Parents/grandparents of child's playmate 2 Babysitter's friend 1 Fueling Fear Highly publicized cases of child victimization serve to heighten fears among children and parents. A 1987 Roper poll found that 76 percent of children "feared being kidnapped" -- their number one concern (Feinberg, 1987). In 1988, Peter Hart found that the second greatest perceived risk of parents regarding their children was "being kidnapped" (37 percent) (Colburn, 1988). In a 1997 survey conducted by the Princeton Survey Research Associates Poll, the top worry of parents is the fear that their child might be kidnaped or become the victim of violent crime. In the same survey, parents' fear that their child might become a victim of sexual abuse ranked fourth, just behind serious accident or illness (Kantrowitz, 1997). In 1991, Mayo Clinic pediatricians Gunnar B. Stickler, M.D., Daniel D. Broughton, M.D., and Anthony Alario, M.D., in conjunction with Margery Salter, Ph.D., published an extensive examination in Clinical Pediatrics (Stickler et al., 1991:527). The authors reported that 72 percent of parents feared "that their child will be kidnapped by a stranger" but noted that, "as in other violent crimes such as rape, physical abuse, and sexual abuse, a child is more likely to be abducted by someone known to the victim than by a stranger. Anticipatory guidance in these areas needs to be aimed more at interpersonal relationships than at 'stranger danger.'" Although intense media focus on the most extreme cases has led millions of Americans to define the missing and exploited children problem in terms of the rarest cases, some media have performed a public service by focusing national attention on the need for research, common definitions, and consistent reporting of missing children. Hidden Victims Child psychologist Robert L. Geiser (1979) observed in his introduction to Hidden Victims that "social problems have an uncanny ability to survive most attempts to remedy them. Their first line of defense is to hide from public awareness and then later to spring onto the scene as full-blown crises." Today, America has awakened to the problem of missing and exploited children. As Daniel J. Travanti observed in "How to Raise a Street Smart Child" (1987), "Ignorance scares a child more than knowledge does." The challenge is to create awareness of the risk faced by children and to avoid incomplete or inaccurate messages. NISMART provides an important starting point for understanding the full range of the problem. Armed with a more accurate picture of those who victimize children, we can provide more effective information to families to help parents keep their children safe. Rhetoric Versus Reality For generations, our fundamental messages to children have contained three basic premises. "Don't take candy from strangers." As indicated above, in at least two of three cases, the offender is not a stranger in the mind of the child. Usually, the victim and offender know each other, at least casually. Child molesters often seek legitimate access to children and then victimize them through a process similar to seduction. This reality does not make the message wrong, only grossly inadequate in providing protection for children, who need more comprehensive information about the dangers they are far more likely to face. "Don't be a tattletale." One of the most stigmatizing names that a child can be called is tattletale. From their earliest moments, we consciously and subconsciously encourage children not to communicate. Thousands of children are hidden victims, and the key to prevention and detection is communication. Children must be taught that if something is happening in their lives that they do not feel right about or that makes them feel uncomfortable, they must tell somebody they trust. "You're just a kid. Be respectful to adults; they know what they're doing." With this final message, we face a delicate challenge. All parents want their children to be polite and respectful to adults. Our message is not that we want children to be disrespectful, but that we must empower them to realize that they have the right to say no to those who would abuse their authority as adults. As educational consultant Stephanie Meeghan aptly expresses during many of the training sessions for teachers that she has held since 1988, "We must make children aware that their safety is more important than good manners." Combating Fear With Facts America's families need not live in fear, but parents need to be fully informed about the dangers their children face and the most effective ways to educate them and guard them from harm. The key to child safety is communication. Children should recognize that "strangers" often do not look strange, and parents should recognize that most abductions and assaults involve an offender and victim who know each other. The exaggerated fears of "stranger danger" generated by lurid tabloid headlines need to be replaced with solid facts garnered from serious research. References Abel, G.G. 1985. The Evaluation of Child Molesters: Final Report to the Center on Antisocial and Violent Behavior. Rockville, MD: National Institute of Mental Health. Colburn, D. 1988. Child safety emphasis of national campaign. The Washington Post (January 22):A12. Feinberg, L. 1987. Poll of schoolchildren reflects satisfaction, positive attitudes. The Washington Post (March 11):A12. Finkelhor, D. 1994 (Summer/Fall). Current information on the scope and nature of child sexual abuse. The Future of Children: Sexual Abuse of Children 4(2):31-53. Finkelhor, D., Hotaling, G., and Sedlak, A. 1990 (May). Missing, Abducted, Runaway, and Thrownaway Children in America, Washington, DC: U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention. Geiser, R.L. 1979. Hidden Victims: The Sexual Abuse of Children. Boston, MA: Beacon Press. Greenfield, L.A. 1997 (February). Sex Offenses and Offenders. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, p. iii. Groth, N.A., Burgess, A.W., Birnbaum, H.J., and Gary, T.S. 1978. A study of the child molester myths and realities. Journal of the American Criminal Justice Association 41:17-22. Hanfland, K.A., Keppel, R.D., and Weis, J.G. 1997 (May). Case Management for Missing Children Homicide Investigation: Executive Summary. Olympia, WA: Office of the Attorney General, State of Washington and U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention, p. 2. How to Raise a Street Smart Child. 1987. Videocassette. New York, NY: Home Box Office Video. Kantrowitz, B. 1997. Off to a good start: Why the first three years are so crucial to a child's development. Newsweek (Spring/Summer):8. Special Issue. Kent, A. 1990 (August 15). The reality of the nightmare. The Times (England). Masquire, K., and Pastore, A.L., eds. 1997. Bureau of Justice Statistics Sourcebook of Criminal Justice Statistics -- 1996. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Assistance, pp. 210-211. National Victim Center. 1992 (April). Rape in America: A Report to the Nation. Arlington, VA: National Victim Center, p. 3. Rayment, T. 1991 (August 18). Stranger danger. The Sunday Times (England). Stickler, G.B., Salter, M., Broughton, D.D., and Alario, A. 1991. Parents' worries about children compared to actual risks. Clinical Pediatrics 30(9):522-528. U.S. Department of Justice, National Center for Missing and Exploited Children, Federal Bureau of Investigation, and University of Pennsylvania School of Nursing. 1990 (August). Case in Point, Special Edition. Arlington, VA: U.S. Department of Justice, National Center for Missing and Exploited Children, Federal Bureau of Investigation, and University of Pennsylvania School of Nursing. Ernest E. Allen is president and chief executive officer of the National Center for Missing and Exploited Children (NCMEC). Since 1984, NCMEC has worked with OJJDP to provide technical assistance to parents, law enforcement, and other agencies working on missing and exploited children issues.