Review and Special Articles

Youth Violence in the United States

Major Trends, Risk Factors, and Prevention Approaches
Linda L. Dahlberg, PhD

ABSTRACT: Violence among youths is an important public health problem. Between 1985 and 1991,
homicide rates among youths 15–19 years of age increased 154% and remain, today, at
historically high levels. This paper reviews the major trends in homicide victimization and
perpetration among youths over the last decade, the key risk factors associated with
violence, and summarizes the many primary prevention efforts under way to reduce
violence. Previous research points to a number of factors that increase the probability of
violence during adolescence and young adulthood. Some of these factors include the early
onset of aggressive behavior in childhood, social problem-solving skill deficits, exposure to
violence, poor parenting practices and family functioning, negative peer influences, access
to firearms, and neighborhoods characterized by high rates of poverty, transiency, family
disruption, and social isolation. Efforts to address some of the primary risk factors for
violence are under way across the United States, but evaluations to confirm program
effectiveness are needed.
Medical Subject Headings (MeSH): adolescence, violence, aggression, risk factors, preven-
tion (Am J Prev Med 1998;14:259 –272) © 1998 American Journal of Preventive Medicine

Y
outh violence is an important national and pub- among youths over the last decade, (2) to review the
lic health problem. Despite a recent decline in key risk factors associated with antisocial, aggressive,
homicide rates across the United States, homi- and violent behavior, and (3) to describe the types of
cide continues to claim the lives of many young people. prevention approaches being used to reduce aggres-
The human and economic toll of violence on young sion and violence and what is presently known about
people, their families, and society is high. Homicide is the effectiveness of these approaches.
the second leading cause of death for persons 15–24
years of age1 and has been the leading cause of death
for African Americans in this age group for over a Prevalence of Violence Among Youth
decade. The economic cost to society associated with
The United States ranks first in the highly industrial-
violence-related illness, disability, and premature death
ized world in homicide rates, exceeding the rates of
is estimated to be in the billions of dollars each year.2
homicide in other industrialized nations by several-
Many efforts to reduce violence among youths are
fold.3 In 1995, there were 22,895 homicides in the
under way. The current challenge of the field is to
United States corresponding to an annual age-adjusted
identify programs that work. Primary prevention efforts
rate of 9.38/100,000.4 The rate of homicide in the
within families, and in schools, neighborhoods, and
United States exceeds the rate in Northern Ireland,
communities across this country appear to be essential
which has the next highest rate (5.85/100,000) by a
to stemming the tide of violence. As preventive mea-
factor of 1.6, and is 17 times the rates of homicide in
sures are tried, it is vital that they be critically reviewed
England/Wales and Japan, countries with the lowest
and evaluated. Monitoring and documenting proven
rates in the industrialized world (.55 and .57/100,000,
strategies will go a long way toward reducing violence
respectively).3
and creating peaceful, healthier communities.
Young people are disproportionately represented as
The purpose of this paper is threefold: (1) to provide
victims of homicide in the United States. Thirty-eight
an overview of violent victimization and perpetration
percent of all homicides in 1995 occurred among
children and youths under 25 years of age.4 The
Division of Violence Prevention, National Center for Injury Preven-
tion and Control, Centers for Disease Control and Prevention, majority of homicide victims (82%) under 25 years of
Atlanta, Georgia. age were boys and young men. In each year since 1985,
Address correspondence to: Linda L. Dahlberg, PhD, NCIPC, 15- to 19-year-old males have constituted an increasing
Division of Violence Prevention, Mailstop K60, Centers for Disease
Control and Prevention, 4770 Buford Highway NE, Atlanta, Georgia proportion of all homicide victims, rising from 6.5% of
30341. the total in 1986 to 12.1% in 1995.4 Homicide victim-

Am J Prev Med 1998;14(4) 0749-3797/98/$19.00 259

© 1998 American Journal of Preventive Medicine PII S0749-3797(98)00009-9
Figure 1. Homicide rates for young males by age group, United States, 1963–1995.

ization is especially high among African-American and been among teenagers (Figure 2). For two decades
Hispanic youths. In 1995, African-American males and (1965–1985), arrest rates for murder were distributed
females 15–24 years of age had homicide rates (74.4/ across the entire 18 –24 age range. During the latter
100,000) that were more than twice the rate of their half of the 1980s, arrest rates for murder shifted sharply
Hispanic counterparts (36.3/100,000) and nearly 14 for those 18 years of age and younger. Findings from a
times the rate of their white, non-Hispanic counterparts recent report reveal that between 1985 and 1994, the
(5.4/100,000).1 arrest rate for murder increased 150% among teenag-
Historically, young people have had higher rates of ers under the age of 18.6 Between 1989 and 1994, the
homicide than any other age group. However, the arrest rate for murder among teenagers (14 –17 years),
exceptional increase in the rates of homicide and alone, increased 41% compared with an 18% increase
homicide perpetration among teenagers causes special among young adults (18 –24 years).7 During the same
concern. Rates of homicide among youths 15–19 years time period, homicide arrest rates decreased 19% for
of age reached record-high levels in the latter half of adults over 25 years of age.
the 1980s and continue to be among the highest ever Arrest rates for murder have risen sharply for both
recorded in the United States for this age group. From minority and white male teens.7 Arrests among young
1963 through 1985, homicide rates among males 15–19 females have also increased substantially since 1985.
years of age were nearly half the rates for the next three Between 1985 and 1994, violent crime arrests for fe-
higher 5-year age groups (Figure 1).5 Between 1985 males under age 18 more than doubled, rising from
and 1991, however, annual homicide rates among 9,000 violent crime index arrests to more than 21,000.6
males 15–19 years old increased 154% (from 13 to Arrest rates for females increased 125% during this
33/100,000), surpassing the rates of youths in the time period compared with an increase of 67% for
25–29 and 30 –34 year age groups, although rates for males. While the increase in arrests for murder by
the two older age groups also increased during this males under the age of 18 surpassed females (158% vs.
time period (from 24.4 to 32.3 and from 22.1 to 25.7, 64%), female arrest rates increased more than male
respectively).5 Homicide rates for males 15–19 years of rates for robbery (110% vs. 51%), aggravated assault
age peaked again in 1993, rising to 34.7/100,000. Rates (134% vs. 88%), and simple assault (141% vs. 102%).6
for all age groups dropped in 1994 and 1995, but for
youths under 25 years of age, homicide rates remain,
Major Causes of Youth Violence
even today, at historically high levels.
Age-specific arrest rates for murder and non-negli- The first step toward preventing violence, according to
gent manslaughter have also increased among youths. the public health approach, is to identify and under-
The sharpest increase in arrest rates for murder has stand the factors that place young people at risk for

260 American Journal of Preventive Medicine, Volume 14, Number 4

Figure 2. Age-specific arrest rates for murder, United States, 1965–1994. Source: Age-Specific Arrest Rates and Race-Specific Arrest
Rates for Selected Offenses, 1965–1992 Uniform Crime Reporting Program, Federal Bureau of Investigation, Washington, D.C.:
December 1993, 1994 unpublished.

violent victimization and perpetration. Once these fac- adults, to have more criminal convictions as adults, and
tors are identified, researchers and prevention special- to engage in more serious criminal acts. As adults, the
ists can then assess which are the most amenable to children of these subjects also exhibited aggressive
intervention. Previous research indicates that there are behavior.9
a number of factors—some residing in the individual Early aggressive behavior is also associated with de-
and others in the social environment—that increase linquent behavior and serious violent offending (e.g.,
the probability of violence during adolescence and aggravated assault, armed robbery) during adoles-
young adulthood (Table 1).

Individual and Familial Factors Table 1. Key Risk Factors for Aggression, Violence, and
The period of adolescence and young adulthood is a Delinquency
time of heightened risk behavior. Even though violence Individual factors
and criminal behavior reach their peak during this History of early aggression
stage of life, violence has many precursors and rather Beliefs supportive of violence
long developmental pathways.8 –11 Previous research Attributional biases
has identified a number of precursors to violence, Social cognitive deficits
Family factors
including early onset of aggressive behavior,8 –10,12 early Problem parental behavior
development of beliefs supportive of aggression,13 in- Low emotional attachment to parents/caregivers
formation-processing/social problem-solving skill defi- Poor monitoring and supervision of children
cits,14,15 exposure to violence in the home,11,16,17 poor Exposure to violence
parental supervision and monitoring,18,19 and weak Poor family functioning
Peer/school factors
family bonding and social supports.20 Negative peer influences
One of the strongest predictors of later aggression Low commitment to school
and criminal involvement is a history of early aggres- Academic failure
sion.8 –10,12,21 Longitudinal research shows that aggres- Certain school environments/practices
sion is stable over time and across generations.9 In one Environment/neighborhood factors
High concentrations of poor residents
study spanning 22 years, children who exhibited aggres- High levels of transiency
sive behavior at age eight were more likely to score High levels of family disruption
higher on aggression based on the Minnesota Mul- Low community participation
tiphasic Personality Inventory at age 30, to report Diminished economic opportunity
higher levels of physical aggression and spouse abuse as Access to firearms

Am J Prev Med 1998;14(4) 261

cence. Longitudinal research on more than 4,000 ing emotional neglect during childhood may persist
youths in three cities, for example, shows that violent over time. Kruttschnitt et al.,36 for example, found that
offending begins at relatively early ages, and for a small children who experienced emotional neglect and
proportion of youths, persists over time. Data from the lacked emotional support systems inside or outside the
Rochester Youth Study indicate that among those who home were more likely to exhibit abusive behavior as
began committing violent offenses before age 9, nearly adults. Some evidence also suggests that neglected
40% became chronic violent offenders by the age of 16, children may actually show higher levels of subsequent
compared with 30% who began committing violent violent behavior than physically abused children.37,38
offenses between the ages of 10 and 12, and 23% who Neglected children may also direct aggression inward
began at age 13 or older.11 Similar data from the and attempt suicide.38
Denver Youth Study show that of those who initiated Low emotional attachment to parents has also been
violent behavior at age 9 or younger, 62% became reported for violent offending and delinquency.
chronic violent offenders during adolescence, and al- Among youths characterized as chronic violent offend-
most half (48%) of those who initiated violent behavior ers in one study, over 40% report low attachment to
between 10 and 12 years of age eventually became parents compared with 16.4% of nonviolent respon-
chronic violent offenders.11 dents.12 There also appears to be a reciprocal relation-
There are a number of factors that contribute to ship between parental attachment and delinquency.
keeping the pattern of aggressive and delinquent be- Youths who have low attachment to parents have higher
havior consistent over time. Researchers using social- rates of delinquency.39 Also, earlier involvement in
cognitive models have found an empirical link between delinquency tends to further reduce levels of attach-
aggression and hostile attributional biases,22–24 norma- ment to parents39 and attachment to school.40,41
tive beliefs supportive of aggression,13 and social prob- Aggression, delinquency, and other problem behav-
lem-solving skill deficits.14,15,25 Children and adoles- iors have also been reported among children growing
cents with cognitive impairment, for example, often up in homes where a parent has an antisocial person-
have difficulty interpreting social situations, make so- ality,42 or a history of criminal behavior, and drug or
cial decisions quickly, and perceive the actions or alcohol abuse.11,32,43 Drug or alcohol abuse on the part
intentions of others as hostile.14,15,22–25 When con- of parents or caretakers is an indicator of family dys-
fronted with aggressive responses from others, highly function and can impair parenting practices. Other
aggressive youths tend to have difficulty arriving at factors, such as individual temperament, poor impulse
non-aggressive solutions, endorse retaliatory aggres- control, or antisocial personality in the parent, can
sion,15,26 –28 believe that aggression reduces aversive interfere with or complicate parental efforts to exercise
treatment by others,29 and believe that aggression effective control or monitoring of children.44 Parenting
increases one’s self-esteem and image.30 practices such as poor monitoring and supervision of
Social-cognitive factors, however, do not exist in children,12,18,19 poor communication, or deficiencies
isolation from other risk factors. Family risk factors, in solving problems or negotiating with children45 are
particularly those associated with parental behavior and positively associated with antisocial, aggressive, and
the family environment, are also key to understanding delinquent behavior. In effect, the absence of effective
why some youths are at greater risk for violence and bonds and controls over behavior puts children at risk
why patterns of aggression and delinquency can be for later violence.44
stable over time. Familial factors can be categorized as: Effective monitoring and social integration within
(1) those that pertain to emotional attachment or the family is important for prosocial development, but
bonding between parents or caretakers and children, other facets of family life can offset the positive effects.
(2) those that pertain to parental problem behaviors Some of the key family factors believed to be associated
(e.g., crime or drug and alcohol use) or parenting with violence include harsh discipline of children, child
practices (e.g., discipline, monitoring, and supervision maltreatment, spousal/partner violence, and a family
of children), and (3) those that pertain to overall family climate of conflict and hostility. Each of these factors
functioning (e.g., communication, cohesion, and fam- has been examined for its contribution to the develop-
ily conflict or violence). ment of aggressive responses in children, to delin-
Research on emotional attachment and bonding quency, to violent offenses during adulthood, and to
shows a greater risk for aggressive behavior and antiso- the cycle of violence or intergenerational transmission
cial behavior in children who experience rejection, of violence. The evidence shows that children who have
neglect, or indifference from parents.18,31–35 Neglectful been physically abused at home are more likely to
or disengaged parents are nonresponsive to the needs exhibit aggression in daycare and school settings,46 – 48
of children and tend to demand little of them. Not only to engage in delinquent and serious violent behavior as
have positive associations been reported between ne- adolescents,11,39,49 and to be arrested for a violent
glect and aggression, but also, the effect of experienc- crime as an adult.37 Empirical evidence to support

262 American Journal of Preventive Medicine, Volume 14, Number 4

intergenerational transmission of violence, however, is of dress, and support system are important facets of
limited by numerous and serious methodologic flaws.38 adolescent thought and behavior.
Several studies have documented a relationship be- The functions of peer groups are generally positive
tween witnessing violence in the home (e.g., violence and they are regarded by most developmental experts
between parents) and increased risk of being both a as instrumental in shaping interpersonal development,
victim and a perpetrator of violence.11,16,17,50,51 Expo- and emotional and social competence. Peer influences,
sure to multiple forms of family violence (e.g., child however, can also be negative, especially if peer de-
maltreatment, spousal/partner violence, a family cli- mands for conformity include strong social pressures
mate of physical fighting and hostility) has a particu- for engaging in risk behaviors. Youths are much more
larly strong effect. Findings from one longitudinal likely to engage in negative activities when those behav-
study show that while 38% of youths from nonviolent iors are encouraged and approved by their friends.61 It
families report involvement in violent behavior, the is the potentially negative influences—the behavioral
rate increases to 60% for youths exposed to one form of patterns that are established, and the adverse conse-
family violence, 73% for youths exposed to two forms of quences that follow from engaging in risk behaviors—
family violence, and 78% for youths exposed to three that receive the most attention from parents, school
forms of family violence.11,49 Other research indicates administrators, researchers, health officials, and policy
that exposure to violence during childhood as a witness makers.
or victim significantly increases the likelihood of adult Negative peer influences are an important risk factor
violence toward spouses or children (see Fagan and for a variety of health consequences, including teenage
Browne, 1994, for a summary)52 and may increase the pregnancy, tobacco use, substance abuse, and delin-
risk of violent behavior during adolescence by as much quency. Research on delinquency, for example, shows
as 40%.44 that associating with delinquent peers increases the risk
Apart from the potential effect on later victimization of serious delinquency and involvement in criminal
and perpetration, exposure to violence may have im- activity severalfold.11 Others have also found that delin-
portant effects on mental health. Post-traumatic stress quency is a group phenomenon.62– 64 Within delin-
and psychological trauma associated with witnessing quent peer groups, negative behavioral patterns are
violence have been reported in a number of studies and modeled, encouraged, and rewarded. The adolescent
reports.53–57 Findings from a recent study of high “status” hierarchy that emerges in these groups can also
school students revealed that exposure to physical lead some adolescents to commit delinquent acts in
violence had a positive and significant independent pursuit of a higher position in the status hierarchy.
association with depression, anger, anxiety, dissocia- The relationship between violent behavior and peer
tion, post-traumatic stress, and total trauma symp- influences is less clear. Delinquency research does not
toms.57 Anxiety, stress, and depression have also been always separate violent acts (e.g., aggravated assault,
linked to school failure and suicidal ideation.58 – 60 homicide) from nonviolent acts (e.g., theft, property
Aggressive behavior patterns and poor problem-solv- crime), making it more difficult to discern the effect of
ing skills learned at home can interfere with develop- group processes on violent events. Membership in
ment of positive peer relationships and academic gangs has been cited as a risk factor for violent victim-
achievement. Growing up in homes without these prob- ization and perpetration;64 – 66 however, research on
lems, however, is no guarantee of nonviolent social gangs shows a great deal of variability in their social
adjustment in other contexts. Peer and school contexts, organization and cohesion.67 While some gangs main-
for example, exert their own influences on social tain high rates of violence and involvement in delin-
development and can create situations that elicit vio- quent activities, others do not. Moreover, research on
lent responses. gang members and nongang members indicates that
similar processes produce frequent and persistent vio-
lent behavior among both groups.20
Peer and School Influences
The effects of negative peer groups can spill into the
Adolescence is a period in which young people experi- wider school setting. Findings from a national school
ence many physical, psychological, and social demands. victimization survey reveal that students who report
During this stage of life, young people begin to distance gang presence in school also report an increased
themselves from parental authority, try on new identi- presence of weapons in school, compared with students
ties with the hope of establishing an individual identity who report no gang presence in school.68 Students
and self-worth, adapt to society’s demands for behav- reporting weapons and gangs in school are also more
ioral maturity, and prepare for adult roles.61 Peers are likely to report being victims of physical attacks, rob-
key to the developmental process. The need to “get bery, and bullying and to report higher rates of witness-
along with” peers; to be accepted by the group; to not ing these events in school.68,69 These students are also
be seen as different; and to have a group language, style more fearful of physical attacks while attending school.

Am J Prev Med 1998;14(4) 263

It is important to recognize that the presence of gangs 20.3% vs. 12.9%, respectively). Eight percent of stu-
and weapons in these schools may be symptomatic of dents reported being threatened or injured with a
other systemic problems surrounding schools (e.g., weapon on school property during the 12 months
neighborhood violence, social disorganization, pover- preceding the survey.
ty). According to the National Research Council, rates
of violence in school are highest in districts marked by
Environmental Influences
higher neighborhood crime rates.20
Certain school practices have been implicated in Just as peers and schools exert their own influences on
fostering negative peer group interactions and the behavior, so too do the environments children inhabit
problems most likely to result from these interactions. and confront on a daily basis. In 1995, 36.4 million
Ability tracking, for example, tends to place academi- people, representing 13.8% of the nation’s population,
cally poor students and those with disruptive behavioral were living below the poverty level.75 Children under
patterns together in classes. Ability grouping has not the age of 18 made up 40% of the nation’s poor and
been shown to improve learning among low-achieving had a rate of poverty (20.8) well above the national rate
students, and indeed, has been associated with many (13.8). African Americans and Hispanics were also
negative social consequences.70 –72 Placement of youths disproportionately represented among the nation’s
in ability groups reinforces feelings of anger, rejection, poor. In 1995, the rates of poverty for African Ameri-
and alienation, and can lead to academic failure.71,72 cans and Hispanics (29.3 and 30.3, respectively) were
Although direct empirical links between ability group- more than twice the national rate and more than three
ing and violence have not been demonstrated, research times the poverty rate for white, non-Hispanic persons
does show that youths who experience academic failure (8.5).
and exhibit a low commitment to school are at in- The relationship between poverty and violence is
creased risk of engaging in violent behavior.12 complex. Although research shows that homicide is
School settings in general, however, may contribute higher in urban areas where poverty is most prevalent,
to disruptive behavior and violence. The National Re- crime and violence have more to do with the dimen-
search Council has cited a number of characteristics of sions of poverty than poverty status per se. There are
the school environment that engender more aggres- numerous dimensions of poverty that relate to high
sion and violence, including undisciplined classrooms, rates of community violence. Crime and violence are
lax enforcement of school rules and policies, tight high in poverty areas, where at least 20% of the
physical space, and conformity to behavioral routines residents are poor.76 These areas are characterized by
that seem to produce feelings of anger, resentment, high concentrations of poor people, high levels of
and rejection in some students.20 It is not clear, how- transiency, family disruption, crowded housing, low
ever, whether these factors lead to violence or whether community participation and organization, and the
high rates of violence lead to these conditions in presence of firearms and drug distribution net-
schools. works.20,76,77 These areas also tend to have high rates of
Much of the violence that does occur in schools, school dropouts; high rates of substance abuse, unem-
while serious, is not fatal. Findings from a recent ployment, and teenage pregnancy; and a disproportion-
national study reveal that school-associated violent ate number of households headed by women.20,76,78
deaths are rare events.73 During the 2-year study period In addition to their demographic characteristics,
(1992–1994), 105 school-associated violent deaths oc- economically poor neighborhoods differ from affluent
curred nationally: 81% were homicides and 19% were neighborhoods in a number of ways. These neighbor-
suicides. Over 40% of the deaths occurred during a hoods have diminished private economic activity. The
school activity (e.g., during regular class sessions), types of public and social services that are available to
while an equal proportion (43.8%) occurred before or residents are limited as are recreation and developmen-
after official school activities. The overall risk of a tal programs for youths.79 Poor neighborhoods also
school-associated violent event was .09/100,000 student tend to be characterized by disorganization or a lack of
years. neighborhood cohesion. Disorganized neighborhoods
School violence, more commonly, takes the form of lack effective social controls. Factors such as high levels
physical fighting. In 1995, 15.5% of students from a of transiency make it difficult for individuals to estab-
nationally representative sample of students (grades lish common values and norms and to develop informal
9 –12) reported being in a physical fight on school support networks.80 As a result, people living in such
property during the 12 months preceding the survey.74 neighborhoods often experience a sense of social iso-
Physical fighting on school property was significantly lation and exhibit lower levels of attachment to the
higher among male students than female students community.42,77,81,82 Chronic unemployment further
(21.0% vs. 9.6%) and among Hispanic and African- isolates people by distancing them from legitimate
American students than white students (21.1% and labor markets and increases the likelihood of illegiti-

264 American Journal of Preventive Medicine, Volume 14, Number 4

Figure 3. Firearm, nonfirearm, and total homicide rates ages 15–34 years, United States, 1965–1995.

mate enterprises.80 When neighborhood social and violence and the experience of stressful life events
economic systems break down, the poorest of neighbor- contributes significantly to aggression in children.89
hoods, in effect, become unable to resist crime and Over the long term, children growing up in these
violence.20,77 communities are at greater risk of abandoning educa-
Poor neighborhoods also have a high degree of tional goals, of becoming teenage parents, and are more
family disruption, as evidenced by the number of likely to adopt lifestyles and behaviors that put them at
households headed by women. Families in impover- risk for violent victimization and perpetration.79
ished neighborhoods are nearly twice as likely as those
elsewhere to have a female head-of-household.83 A
Role of Firearms
positive association between family disruption and high
rates of violence has been reported.84 – 87 Family disrup- Violence is too complex a problem to single out any
tion also seems to predict crime and violence indepen- one factor, but the role of firearms is too important to
dent of other social and demographic characteristics.88 be ignored. Firearms are one of the major risk factors
Studies conducted by Sampson,84,85 for example, re- for fatal violent victimization and perpetration. The
vealed that rates of violent victimization were two to age-adjusted homicide mortality attributable to fire-
three times higher among residents of neighborhoods arms increased from 4.97/100,000 in 1985 to 6.71 in
with high levels of family disruption compared with 1995, when firearm homicides accounted for 69% of all
those in areas with low levels, independent of predictive homicides.4 Over 90% of the increase in the number of
factors such as percent African-American and poverty. homicides since the mid-1980s is associated with fire-
Further, the percentage of families headed by women arms, and the recent increases in youth homicide are
explained a large part of the relationship between related, in large part, to increases in homicides involv-
percent African-American and crime. ing firearms.5 Nonfirearm homicides have remained
The net effect of the many social and demographic stable for several years (Figure 3).
characteristics of poor neighborhoods is diminished Firearms have been shown to be far more lethal than
opportunity for the youngest residents. Young people other weapons in aggravated assaults,90 from 2 to 5
growing up in impoverished conditions have few posi- times more lethal than knife attacks.91,92 Furthermore,
tive role models to offset the negative influences in the as Cook90 estimates, between 3 and 5.7 nonfatal gun-
environment. It is also difficult for families, most often shot injuries occur for every firearm homicide. The
single mothers, to reduce children’s level of exposure health-care costs and years of potential life lost from
to the unhealthy lifestyles that characterize these com- firearm injuries and fatalities exceed any other type of
munities or to reduce the level of exposure to violence fatal injury and any of the four leading causes of
in these communities. In the short term, exposure to death.93

Am J Prev Med 1998;14(4) 265

 Young people are reportedly carrying firearms and much more information is needed to fully understand the
other weapons at high rates. In 1990, 4.1% or 1 in 24 relationship between weapon carrying and violence.
students in a nationally representative sample of stu-
dents (grades 9 –12) reported carrying a firearm for
Role of Alcohol and Drugs
fighting or self-defense at least once during the 30 days
preceding the survey.94 In 1995, 7.6% or nearly 1 in 12 Alcohol and illicit drug use have been identified as risk
students reported carrying a firearm for fighting or factors for violence. Case series in the United States and
self-defense.74 The rate among male high school stu- in other nations show that large proportions (22%–
dents was 12.3%, with higher rates among minority 60%, mean 45%) of homicide victims have positive
males (African-American males, 18.7%; Hispanic blood alcohol levels at the time of death. Alcohol
males, 16.8%).74 In another survey, conducted in 10 involvement has been reported among offenders as
selected inner-city high schools in four states, 35% of well. Many violence victims also have evidence of other
male students and 11% of female students reported drugs on toxicology tests, including cocaine (13%–
carrying a gun.95 Findings from a recent study by the 33%), barbiturates (8%), heroin (3%–5%), and other
National Institute of Justice, involving 4,000 juvenile illicit drugs.102 Large surveys of high-school students
arrestees in 11 cities, found that 40% of juvenile males have also reported associations between illicit drug use
reported possessing a firearm at some time; over 33% and involvement in violence.103–106
admitted owning a firearm in the previous month; 22% The contribution of alcohol and substance use to the
reported carrying a gun all or most of the time; 55% current youth violence problem, however, is less clear.
reported being threatened with a gun; 50% reported Demographic patterns of violent behavior in youths, for
having a gun fired at them; 11% had been injured by a example, do not match those for substance use. Some
gunshot; and 38% believed that it was okay to shoot of the patterns presented in a recent review107 are
someone who hurts you.6 worth noting here. Aggressive or violent behavior be-
gins at an earlier age than drug use and also tends to
Arrest rates for weapons possession among youths
decline before drug use does.103,104,108 Males are more
under 18 years of age are also at an historic high.6
likely to use substances than females, but the difference
Between 1985 and 1994, arrests for weapons-law viola-
in patterns of substance use is relatively small, com-
tions among youths (,18 years) increased 103%, com-
pared with the difference between males and females in
pared with an increase of 26% among adults. Most of
rates of violent victimization and perpetration.4,103,104
the increase in weapons-law violations among youths
In a similar vein, African Americans have much higher
has been in the violent crime categories.6 Arrest counts
rates of violence than their white counterparts, but
for weapons-law violations, however, are likely to under-
much lower rates of substance use.4,62
estimate the extent of weapon involvement among To understand the relationship among alcohol,
youths because Federal Bureau of Investigation coding drugs, and violence, some have proposed separating
procedures require an arrest to be classified by the most out the psychopharmacologic effects (i.e., influence of
serious charge involved. Firearms associated with homi- the substance on behavior), economic effects (i.e.,
cides, for example, would not be included in arrest violence committed to obtain money to purchase
counts for weapons-law violations. drugs), and the systemic effects (i.e., violence that
Young people also report having easy access to guns. arises from the sale and distribution of drugs).109 A
A 1993 Harris/Joyce Foundation survey found that 35% thorough review of the research on each of these
of students polled claimed that it would take them “less factors is available elsewhere.107 The major findings
than one hour” to obtain a firearm, and nearly 60% from this review, however, show a moderately strong
responded that firearms are for them “easily obtain- relationship between violence and substance use, but
able.”96 In a study of inner-city students, 23% reported the relationship has more to do with shared influences
that guns were easily accessible in the neighborhoods in on a variety of deviant behaviors than with any causal
which they lived.95 A study of half of the public high relationship between the two. The few longitudinal
schools in Seattle, Washington, also revealed high rates studies that have examined the relationship between
of reported access to firearms. Over one-third of the alcohol, drugs, and violence have failed to find a clear
students (34%) reported easy access to handguns and causal link between them.110 –114 For example, White et
6.4% reported owning a handgun.97 al.’s longitudinal study of males 12 to 18 years of age
Rates of firearm homicide, weapons possession, and found early aggressive behavior to be predictive of later
reported firearm carriage suggest that young people alcohol-related aggression, and early aggression to pre-
have easier access to firearms than in the recent past, or dict increases in alcohol use, but alcohol use did not
greater willingness to use them, or both. Although a lead to increases in aggressive behavior.114
number of studies point to self-protection as the pri- Seeking money to purchase drugs also does not
mary motivating factor for carrying firearms,98 –101 appear to be a major cause of adolescent violence, in

266 American Journal of Preventive Medicine, Volume 14, Number 4

part, because the use of addictive drugs is much lower in schools and communities across the United States
in adolescent populations than in adult populations.107 than ever before.
Evidence to support systemic effects is mixed. A widely
held belief among some researchers115 is that recent
Approaches to Prevent Violence
increases in youth violence are related to the drug
economy, particularly the crack cocaine market that The majority of intervention programs taking place in
emerged in the mid-1980s in many large American this country focus on changing individual attitudes,
cities. Citing a doubling of the homicide rate among beliefs, and behavior. Cognitive-behavior, behavior
youths, doubling of the number of homicides commit- modification, and social-skills training are some of the
ted with firearms, and doubling of the arrest rate of more common approaches being used to reduce anti-
nonwhites for drug offenses, Blumstein115 argues that social and aggressive behavior in children and violent
youths, particularly, minority youths, were recruited victimization and perpetration among youths.116 Cog-
into the illicit drug industry. Youths were desired nitive-behavioral interventions train participants in an-
because they work more cheaply than adults, face less ger management, cognitive self-control, moral reason-
severe sanctions if arrested, and are more willing to ing, social perspective-taking, and attitude change.
take risks. Similar to other participants in the drug Social-skills training programs are similar to cognitive-
trade, youths armed themselves with firearms as a behavioral programs, but place a greater emphasis on
means for resolving disputes as well as for self-protec- behavioral skill development and include techniques
tion. One of the consequences, according to Blumstein, such as modeling, rehearsal, and discussion.116 Behav-
was a “diffusion of guns” in both the community and ioral modification approaches focus more on direct rein-
among youths in schools. The introduction of firearms, forcement, modeling, and contingency contracting.
even among those not directly tied to the drug trade, Other individual approaches include rites-of-passage
resulted in more violent outcomes, as evidenced by or manhood development programs. These programs
homicide rates and firearm homicide rates among focus on self-esteem and emphasize cultural history,
youths during the latter half of the 1980s. pride, moral guidance, and personal responsibility.
Evidence to clearly support the link between the Manhood development programs are sometimes
drug trade and violence, however, is lacking. Not only is paired with mentoring programs to provide youths with
it difficult to gain access to the populations involved in additional guidance and support.117 Psychotherapy and
the illicit drug trade, but the relationship, itself, is intensive casework are also used to alter individual
difficult to study. According to the National Research behavior, particularly with youths exhibiting more mal-
Council, violence associated with the drug trade ap- adaptive behavioral patterns. Together, individual-level
pears to stem from the recruitment of individuals with approaches try to change information-processing skills,
more violent histories and with the location of the drug social cognitions, and perceptions of risk, and provide
trade in communities already experiencing high rates young people with information on how to manage
of violence.20 The implication is that a certain propor- feelings and situations and the tools to resolve potential
tion of adolescent violence would exist even without the conflicts in a nonviolent manner.
drug trade. Despite the compelling connection be- Programs that focus on changing the nature of
tween the drug trade and violence, much more infor- peer-group interactions, peer-group norms, and family
mation is needed to sort out the specific contribution functioning are less common than individual ap-
of the illicit drug industry to youth violence. proaches, although programs such as peer mediation
have increased in frequency in schools across the
United States. Some of the peer interventions that have
Prevention of Youth Violence been tried with youths focus on changing peer-group
norms. The goal of these programs is to change nega-
Identifying the factors that increase or reduce the risk tive peer influences into positive or prosocial peer
for violent victimization and perpetration is an impor- influences (i.e., create a positive peer culture). Pro-
tant first step toward preventing youth violence. The grams that attempt to reduce or prevent associations
interplay of individual, peer, familial, and environmen- with antisocial peers have also been tried. Other peer
tal factors, however, speak to the complexity of the interventions, such as peer mediation, utilize peers or
problem of violence in the United States and to the age mates to resolve conflict disputes.
many challenges faced by prevention specialists. In Family interventions focus on a number of the factors
many respects, it is far easier to identify the factors that related to aggression and antisocial behavior in chil-
place young people at risk for violent victimization than dren. Family intervention approaches provide educa-
it is to design interventions and programs to reduce this tion and training to parents with the goal of teaching
risk. Researchers and intervention specialists have risen participants how to effectively discipline, monitor, and
to this challenge and now more programs are in place supervise children. Some of these programs also try to

Am J Prev Med 1998;14(4) 267

improve the family environment by focusing on family or no effects. Unfortunately, a sizable gap exists be-
relationships, communication, and problem-solving. tween the most commonly used approaches and the
Acknowledging that external factors and demands in- most frequently evaluated approaches. Data to support
fluence family functioning, some approaches also em- approaches such as social-skills training, mentoring,
phasize social support systems and try to provide fami- rites-of-passage programs, peer mediation, conflict res-
lies with skills and resources to address external olution, teacher management practices, and school
demands. security are clearly needed. Evaluation efforts, however,
In addition to individual, peer, and family interven- are under way to help determine if several of these
tion approaches, efforts are under way to change approaches are effective in reducing aggression and
attitudes, beliefs, and behavior by altering the settings violence121 and some are indicating moderate positive
(e.g., schools and neighborhoods) that young people effects.122 Still needed are evaluation data on programs
inhabit on a daily basis. Interventions aimed at chang- addressing larger neighborhood, community, or socio-
ing school settings focus on teacher management prac- economic factors.
tices, school administrative policies and rules, and
school security (e.g., metal detectors, surveillance cam-
eras, campus police). Programs such as antiviolence
awareness campaigns, safe-haven or community youth Conclusion
programs, and neighborhood helpers have been used The United States has witnessed unprecedented rates
to foster prosocial attitudes and behavior in neighbor- of violent victimization and perpetration among the
hoods. In several communities across the country, nation’s youth over the last decade. This surge in youth
empowerment zones have recently been established to crime and violence occurred at a time when the popu-
deal with a multitude of community problems, includ- lation of youths was on the decline.7 An estimated 39
ing crime and violence. Efforts to change the physical million children under 10 years of age now make up
environments of communities have also been tried the “echo-boom.”83 Some researchers have predicted
(e.g., increased lighting, neighborhood police patrols that as these children move into adolescence, the peak
and substations).
years for crime and violence, this country will see yet
another surge in crime and violence.7 This predicted
Determining Effectiveness surge is based not simply on the numbers of children
that make up this cohort, but rather on the presence of
The fundamental question from the perspective of many key risk factors. Nearly one-quarter of the echo-
public health and policy is: “Do these programs work?
boom children under 6 years of age, alone, are living in
Do any of the individual, peer, family, or setting ap-
poverty.75 Over half of the echo-boom children are
proaches effectively reduce aggressive and violent be-
living without the benefit of full-time parental supervi-
havior?” At present, the answer to these questions is not
sion.83 Many are attending schools in neighborhoods
known. Despite the proliferation of programs in
plagued with crime and violence, and many are at risk
schools and communities nationwide, the most effec-
for a number of negative health consequences.
tive strategies to prevent youth violence have yet to be
Intervening with youths at risk for violence should be
determined. For example, although 39% of states and
61% of districts require conflict resolution training in a national priority. Primary prevention efforts are
schools,118 few of these programs have been evaluated clearly needed to disrupt the developmental pathways
for their effectiveness in reducing violent behavior. that lead to violent behavior in adolescence and young
Reviews on the state of the field have recently con- adulthood. The available evidence suggests that early
cluded that promising violence prevention programs intervention, before violence becomes part of a young
exist, but evaluations to confirm strong, positive effects person’s behavioral repertoire, is key to fostering proso-
are lacking.116,119,120 cial behavior and prosocial development. Given that
Results from interventions that have been evaluated violence is the result of a complex interplay of individ-
are mixed. Findings from one of the comprehensive ual, peer, familial, and environmental factors, primary
reviews of the field116 indicate positive effects for cog- prevention efforts are necessary on several levels, from
nitive-behavioral interventions, family interventions, the individual, to the family, to the neighborhoods and
school climate programs that include parental involve- communities where young people live. Finally, it is
ment, and interventions that reduce exposure to media important to measure progress at each level and to use
violence. Psychotherapy, intensive casework, peer inter- science to guide efforts. Determining “what works” will
ventions that attempt to redirect negative peer activities go a long way toward reducing violence among our
(e.g., Guided Group Interaction), and behavioral mod- youth and will move us forward in creating peaceful,
ification programs have demonstrated weak, negative, healthier communities.

268 American Journal of Preventive Medicine, Volume 14, Number 4

References 17. Finkelhor D, Gelles RJ, Hotaling GT, Straus MA, eds.
1. Anderson RN, Kochanek KD, Murphy SL. Report of The dark side of families: current family violence re-
final mortality statistics, 1995. Monthly Vital Statistics search. Beverly Hills, California: Sage Publications;
Report 45. 1997;11(suppl2). 1983.
2. Miller, TR, Cohen, MA, Rossman, SB. Victim costs of 18. Loeber R, Stouthamer-Loeber M. Family factors as cor-
violent crime and resulting injuries. Health Affairs 1993; relates and predictors of juvenile conduct problems and
12:186 –97. delinquency. In: Tonry M, Morris N, eds. Crime and
3. Krug EG, Powell KE, Dahlberg LL. Firearm-related Justice: an annual review of research, vol. 7. Chicago,
deaths in the United States and 35 other high- and Illinois: University of Chicago Press; 1986:29 –150.
upper-middle income countries. Int J Epidemiology 19. Patterson GR, Southamer-Loeber M. The correlation of
1998;27:214 –221. family management practices and delinquency. Child
4. Centers for Disease Control and Prevention. National Dev 1984;55:1299 –1307.
Summary of Injury Mortality Data, 1985–1995. Atlanta, 20. Reiss AJ, Roth JA, eds. Understanding and preventing
GA: National Center for Injury Prevention and Control; violence: panel on the understanding and control of
1997. violent behavior, vol. 1. Washington, DC: National Acad-
5. Centers for Disease Control and Prevention. Homicide emy Press; 1993.
among 15–19-year-old males: United States, 1963–1991. 21. Parke RD, Slaby RG. The development of aggression. In:
MMWR 1994;43(40):725–27. Mussen PH, Hetherington EM, eds. Handbook of child
psychology, vol. 4: socialization, personality, and social
6. Snyder HN, Sickmund M, Poe-Yamagata E. Juvenile
development. New York: John Wiley & Sons; 1983:547–
offenders and victims: 1996 update on violence. Wash-
641.
ington, DC: US Department of Justice, Office of Juvenile
22. Hudley, CA. Perceptions of intentionality, feelings of
Justice and Delinquency Prevention; 1996;1–32.
anger, and reactive aggression. In: Furlong MJ, Smith
7. Fox JA. Trends in juvenile violence: a report to the
DC, eds. Anger, hostility, and aggression: assessment,
United States Attorney General on current and future
prevention, and intervention strategies for youth. Bran-
rates of juvenile offending. Washington, DC: US Depart-
don, Vermont: Clinical Psychology Publishing Com-
ment of Justice, Bureau of Justice Statistics; 1996:1–15.
pany; 1994:39 –56.
8. Farrington DP. Longitudinal research on crime and
23. Hudley C, Graham S. An attributional intervention to
delinquency. In: Morris N, Tonry M, eds. Crime and
reduce peer directed aggression among African-Ameri-
justice: an annual review of research, vol. 1. Chicago,
can boys. Child Dev 1993;64:124 –38.
Illinois: University of Chicago Press; 1979:289 –348.
24. Nasby W, Hayden B, De Paulo BM. Attributional bias
9. Huesmann LR, Eron LD, Lefkowitz MM, Walder LO. among aggressive boys to interpret unambiguous social
Stability of aggression over time. Dev Psychol 1984;20: stimuli as displays of hostility. J. Abnorm Psychol 1980:
1120 –34. 89:459 – 68.
10. Olweus D. Stability of aggressive reaction patterns in 25. Dodge KA. Social cognition and children’s aggressive
males: a review. Psychol Bull 1979;86:852–75. behavior. Child Dev 1980;51:162–70.
11. Thornberry TP, Huizinga D, Loeber R. The prevention 26. Guerra NG. Consequential thinking and self-reported
of serious delinquency and violence: implications from delinquency in high school youth. Criminal Justice
the program of research on the causes and correlates of Behav 1980;16:440 –54.
delinquency. In: sourcebook on juvenile offenders. 27. Guerra NG, Slaby RG. Evaluative factors in social prob-
Washington, DC: US Department of Justice; 1995:213– lem-solving by aggressive boys. J Abnorm Child Psychol
37. 1989;17:277– 89.
12. Farrington DP. Childhood aggression and adult vio- 28. Richard BA, Dodge KA. Social maladjustment and prob-
lence: early precursors and later-life outcomes. In: Pep- lem-solving in school-aged children. J Consult Clin
ler DJ, Rubin KH, eds. The development and treatment Psychol 1982;50:226 –33.
of childhood aggression. Hillsdale, New Jersey: Law- 29. Perry DG, Perry LC, Rasmussen P. Cognitive social
rence Erlbaum; 1991:5–29. learning mediators of aggression. Child Dev 1986;57:
13. Guerra NG, Huesmann LR, Hanish L. The role of 700 –11.
normative beliefs in children’s social behavior. In: Eisen- 30. Slaby RG, Guerra NG. Cognitive mediators of aggression
berg N, ed. Review of personality and social psychology, in adolescent offenders: I. Assessment. Dev Psychol
vol. 15: social development. Thousand Oaks, California: 1988;24:580 – 8.
Sage Publications; 1995:140 –58. 31. Bousha DM, Twentyman CT. Mother-child interactional
14. Dodge KA, Frame CL. Social cognitive biases and defi- style in abuse, neglect, and control groups: naturalistic
cits in aggressive boys. Child Dev 1982;53:620 –35. observations in the home. J Abnorm Psychol 1984;93:
15. Dodge KA, Newman JP. Biased decision-making pro- 106 –14.
cesses in aggressive boys. J Abnorm Psychol 1981:90: 32. Farrington DP. The family backgrounds of aggressive
375–9. youths. In: Hersov LA, Berger M, Shaffer D, eds. Aggres-
16. Jaffe P, Wolfe DA, Wilson SK, Zak L. Similarities in sion and antisocial behavior in childhood and adoles-
behavioral and social maladjustment among child vic- cence. Oxford, England: Pergamon Press; 1978:73–93.
tims and witnesses to family violence. Am J Orthopsychi- 33. Henggeler SW, Melton GB, Smith LA. Family preserva-
atry 1986;56:142– 6. tion using multisystemic therapy: an effective alternative

Am J Prev Med 1998;14(4) 269

 to incarcerating serious juvenile offenders. J Consult intimates: physical aggression between women and men
Clin Psychol 1992;60:953– 61. in intimate relationships. In: Reiss AJ, Roth JA, eds.
34. McCord J. A forty year perspective on the effects of child Understanding and preventing violence: panel on the
abuse and neglect. Child Abuse Neglect 1983;7:265–70. understanding and control of violent behavior, vol. 3:
35. Tolan PH. Socioeconomic, family, and social stress social influences. Washington, DC: National Academy
correlates of adolescent antisocial and delinquent be- Press; 1994:114 –292.
havior. J Abnorm Child Psychol 1988;16:317–31. 53. Fitzpatrick KM, Boldizar JP. The prevalence and conse-
36. Kruttschnitt CL, Ward D, Scheble MA. Abuse-resistant quences of exposure to violence among African-Ameri-
youth: some factors that may inhibit violent criminal can youth. J Am Acad Child Adol Psychiatry 1993;32:
behavior. Social Forces 1987;66:501–19. 424 –30.
37. Widom CS. Child abuse, neglect, and violent criminal 54. Pynoos RS, Eth S. Children traumatized by witnessing
behavior. Criminology 1989;27:251–72. acts of personal violence: homicide, rape, or suicidal
38. Widom CS. Does violence beget violence? A critical behavior. In: Eth S, Pynoos RS, eds. post-traumatic stress
examination of the literature. Psychol bull 1989;106:3– disorder in children. Washington, DC: American Psychi-
28. atric Press; 1985:19 – 43.
39. Huizinga D, Loeber R, Thornberry TP. Urban delin- 55. Taylor L, Zuckerman B, Harik V, Groves BM. Exposure
quency and substance abuse: initial findings. Washing- to violence among inner city parents and children. Am J
ton, DC: US Department of Justice, Office of Juvenile Dis Child 1992;146:487.
Justice and Delinquency Prevention; 1994;1–27. 56. Berton MW, Stabb SD. Exposure to violence and post-
40. Thornberry, TP, Lizotte AJ, Krohn MD, Farnworth M, traumatic stress disorder in urban adolescents. Adoles-
Jang SJ. Delinquent peers, beliefs, and delinquent be- cence 1996;31:489 –98.
havior: a longitudinal test of interactional theory. Crim- 57. Singer MI, Menden Anglin T, Song LY, Lunghofer L.
inology 1994;32:601–37. Adolescents’ exposure to violence and associated symp-
41. Thornberry TP, Lizotte AJ, Krohn MD, Farnworth M, toms of psychological trauma. JAMA 1995;273:477– 82.
Jang SJ. Testing interactional theory: an examination of 58. Berman Al, Jobes DA. Adolescent suicide assessment
reciprocal causal relationships among family, school, and intervention. Presented at the annual meeting of
and delinquency. J Criminal Law Criminol 1991;82:3– the American Psychological Association, Washington,
35. DC; 1991.
42. American Psychological Association Commission on Vi- 59. Dryfoos JG. Adolescents at risk: prevalence and preven-
olence. Violence and youth: psychology’s response, vol. tion. New York: Oxford University Press; 1990.
1. Washington, DC: American Psychological Association; 60. Pynoos RS, Nader K. Psychological first-aid and treat-
1993. ment approach to children exposed to community vio-
43. McCord J. A longitudinal view of the relationship be- lence: research implications. J Trauma Stress 1988;1:
tween paternal absence and crime. In: Gunn J, Far- 445–73.
rington DP, eds. Abnormal offenders, delinquency and 61. Ingersoll GM. Adolescents in school and society. Lexing-
the criminal justice system. New York: John Wiley & ton, Massachusetts: DC Heath & Co; 1982.
Sons; 1982:113–28. 62. Elliott DS, Huizinga D, Ageton SS. Explaining delin-
44. Elliott DS. Serious violent offenders: onset, developmen- quency and drug use. Beverly Hills, California: Sage
tal course, and termination: The American Society of Publications; 1985.
Criminology 1993 Presidential Address. Criminology 63. Reiss AJ. Co-offender influences on criminal careers. In:
1994;32:1–21. Blumstein A, Cohen J, Roth J, Visher C, eds. Criminal
45. Patterson GR, Chamberlain P, Reid JB. A comparative careers and “career criminals.” Washington, DC: Na-
evaluation of a parent-training program. Behav Ther tional Academy Press; 1986:121– 60.
1982;13:638 –50. 64. Thornberry TP, Krohn MD, Lizotte AJ, Chard-Wier-
46. George C, Main M. Social interactions of young abused schem D. The role of juvenile gangs in facilitating
children: approach, avoidance, and aggression. Child delinquent behavior. J Res Crime Delinquency 1993;30:
Dev 1979;50:306 –18. 55– 87.
47. Kinard EM. Emotional development in physically 65. Greydanus DE, Farrell EG, Sladkin K, Rypma CB. The
abused children. Am J Orthopsychiatry 1980;50:686 –96. gang phenomenon and the American teenager. Adolesc
48. Reidy TJ. The aggressive characteristics of abused and Med State Art Rev 1990;1:55–70.
neglected children. J Clinical Psychol 1977;33:1140 –5. 66. Jankowski MS. Islands in the street: gangs and American
49. Thornberry TP. Violent families and youth violence. urban society. Berkeley, California: University of Califor-
Office of Juvenile Justice and Delinquency Prevention, nia Press, 1991.
Fact Sheet No. 21. Washington, DC: Department of 67. Huff CR, ed. Gangs in america, 2nd ed. Thousand Oaks,
Justice; 1994. California: Sage Publications, 1996.
50. Hotaling GT, Sugarman DB. An analysis of risk markers 68. National Center for Education Statistics. Gangs and
in husband to wife violence: the current state of knowl- victimization at school: Education policy issues: statisti-
edge. Violence Victims 1986;1:101–24. cal perspectives. Washington, DC: US Department of
51. Kratcoski PC. Youth violence directed toward significant Education; 1995.
others. J Adolesc 1985;8:145–57. 69. National Center for Education Statistics. Student victim-
52. Fagan J, Browne A. Violence between spouses and ization at school, 1993. Washington, DC: US Depart-

270 American Journal of Preventive Medicine, Volume 14, Number 4

 ment of Education, Office of Educational Research and eds. The social ecology of crime. New York: Springer-
Improvement; 1995. Verlag Press; 1986:25– 46.
70. Good TL, Marshall S. Do students learn more in heter- 86. Schurerman, L, Kobrin S. Community careers in crime.
ogeneous or homogeneous groups? In: Peterson PL, In: Reiss AJ, Tonry M, eds. Communities and crime.
Wilkinson LC, Hallinan M, eds. The social context of Chicago, Illinois: University of Chicago Press, 1986:67–
instruction: group organization and group process. New 100.
York: Academic press; 1984. 87. Smith DR, Jarjoura GR. Social structure and criminal
71. Oakes J. Multiplying inequalities: the effects of race, victimization. J Res Crime Delinquency 1988;25:27–52.
social class, and tracking on opportunities to learn math 88. Sampson RJ, Lauritson JL. Violent victimization and
and science. Santa Monica, California: The Rand Cor- offending: individual-, situational-, and community-level
poration; 1990. risk factors. In: Reiss AJ, Roth JA, eds. Understanding
72. Rosenbaum JE. Social implications of educational and preventing violence: panel on the understanding
grouping. In: Berliner DC, ed. Review of research in and control of violent behavior, vol. 3: social influences.
education, vol. 8. Washington, DC: American Educa- Washington, DC: National Academy Press; 1994:1–114.
tional Research Association; 1980. 89. Attar BK, Guerra NG, Tolan PH. Neighborhood disad-
73. Kachur SP, Stennies GM, Powell KE, et al. School- vantage, stressful life events, and adjustment in urban
associated violent deaths in the United States, 1992– elementary-school children. J. Clin Child Psychol 1994;
1994. JAMA 1996;275:1729 –33. 23:391– 400.
74. Kann L, Warren CW, Harris WA, et al. Youth risk 90. Cook PJ. The technology of personal violence. In: Tonry
behavior surveillance, 1995. MMWR 1996;45(SS-4): M, ed. Crime and justice: a review of research, vol. 14.
6,7,32,35. Chicago, Illinois: University of Chicago Press; 1991:1–72.
75. Baugher E, Lamison-White L. Poverty in the United 91. Wright JD, Rossi PH, Daly K. Under the gun: weapons,
States, 1995. US Bureau of the Census, Current Popula- crime and violence in America. Hawthorne, New York:
tion Reports, Series P60-194. Washington, DC: US Gov- Aldine Publishing Co; 1983.
ernment Printing Office; 1996. 92. Zimring FE. Is gun control likely to reduce violent
76. Lamison-White L. Poverty areas in the United States, killings? Univ Chicago Law Rev 1968;35:721–37.
1995. US Department of Commerce, Bureau of the
93. Max W, Rice DP. Shooting in the dark: estimating the
Census, Housing and Household Economics Division,
cost of firearm injuries. Health Affairs Winter 1993;12:
Washington, DC: US Government Printing Office; 1996.
171– 85.
77. Wilson WJ. The truly disadvantaged: the inner city, the
94. Centers for Disease Control and Prevention. Weapon-
underclass, and public policy. Chicago, Illinois: Univer-
carrying among high school students: United States,
sity of Chicago Press; 1987.
1990. MMWR 1991;40:681– 4.
78. Eller TJ. Who stays poor? Who doesn’t? Dynamics of
95. Sheley JF, McGee ZT, Wright JD. Gun-related violence
economic well-being: poverty, 1992–1993. US Depart-
in and around inner-city schools. Am J Dis Child 1992;
ment of Commerce, Bureau of the Census, Current
146:677– 82.
Population Reports, Household Economic Studies, Pub-
lication no. P70-55. Washington, DC: US Government 96. LH Research, Inc. A survey of experiences, perceptions,
Printing Office; June 1996. and apprehensions about guns among young people in
79. National Research Council. Panel on High Risk Youth, America: report to the Harvard School of Public Health.
Commission on Behavioral and Social Sciences and Joyce Foundation Grant, Boston; 1993.
Education. Losing generations: adolescents in high-risk 97. Callahan CM, Rivara FP. Urban high school youth and
settings. Washington, DC: National Academy Press; handguns: a school-based survey. JAMA 1992;267:3038 –
1993. 42.
80. Elliott DS. Youth violence: an overview. presented at The 98. Arria AM, Wood NP, Anthony JC. Prevalence of carrying
Aspen Institute’s Children’s Policy Forum: Children and a weapon and related behaviors in urban school chil-
Violence Conference, February 18 –21; 1994. dren, 1989 to 1993. Arch Pediatric Adolesc Med 1995;
81. Fernandez RM, Harris D. Social isolation and the un- 149:1345–50.
derclass. Chicago, Illinois: Northwestern University, 99. Callahan CM, Rivara FP, Farrow JA. Youth in detention
Center for Urban Affairs and Policy Research; 1990. and handguns. J Adolesc Health 1993;14:350 –5.
82. Harrell AV, Peterson GE. Drugs, crime, and social 100. McNabb SJ, Farley TA, Powell KE, Rolka HR, Horan JM.
isolation: barriers to urban opportunities. Washington, Correlates of gun-carrying among adolescents in south
DC: Urban Institute Press; 1992. Louisiana. Am J Prev Med 1996;12:96 –102.
83. US Department of Commerce. Marital status and living 101. Sheley JF, Wright JD. Motivations for gun possession and
arrangements: Current Population Survey. Publication carrying among serious juvenile offenders. Behav Sci
no. PPL-52. Washington, DC: Bureau of the Census, US Law 1993;11:375– 88.
Government Printing Office; 1996. 102. DiGuiseppi C. Counseling to prevent violent injuries.
84. Sampson RJ. Neighborhood and crime: the structural Washington, DC: US Preventive Services Task Force, US
determinants of personal victimization. J. Res Crime Department of Health and Human Services; 1994.
Delinquency 1985;22:7– 40. 103. Johnston LD, O’Malley PM, Bachman JG. National
85. Sampson RJ. Neighborhood family structure and the survey results on drug use from the monitoring the
risk of criminal victimization. In: Byrne J, Sampson RJ, future study, 1975–1992, vol. 1: secondary school stu-

Am J Prev Med 1998;14(4) 271

 dents. Rockville, Maryland: National Institute on Drug tion of alcohol use and aggression in adolescence. J
Use; 1993. Studies Alcohol 1993;54(suppl11):62–75.
104. Johnston LD, O’Malley PM, Bachman JG. National 114. White HR, Hansell S, Brick J. Alcohol use and aggres-
survey results on drug use from the monitoring the sion among youth. Alcohol Health Res World 1993;17:
future study, 1975–1992, vol. 2: college students and 144 –50.
young adults. Rockville, Maryland: National Institute on 115. Blumstein A. Violence by young people: why the deadly
Drug Use; 1993. nexus? Nat Inst Justice J 1995;August:2–9.
105. Lauritsen JL, Laub JH, Sampson RJ. Conventional and 116. Tolan PH, Guerra NG. What works in reducing adoles-
delinquent activities: implications for the prevention of cent violence: an empirical review of the field. Boulder,
violent victimization among adolescents. Violence Vic- Colorado: The Center for the Study and Prevention of
tims 1992:7:91–108. Violence, Institute for Behavioral Sciences, University of
106. Valois RF, Vincent ML, McKeown RE, et al. Adolescent Colorado; 1994:1–94.
risk behaviors and the potential for violence: a look at
117. Ringwalt CL, Graham LA, Paschall MJ, Flewelling RL,
what’s coming to campus. J Am Coll Health 1993;41:
Browne DC. Supporting adolescents with guidance
141–7.
and employment (SAGE). Am J Prev Med 1996;
107. Osgood, DW. Drugs, alcohol, and adolescent violence.
12(Suppl 5):31– 8.
Boulder, Colorado: The Center for the Study and Pre-
118. Collins JL, Small ML, Kann L, Paterman BC, Gold RS,
vention of Violence, Institute for Behavioral Sciences
Kolbe LJ. School health education. J School Health
University of Colorado; 1995:1–59.
195;65:302–11.
108. Elliott DS, Huizinga D, Menard S. Multiple problem
youth: delinquency, substance use, and mental health 119. Altman E. Violence prevention curricula: summary of
problems. New York: Springer-Verlag; 1989. evaluations. Springfield, Illinois: Illinois Council for the
109. Goldstein PJ. The drugs-violence nexus: a tripartite Prevention of Violence; 1994.
conceptual framework. J Drug Issues 1985;15:493–506. 120. Cohen S, Wilson-Brewer R. Violence prevention for
110. Dembo R, Williams L. Getreu A, et al. A longitudinal young adolescents: the state of the art of program
study of the relationships among marijuana/hashish evaluation. Washington, DC: Carnegie Council on Ado-
use, cocaine use, and delinquency in a cohort of high lescent Development; 1991.
risk youths. J Drug Issues 1991;21:271–312. 121. Powell KE, Hawkins DF, eds. Youth violence prevention:
111. Kandel D, Simcha-Fagan O, Davies M. Risk factors for descriptions and baseline data from 13 evaluation
delinquency and illicit drug use from adolescence to projects. Am J Prev Med 1996;12(Suppl 5).
young adulthood. J Drug Issues 1986;16:67–90. 122. Grossman DC, Neckerman HJ, Koepsell TD, et al.
112. White HR. Early problem behavior and later drug Effectiveness of a violence prevention curriculum
problems. J Res Crime Delinquency 1992;29:412–29. among children in elementary school: a randomized
113. White HR, Brick J, Hansell S. A longitudinal investiga- controlled trial. JAMA 1997;277:1605–11.

272 American Journal of Preventive Medicine, Volume 14, Number 4