Objective. To measure exposure to different types of violence among school-aged children in a primary care setting.
Design. Child interviews using an instrument measuring 4 types of exposure (direct victimization, witnessing, hearing reports, media). Violent acts measured include being beaten up, chased/threatened, robbed/mugged, stabbed/shot, killed.
Setting. Pediatric primary care clinic of large urban hospital.
Patients. Convenience sample of 175 children 9–12 years old and their mothers. A total of 53% of the children were boys, 55% were Hispanic, and 40% received public assistance.
Results. All children had been exposed to media violence. A total of 97% (170/175) had been exposed to more direct forms of violence; 77% had witnessed violence involving strangers; 49% had witnessed violence involving familiar persons; 49% had been direct victims; and 31% had witnessed someone being shot, stabbed, or killed. Exposure to violence was significantly associated with being male.
Conclusion. Most school-aged children who visited a pediatric primary care clinic of a large urban hospital had directly experienced violence as witnesses and/or victims.
- exposure to violence
- witness to violence
- victims of violence
- school-aged children
- pediatric primary care clinic
The United States is reported to be the most violent country in the western world.1 From 1985 to 1993, victimization and exposure to violence among children and youth dramatically increased.2,3 Although, in recent years, there has been a general decrease in violent crime rates around the country, the proportion of youth involved in crime compared with other age groups has risen.4 In 1996, about a third of all victims of violent crime were 12 to 19 years old and almost half were <25 years old.2 More than 750 children between 4 and 14 years old were killed by firearms in 1994.5 About 1.6 million children are abused or neglected every year leading to >1000 deaths.6
A larger, but often overlooked, group of children are the silent witnesses, who may not be physically harmed but may be psychologically affected by witnessing violent events.6–10 Published data from school-based studies in urban areas consistently show that inner-city school-aged children witness a high dose of community stabbings, shootings, and beatings.7,10–13 More than a quarter of African-American children surveyed in a Chicago elementary school11 had seen someone shot or stabbed. In a moderately violent neighborhood in Washington, DC, about 61% of school-aged children had witnessed some form of community violence.12Recent evidence shows that even children in suburban communities are exposed to violence. A study in a suburban middle school in Pennsylvania reported that 57% of the sixth graders surveyed had witnessed a robbing, beating, stabbing, shooting, or murder.14
Despite these reports, research in this field by pediatricians is lacking. Previous research has focused on violence in schools and in neighborhoods with high crime rates. This approach may have overestimated exposure of children to violence because large numbers of children may have been simultaneously exposed to the same violent events (eg, a single school-based setting or neighborhood event). Few data are available on rates of exposure to violence among children in a primary health care clinic, a setting more clinically relevant to the pediatrician. In 1 prior study, mothers of preschool children visiting a pediatric primary care clinic were asked about their children's exposure to violent acts.15 Our review found no prior research in this area among school-aged children attending a pediatric primary care clinic.
Exposure to violence may be an area that pediatricians may not ask about in a regular well-child visit. Therefore, they may fail to recognize that the children they see in the office may have had exposure to violence. Pediatricians may limit their focus to exposure to television violence. Noting that the average American child watches 23 to 27 hours of television and 1 to 2 films a week, which are increasingly more violent in nature,16 pediatricians may give advice to parents about monitoring what their children watch on television, unaware that more direct exposure occurs in the neighborhood, in school, or at home.
This study was conducted to assess the degree to which urban school-aged children who visit a large pediatric primary care clinic are exposed to different types of violence. By collecting first-hand data on the violent experiences of school-aged children who reside in this socially and economically diverse borough, we minimized the bias that results from school-based studies whose subjects may be collectively exposed to violent incidents that happen in or near the school. This report is the first part of a larger study that aims to look at exposure and victimization to violence and psychosocial adjustment among school-aged children who visit a pediatric primary care clinic. As an initial query, we asked the question: Is exposure to violence among school-aged children visiting a primary care clinic limited to violence on television or are school-aged children significantly exposed to other more direct forms of violence than media?
The study population consisted of a convenience sample of children 9 to 12 years old and their mothers who were recruited from the pediatric primary care clinic of a large urban hospital during the months of January 1997 to February 1998. Mothers and children were approached by a trained interviewer while waiting to be seen by a pediatrician during health care maintenance visits. Children were excluded if their mothers reported that they were mentally retarded, had a learning disability or were in special education, or had a physical disability or chronic illness. Non-English speaking families were also excluded because of the unavailability of a validated non-English version of the questionnaire. Only 1 child per household was enrolled. Parental consent and child assent forms were obtained before their participation in the study in accordance with procedures approved by our institutional review board.
This cross-sectional study consisted of a maternal interview to determine family background characteristics and a child interview about exposure to violence. Children were interviewed alone.
Children's Report of Exposure to Violence (CREV)
Children were interviewed using the CREV, a 29-item questionnaire developed by Cooley et al17 to measure exposure to community violence. The CREV was designed for children 9 to 15 years old, and has good test-retest reliability, internal consistency, and construct validity. It measures 4 types of exposure to violence: media (television or film exposure), report (hearing other people's reports of occurrence), witness (violence that is directly witnessed), and victim (violence that is directly experienced). The report and witness categories are further divided into 2 groups based on whether the victim of the violent act was a familiar person or a stranger. In our questionnaire, we modified the definitions of the terms familiar person and stranger. We defined familiar person as someone well-known to the child, such as a family member or friend. Stranger was defined as any person not well-known to the child. In the original questionnaire, the term familiar person referred to a wide spectrum of people from family members to distant acquaintances. The modified questionnaire had good internal consistency reliability (overall Cronbach α of .89).
For each type of exposure to violence, the child was asked about his/her experience with 5 violent acts: being beaten up, chased or threatened, robbed or mugged, stabbed or shot, and killed. Thus, there were 5 questions for each type of exposure: media, reported/stranger, reported/familiar person, witness/stranger, witness/familiar person; and 4 questions for the victim scale (which excludes being killed). For example, the 5 questions belonging to the media scale were as follows: “How often have you watched someone on television or in the movies being beaten up? … chased or threatened? … robbed or mugged? … stabbed or shot? … killed?” Questions belonging to the witness/stranger scale were as follows: “How often have you witnessed a stranger or a person not well-known to you being beaten up? … chased or threatened? … robbed or mugged? … stabbed or shot? … killed?” No specific questions were asked to further define or identify the precise nature of the exposure to violence (who did it, when, and how it happened).
Each question in the CREV was answered using a 5-point scale: 0 (no/never), 1 (1 time), 2 (a few times), 3 (many times), to a maximum of 4 (almost everyday). As with the original CREV, a total score was obtained by summing the scores for all questions. The range of possible total scores was from 0 (minimum) to 116 (maximum).
For most of the study, we assessed exposure to violence categorically (whether a child was exposed to a violent event or not). In so doing, we came up with prevalence rates of the different types of exposure in our sample. We determined the percentage of children who had been exposed to each of the 5 violent acts surveyed through the 4 types of exposure. For each type of exposure, a cumulative percentage of children exposed to any of the 5 acts was also calculated. χ2 analysis was used to compare sociodemographic groups in terms of their exposure to violence. In addition, CREV total scores, which would reflect frequency of exposure, were used as continuous variables in the analysis of variance, and multiple regression analyses to assess further the correlates of exposure to violence. P values <.05 were considered significant.
Of 225 mother-child pairs approached, 175 (78%) agreed to participate and completed the survey. Table 1 describes the study population. Slightly more than half (53%) were boys. Mean age of the children was 10.8 years. Among the mothers, 55% were Hispanic and 33% African-American, 29% did not finish high school, 40% received public assistance, 52% were unemployed, and 44% were either married or living with partner.
Exposure to Violence
Table 2 summarizes the exposure to the different types of violence among the 175 children who participated in the study. The last row represents the cumulative percentage of children who had been exposed to any of the 5 violent acts surveyed. All children had been exposed to media violence; 93% had heard reports of violent acts involving a familiar person or stranger; 79% had witnessed 1 or more of the 5 violent acts surveyed; and about half (49%) had been direct victims of violence. All but 5 (170/175) or 97% of the children had also been exposed to more direct forms of violence other than media (being victims, witnesses, or hearing other people's reports). Over one quarter (27%) of the children had witnessed someone being shot or stabbed, and 18% had actually witnessed someone being killed. About 31% of the children said they had witnessed someone being shot, stabbed, or killed. Most of the children who were victims of violence had been beaten up (37 of 49 children).
Table 3 characterizes the relationship of the child-witness to the victim of the violent act. About three-quarters (77%) of the 175 children had witnessed violence involving strangers and almost half (49%) of all subjects had witnessed violence involving familiar people. Rates of exposure to stranger-related violence were about double the rates for familiar persons except for killings where it was triple. Seventeen percent had witnessed a stranger being killed, while 5% had witnessed a familiar person being killed.
Factors Related to Exposure to Violence
Using bivariate analysis, male children were significantly more likely to be victims of and witnesses to violence than females. Sixty-three percent (56/93) of the boys had been victims of violent acts compared with 33% (27/82) of the girls (χ2 = 16.23, df = 1; P<.0001). Eighty-seven percent (81/93) of the boys as compared with 70% (57/82) of the girls had been witnesses to violent acts (χ2 = 8.08, df = 1;P < .004). Using analysis of variance, boys had significantly higher CREV total scores compared with girls (27.70 vs 22.15; F = 8.94, df = 1173;P = .0032). Other variables including age, ethnicity, socioeconomic status, maternal education, employment, and marital status were not significantly associated with CREV total scores. Multiple regression analysis was further used to examine which sociodemographic variables were independently associated with exposure to violence when the other variables were taken in consideration (F = 3.01, adjusted R2= .06; P < .05). The results were similar to the bivariate analysis in that being a male child was the only variable among those that were measured that was significantly associated with CREV total scores (β = .21; P = .006).
The findings from this study parallel findings in school-based studies about the high rates of exposure to violence among urban school-aged children. As expected, all children were exposed to media violence. However, 97% had also been exposed to more direct forms of violence through hearing reports, witnessing events, and being victims themselves. Because the population was derived from users of a primary care clinic, there may be limitations in the generalizability of the study to a larger population. Previous reports have shown that there is a higher rate of psychosocial morbidity among those who visit a primary care clinic.18 However, the consistency of the findings with those found in school and housing-based studies in other urban areas is relevant and noteworthy and suggests a phenomenon that is affecting school-aged children in many urban areas11–13even when taken out of the cluster sample provided in a school setting. More importantly, it supports the idea that pediatricians need not look further out in the community to encounter children who have been exposed to violence. The primary care clinic may be a venue for screening patients for exposure to violence. Recognizing this, the American Academy of Pediatrics' Task Force on Violence through its policy statement has called for a more active role for pediatricians in youth violence prevention and management, specifically in 4 major areas: clinical services, community advocacy, research, and education.19
The data from the study also demonstrate that exposure of urban school-aged children to violence is not a rare event, but that many children are exposed to multiple episodes of violent acts. This exposure to violence may provide a constant threat to the safety of children and may influence child-rearing practices.20Children in highly violent areas, for example, may be taught to watch television lying prone to avoid being hit by stray bullets through the windows.21 The walk home from school may become an arena of fear and distress for parents and children. Mothers in violent communities may refrain from bringing their children to playgrounds denying them the chance to play with other children and explore the environment.20 An interesting finding is that the children reported witnessing violence toward strangers more than toward familiar people. It may be hypothesized that even at this early age of 9 to 12 years, community violence may actually be a reality that many children experience.
The low R2 (.06) in the multiple regression analysis underscores the complexity of the issue regarding the correlates of exposure to violence. Although we are not trying to make a model for violence exposure in children, we have considered some sociodemographic characteristics that we feel may contribute to exposure to violence, including race, socioeconomic status, maternal education and marital status, finding the strongest association with exposure to violence with gender. Undoubtedly, many other unmeasured factors may contribute to violence exposure among children.
A substantial number of families approached (22%) refused to participate in the study. If families exposed to a high degree of violence are more reluctant to let their children participate, then this study may underestimate rates of exposure.
This study measures only 5 violent acts: being beaten up, chased or threatened, robbed or mugged, shot or stabbed, and killed. The questionnaire used does not measure other acts of violence such as rape. The forms of violence assessed may partly account for the larger percentage of boys compared with girls who were victims and witnesses of violence.
This study explored only the experiences of children on the receiving end of violent acts and provided no information about children who were perpetrators of violent acts. The Youth Risk Behavior Surveillance Survey22 reports that almost one fifth of students nationwide carried a weapon to school during the previous 30 days and as many as 37% of students had been in a physical fight during the previous 12 months. In 1997, juvenile offenders were involved in about 12 percent of all murders in the United States.23 Victims and witnesses of violence are more likely to be perpetrators themselves.6
In summary, the findings from this clinic-based study are consistent with the findings from previous school-based studies. In our primary care setting, almost all of the urban school-aged children interviewed had been exposed to types of violence that are more direct than media exposure, with a substantial number exposed to multiple episodes of violence. Male children had significantly greater exposure to violence than females.
Additional studies should explore the extent of the problem among children in other settings and age groups and the impact of exposure to violence on children. Pediatricians should be aware that violence happens not only on television, but may be a problem in many of their patients' lives.
This work was supported in part by the Maternal and Child Health Bureau fellowship training grant in Behavioral Pediatrics (MCJ-9096).
We thank the New York City Health and Hospitals Corporation and the Jacobi Medical Center for their cooperation. Special thanks to Laurie Bauman, PhD, for her invaluable support and comments during the course of the project.
- Received March 17, 2000.
- Accepted June 29, 2000.
Reprint requests to (O.H.P.) Children's Evaluation and Rehabilitation Center, Rose F. Kennedy Center, 1410 Pelham Pkwy S, Bronx, NY 10461. E-mail:
Presented in part at the 38th Annual Meeting of the Ambulatory Pediatric Association; May 5, 1998; New Orleans, LA.
Dr Purugganan is currently with the Children's Evaluation and Rehabilitation Center, Rose F. Kennedy Center, Bronx, New York.
- CREV =
- Children's Report of Exposure to Violence
- ↵Fox JA, Zawitz M. Homicide trends in the United States. Washington, DC: US Department of Justice: Bureau of Justice Statistics. Available at: http://www.ojp.usdoj.gov/bjs/homicide/homtrnd.htm October 17, 1999
- Mann Rinehart P, Borowsky I, Stolz A, et al. Youth Violence. Lessons From the Experts. Minneapolis, MN and San Francisco, CA: Division of General Pediatrics and Adolescent Health, Department of Pediatrics, University of Minnesota and the Division of Adolescent Medicine, Department of Pediatrics and Institute of Health Policy Studies, School of Medicine, University of California, San Francisco; 1998
- Campbell C,
- Schwarz D
- Willis E, Strasburger V. Media violence. The Pediatric Clinics of North America: Violence Among Children and Adolescents. 1998;45:319–332
- American Academy of Pediatrics, Task Force on Violence
- Osofsky J. The effects of exposure to violence on young children. Am Psychol. 1995:782–788
- Centers for Disease Control and Prevention
- Snyder HN, Sickmund M. Juvenile Offenders and Victims: 1999 National Report. Washington, DC: US. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention; 1999
- Copyright © 2000 American Academy of Pediatrics