Effectiveness of Universal School-Based Programs to Prevent Violent and Aggressive Behavior A...

EPARAEGL

A

I

Ympyaacg“rhri

FWCCtUJDIl

b

CtR

S

ffectiveness of Universal School-Based Programs torevent Violent and Aggressive BehaviorSystematic Review

obert Hahn, PhD, MPH, Dawna Fuqua-Whitley, MA, Holly Wethington, PhD, Jessica Lowy, MPH, CHES,lex Crosby, MD, MPH, Mindy Fullilove, MD, Robert Johnson, MD, Akiva Liberman, PhD,ve Moscicki, ScD, MPH, LeShawndra Price, PhD, Susan Snyder, PhD, Farris Tuma, ScD, Stella Cory, MD, MPH,lenda Stone, PhD, Kaushik Mukhopadhaya, PhD, Sajal Chattopadhyay, PhD,inda Dahlberg, PhD, Task Force on Community Preventive Services

bstract: Universal, school-based programs, intended to prevent violent behavior, have been used atall grade levels from pre-kindergarten through high school. These programs may betargeted to schools in a high-risk area—defined by low socioeconomic status or high crimerate—and to selected grades as well. All children in those grades receive the programs intheir own classrooms, not in special pull-out sessions. According to the criteria of thesystematic review methods developed for the Guide to Community Preventive Services (Com-munity Guide), there is strong evidence that universal, school-based programs decrease ratesof violence among school-aged children and youth. Program effects were consistent at allgrade levels. An independent, recently updated meta-analysis of school-based programsconfirms and supplements the Community Guide findings.(Am J Prev Med 2007;33(2S):S114–S129) © 2007 American Journal of Preventive Medicine

syhocctp

naqihosycs

abcifsbf

ntroduction

outh violence is a substantial public health prob-lem in the United States. In a representativenational survey in 2003, U.S. adults reported

ore than 1.56 million incidents of victimization byerpetrators estimated to be aged between 12 and 20ears—indicating that violent crimes were committed atrate of approximately 4.2 for every 100 youths in that

ge group.1,2 Victims report that youth perpetratorsommit violent acts at a higher rate than any other ageroup.1,2 Two thirds of reports by victims were ofsimple assaults” (i.e., attacks without a weapon and notesulting in an injury requiring more than 2 days ofospitalization); the remaining victimizations were “se-ious violent crimes” (i.e., aggravated assaults, robber-es, or rapes, but not murder, because these were victim

rom the National Center for Health Marketing (Hahn, Fuqua-hitley, Wethington, Lowy, Snyder, Cory, Stone, Mukhopadhaya,hattopadhyay), and National Center for Injury Prevention andontrol (Crosby, Dahlberg), Centers for Disease Control and Preven-

ion, Atlanta, Georgia; Mailman School of Public Health, Columbianiversity (Fullilove), New York, New York; National Institute of

ustice (Liberman), Washington, DC; University of Medicine andentistry of New Jersey (Johnson), Newark, New Jersey; and National

nstitute of Mental Health (Moscicki, Price, Tuma), Bethesda, Mary-and

The names and affiliations of the Task Force members are listed at theeginning of this supplement and at www.thecommunityguide.org.Correspondence and reprint requests: Robert A. Hahn, PhD,

ommunity Guide Branch, Centers for Disease Control and Preven-

iion, 1600 Clifton Road, MS E-69, Atlanta GA 30333. E-mail:[email protected].

114 Am J Prev Med 2007;33(2S)© 2007 American Journal of Preventive Medicine • Publish

urveys). Over the last 25 years, youths aged 10 to 17ears, who constitute less than 12% of the population,ave been involved as offenders in approximately 25%f serious violent victimizations.3 Homicide and sui-ide, respectively, are the fourth and fifth leadingauses of death among children aged 5 to 14 years, andhe second and third leading causes of death amongeople aged 15 to 24 years.4

Risk factors for youth violence include low socioeco-omic status (SES), poor parental supervision, harshnd erratic discipline, and delinquent peers.5 Delin-uent youths commonly have other problems as well,6

ncluding drug abuse, difficulties at school, and mentalealth problems (as indicated by being in the top 10%f the distribution of externalizing and internalizingymptoms in the Child Behavior Checklist7). Theseouths are threats not only for the direct harm they mayause, but also because they may play roles in theocialization of other potential delinquents.8

The prevention of youth violence and aggression, inddition to being of value in itself, is also desirableecause early violent and aggressive behavior is an indi-ator of later problem behaviors. “Early antisocial behav-or may be the best predictor of later delinquency. . . . Inact, early aggression appears to be the most significantocial behavior characteristic to predict delinquentehavior before age 13.”5 Researchers categorize riskactors for early delinquency, including violent behav-
or, within an ecologic framework as individual, family,
0749-3797/07/$–see front mattered by Elsevier Inc. doi:10.1016/j.amepre.2007.04.012

paapd

ydoleytcawy

tv17tci2dwsor

sb1ylsoltieasl

nvgcoar

(r

s8vstbogc7asompsrms

tbtEHsbimttairnaarmap

PltpaiaoctosnI

A

eers, school, and community.9 Factors in all categoriesre thought to contribute to the development of earlynd chronic violent behavior, and all are thought torovide opportunities for intervention to reduce theevelopment of these behaviors.5,10

Rates of the violent victimization of youth by otherouth or by adults are also substantial. During the lastecade, the highest rates of homicide in the U.S.ccurred among people aged 15 to 24 years.11 In the

ate 1990s, reliable data from 14 established marketconomies showed that young people aged 10 to 29ears in the U.S. were homicide victims more than siximes as often as their New Zealand counterparts (theountry with the next highest rate).12 In 2002, inddition to fatal violent crimes, nonfatal violent crimesere committed against approximately 5.6 per 100ouths aged 12 to 19 years.1

Violence in the school environment per se was nothe focus of this review. The most serious forms ofiolent crime rarely occur in schools. Between July 1,992 and June 30, 2000, school-aged youth were at least0 times more likely to be murdered away from schoolhan at school.13 Much nonfatal crime, however, in-luding both nonviolent and violent crime, occursn school facilities or on the way to or from school.13 In003, although rates of violent crime at school hadeclined over the past decade, 740,000 violent crimesere committed against youth aged 12 to 18 years at

chool—approximately 1.3 per 100 youth, and 150,000f these crimes were classified as “serious” (i.e., rape,obbery, sexual assault, and aggravated assault).13

Less serious violent behavior is also prevalent amongtudents. In 2003, 33.0% of high school students reportedeing in a physical fight at least once in the past year,2.8% on school property. Of students aged 12 to 18ears, 7.2% reported having been bullied at school in theast 6 months.13 Approximately 5.4% of high schooltudents surveyed in 2005 reported carrying a gun at leastnce in the past month—a risk factor for serious vio-

ence.14 In high schools in 2005, 7.9% of students werehreatened or injured with a weapon on school propertyn the same year.15 Teachers also reported being threat-ned and injured.13 Violence at school may make studentsfraid to attend school; in 2005, 6.0% of high schooltudents reported feeling too unsafe to go to school ateast once in the past 6 months.14

The purpose of this review was to assess the effective-ess of universal school-based programs in preventingiolent behavior. Thus, studies of school-based pro-rams were reviewed only if they assessed violent out-omes or proxies for violent outcomes. The effects onther outcomes were not assessed systematically, butre reported if they were addressed in the studieseviewed.

The School Health Policies and Programs StudySHPPS)16 indicates that most schools in the U.S.
eport having implemented programs to educate their i
ugust 2007

tudents about violence and nonviolent behavior. In8% of schools, funding or staff development foriolence prevention is available. More than 90% ofchools (elementary, middle, junior and senior high)each about anger management, bullying, prosocialehavior (such as cooperation, praise, or support ofthers), communications skills, decision-making skills,oal-setting skills, and other techniques for avoidingonflict and violence. In 52% of middle schools and7% to 83% of high schools, instruction about sexualssault and dating violence is offered; less than 50% ofchools teach about gun safety. Despite the high levelsf reported instruction, teachers indicate spending aedian of only 4 to 5 hours per year in violence

revention instruction.16 (The number of programessions in the interventions included in the presenteview ranged from 5 to 480, with a median of 36,aking it unlikely that the reviewed programs are

imilar to those surveyed by SHPPS.)Although not always explicit in study reports, several

heories are offered as the basis for universal school-ased programs.17 Diverse approaches to the preven-ion of violent behavior are reviewed by Goleman inmotional Intelligence18; see also Health Behavior andealth Education: Theory, Research and Practice19 for a

ummary of behavior change theories. Theories ofehavior change vary in their focus on individuals;

nterpersonal relations; the physical and social environ-ent, including social norms; and combinations of

hese. Several programs focus on providing informa-ion about the problem of violence and approaches tovoiding it, on the assumption that providing thisnformation to students will lead to its application andeduced violence, and that information is necessary, ifot sufficient, to change behavior. Some programs20

re designed on the converse theory that self-conceptnd self-esteem derive from positive action and itsewards, so that if children’s behavior can be madeore positive and sociable, they will develop a better

ttitude toward themselves and then continue to makeositive choices.Many programs, such as Responding in Peaceful and

ositive Ways21 and Students for Peace,22 cite socialearning theory23 as the foundation of their interven-ion design. The Second Step24,25 program, for exam-le, posits that changing the way children experiencend think about social problems and providing practicen this new way of thinking will improve their behaviorround conflict and aggression; the program focusesn developing children’s competence in empathy, so-ial problem solving, and impulse control. Modeling ofhese competencies by teachers is another componentf this program, and efforts are also made to engagechoolwide and parental support and involvement. Theames of some programs reflect a similar approach:nterpersonal Cognitive Problem-Solving26 and Provid-
ng Alternative Thinking Strategies.27
Am J Prev Med 2007;33(2S) S115

wmpt

T

TomFtt(n(mC(bm

H

TitfseeMtueob

I

AR2

T

O

IR

R

R

R

R

VR

R

R

R

R

R

R

R

R

a

b

o 000 cht

S

Some programs are founded on the theory that theyill be most effective if they modify the broader environ-ent of the child. The multicomponent PeaceBuilders

rogram28 involves the entire school—outside and insidehe classroom—as well as parents and the community.

he Guide to Community Preventive Services

he systematic review in this report represents the workf the independent, nonfederal Task Force on Com-unity Preventive Services (the Task Force). The Task

orce is developing the Guide to Community Preven-ive Services (the Community Guide) with the support ofhe U.S. Department of Health and Human ServicesDHHS) in collaboration with public and private part-ers. The Centers for Disease Control and PreventionCDC) provides staff support to the Task Force for develop-ent of the Community Guide. The book, The Guide toommunity Preventive Services: What Works to Promote Health?available at www.thecommunityguide.org) presents theackground and the methods used in developing the Com-

able 1. Selected Healthy People 2010 30 objectives related to

bjective P

njury preventioneduce hospitalization for nonfatal head injuriesper 100,000 population (Objective 15-1)

A

educe hospitalization for nonfatal spinal cordinjuries per 100,000 population (15-2)

A

educe firearm-related deaths per 100,000population (15-3)

A

educe nonfatal firearm-related injuries per100,000 population (15-5)

A

educe hospital emergency department visits per1000 population (15-12)

A

iolence and abuse preventioneduce homicides per 100,000 population(15-32)

A

educe maltreatment of children per 1000children aged �18 years (15-33a)

C

educe child maltreatment fatalities per 100,000children aged �18 years (15-33b)

C

educe the rate of physical assault by current orformer intimate partners per 1000 people aged�12 years (15-34)

A

educe the annual rate of rape or attemptedrape per 1000 people aged �12 years (15-35)

A

educe sexual assault other than rape per 1000people aged �12 years (15-36)

A

educe physical assaults per 1000 people aged�12 years (15-37)

A

educe physical fighting among adolescents(students in grades 9 through 12, fightingduring the previous 12 months) (15-38)

A

educe weapon carrying by adolescents on schoolproperty (students in grades 9 through 12,carrying during the past 30 days) (15-39)

A

Age adjusted to year 2000 standard population.Note that objective 15-33a is per 1000 children aged �18 years, whebjectives would be reduction of child maltreatment to 1290 per 100,o 1.6 per 100,000.

unity Guide.29 g

116 American Journal of Preventive Medicine, Volume 33, Num

ealthy People 2010 Goals and Objectives

he intervention reviewed here may be useful in reach-ng several objectives specified in Healthy People 2010,30

he disease prevention and health promotion agendaor the U.S. These objectives identify some of theignificant preventable threats to health, and focus thefforts of public health systems, legislators, and lawnforcement officials for addressing those threats.any of the proposed Healthy People objectives in Chap-

er 15, “Injury and Violence Prevention,” relate toniversal school-based programs and their proposedffects on violence-related outcomes. Violence-specificbjectives with potential relevance to universal school-ased programs are listed in Table 1.

nformation from Other Advisory Groups

National Institutes of Health State of the Scienceeview of the Prevention of Youth Violence held in005 recommended several universal school-based pro-

l-based violence prevention programs

ation Baseline 2010 objective

60.6 (1998)a 45.0

4.5 (1998)a 2.4

11.3 (1998)a 4.1

24.0 (1997)a 8.6

131 (1997)a 126

6.5 (1998)a 3.0

en 12.9b (1998) 10.3

en 1.6b (1998) 1.4

scents/adults 4.4 (1998) 3.3




scents 36% (1999) 32%

scents 6.9% (1999) 4.9%

bjective 15-33b is per 100,000 children aged �18 years. Comparableildren aged �18 years and reduction of child maltreatment fatalities

schoo

opul

ll

ll

ll

ll

ll

ll

hildr

hildr

dole

dole

dole

dole

dole

dole

reas o

rams for the prevention of youth violence.31 The

ber 2S www.ajpm-online.net

http://www.thecommunityguide.org

SlgSpySrGTCRii

aagerrSSncmi

M

Ieridhstasii

rddoidfsl

dTmtp

T

Ipaaavfvtcrribtoem

ethbdpvwhbemsesi

C

Ttwpav

egicTsawrasrvr

A

urgeon General’s 2001 report on youth violence10

isted several school-based violence prevention pro-rams as model or promising programs: the Seattleocial Development Project32 was named as a modelrogram for the primary prevention of violence amongouth; the Bullying Prevention program,33 Families andchools Together (FAST) Track34 and its PATHS cur-iculum,27 I Can Problem Solve,26 the Good Behaviorame,35 and Linking the Interests of Families andeachers (LIFT)36 were recommended as promising.onversely, the report noted that the Drug Abuseesistance Education (DARE) program, which may

nclude a violence reduction component, is ineffectiven reducing violence or drug abuse.10

The Center for the Study and Prevention of Violencelso recommends the Bullying Prevention Program37

nd PATHS curriculum27 as model “Blueprint” pro-rams that meet its highest standards of evaluationvidence for experimental design, substantial effect,eplication, and sustainability. Additionally, the Centerecommends the Good Behavior Game, FAST Track,eattle Social Development Project, I Can Problemolve, and LIFT as promising programs.38 The Commu-ity Guide differs from these others insofar as it drawsonclusions on intervention types (which may includeany different specific programs) rather than specific

ntervention programs.

ethods

n the Community Guide, evidence is summarized on (1) theffectiveness of interventions in altering selected health-elated outcomes, and (2) positive or negative effects of thentervention other than those assessed for the purpose ofetermining effectiveness, including positive or negativeealth and nonhealth outcomes.39 When an intervention ishown to be effective, information is also included about (3)he applicability of evidence (i.e., the extent to which avail-ble effectiveness data might apply to diverse populationegments and settings), (4) the economic impact of thentervention, and (5) barriers to implementation of thentervention.

As with other Community Guide reviews, the process used toeview evidence systematically and then translate that evi-ence into conclusions involves forming a systematic reviewevelopment team; developing a conceptual approach torganizing, grouping, and selecting interventions; selecting

nterventions to evaluate; searching for and retrieving evi-ence; assessing the quality of and abstracting informationrom each study; assessing the quality of and drawing conclu-ions about the body of evidence of effectiveness; and trans-ating the evidence of effectiveness into recommendations.

This section summarizes how these methods were used ineveloping the review of universal school-based programs.he review was produced by the systematic review develop-ent team (the team; see author list) and a multidisciplinary

eam of specialists and consultants representing a variety of
erspectives on violence (see Acknowledgments section). a
ugust 2007

he Intervention

n this review, the effectiveness of universal school-basedrograms in preventing violence and aggressive behaviormong children and youth of preschool and school age wasssessed. Evaluated school programs were designed to teachll students in a given school or grade about the problem ofiolence and its prevention or about one or more of theollowing topics or skills intended to reduce aggressive oriolent behavior: emotional self-awareness, emotional con-rol, self-esteem, positive social skills, social problem solving,onflict resolution; and team work. In this review, violenceefers to both victimization and perpetration. Studies wereeviewed only if the reduction of violent or aggressive behav-or was an objective of the program, although it need not haveeen the only or principal objective. In this review, programshat include not only universal classroom interventions, butther components as well, such as modification of the schoolnvironment, community involvement, or parental involve-ent, are referred to as “multicomponent.”Universal programs are defined to mean programs deliv-

red to all children in a given school or grade, not only tohose who had already manifested violent or aggressive be-avior or risk factors for these behaviors. Although they maye part of an overall violence reduction strategy, programsirected exclusively to youths who had already manifestedroblems of violence or were considered at high risk ofiolence were not the focus of this review. Universal programsere included that were targeted by grade and school inigh-risk areas, defined by low SES or high crime rate,ecause, in those selected settings, the programs are deliv-red to all children. Also included were programs imple-ented in special schools, such as schools for children with

pecific disabilities. Our review included pre-kindergarten,lementary, middle, junior high, and senior high schools;chool settings and targeting were noted to assess differencesn effectiveness associated with these variables.

onceptual Approach and Analytic Framework

he general methods for conducting systematic reviews forhe Community Guide have been described in detail else-here.39,40 This section briefly describes the conceptual ap-roach and the determination of outcomes considered inssessing the effects of universal school-based programs oniolence.

The conceptual model, or analytic framework, used tovaluate the effectiveness of universal school-based pro-rams in reducing violence (Figure 1) depicts the flow ofnfluences from the intervention to two broad outcomeategories: violence by youths and victimization of youths.he working hypotheses are that, through influences of

chool personnel, school programs affect student skills,ttitudes, behaviors, and the school environment, each ofhich supports the others. These changes result in aeduction of negative social behaviors (including violencend aggression), improved school behavior, and improvedocial behavior outside of school, leading to an overalleduction in the perpetration of and victimization byiolence. These reductions in violence lead, in turn, toeductions in physical and psychological harm to victims
nd those around them.

O

Sv

Swt

S

EMavCtireepa

e

s

AE

ErtavofistssipeeSotwmcsavc

a

BTaIlIbAR

Fccs

S

utcomes

tudies that assessed any of the following directly measurediolent outcomes were reviewed:

● Self- or other-reported or observed aggression or vio-lence, including violent crime.

● Aggression or violence as observed by the researcher.

tudies examining any of five proxies for violent outcomes,hich may include clearly violent behavior as well as behavior

hat is not clearly violent were also reviewed:

● Measures of conduct disorder (the psychiatric condition,in which “the basic rights of others or major age-appropriate societal norms or rules are violated”).41

● Measures of externalizing behavior (i.e., rule-breakingbehaviors and conduct problems, including physical andverbal aggression, defiance, lying, stealing, truancy, de-linquency, physical cruelty, and criminal acts).42

● Measures of acting out (i.e., aggressive, impulsive, ordisruptive class behaviors) or conduct problems (in-cludes talking in class, stealing, fighting, lying, notfollowing directions, teasing, and damaging property).

● Measures of delinquency, which may include violentbehavior as well as behavior not regarded as violent.

● School records of suspensions or disciplinary referrals.

earch for Evidence

lectronic searches for literature were conducted in theEDLINE, EMBASE, ERIC, Applied Social Sciences Index

nd Abstracts, NTIS (National Technical Information Ser-ice), PsycINFO, Sociological Abstracts, NCJRS (Nationalriminal Justice Reference Service), and CINAHL (Cumula-

ive Index to Nursing and Allied Health Literature) databasesn June and July 2002, and updated in December 2004. Theeferences listed in all retrieved articles were reviewed, andxperts on the systematic review development team andlsewhere were consulted. The studies in this review wereublished as journal papers, governmental reports, books,nd book chapters.

Articles published prior to December 2004 were consid-

igure 1. Analytic framework: School-based programs for vlassroom instruction and/or behavior modification techniqomponents; rounded rectangles, intermediate outcomes; dolid lines, hypothetical causal pathways.

red for inclusion in the systematic review if they evaluated azU


universal school-based program;assessed at least one of the violentoutcomes specified in the analyticframework; were conducted incountries with high-income econ-omies (as defined by the WorldBank);a reported on a primarystudy rather than, for example, aguideline or review; and com-pared a group of people exposedto the intervention with a com-parison group that had not beenexposed or had been less ex-posed. Studies with a total samplesize of �20 students were ex-cluded because results from suchstudies were regarded as unreli-able. While searching for evi-dence, the team also sought infor-mation about effects on otheroutcomes not related to violence,

uch as changes in school performance and drug use.

ssessing the Quality and Summarizing the Body ofvidence on Effectiveness

ach study that met the inclusion criteria was read by twoeviewers who used standardized criteria (available at www.hecommunityguide.org/methods/abstractionform.pdf) tossess the suitability of the study design and threats toalidity.39,40 Disagreements between the reviewers were rec-nciled by consensus among the team members. Our classi-cation of the designs of studies reviewed is in accord with thetandards of the Community Guide review process, and some-imes differs from the classification used in the originaltudies. On the basis of the number of threats to validity,tudies were assigned a number of penalties and character-zed as having good, fair, or limited execution for theurposes of this review.39 Studies with good or fair quality ofxecution, and any level of design suitability (greatest, mod-rate, or least), were included in the body of evidence.tudies with greatest design suitability are those in which datan exposed and control populations are collected prospec-ively; studies with moderate design suitability are those inhich data are collected retrospectively or in which there areultiple pre- or post- measurements, but no concurrent

omparison population; and studies with least suitable de-igns are those in which there is no comparison populationnd only a single pre- and post- measurement in the inter-ention population. In this review, a sensitivity analysis wasonducted, which excluded studies with least suitable designs.

Countries with high-income economies (as defined by the Worldank) are Andorra, Antigua and Barbuda, Aruba, Australia, Austria,he Bahamas, Bahrain, Barbados, Belgium, Bermuda, Brunei, Can-da, Cayman Islands, Channel Islands, Cyprus, Denmark, Faeroeslands, Finland, France, French Polynesia, Germany, Greece, Green-and, Guam, Hong Kong (China), Iceland, Ireland, Isle of Man,srael, Italy, Japan, Republic of Korea, Kuwait, Liechtenstein, Luxem-ourg, Macao (China), Malta, Monaco, Netherlands, Netherlandsntilles, New Caledonia, New Zealand, Norway, Portugal, Puertoico, Qatar, San Marino, Singapore, Slovenia, Spain, Sweden, Swit-

ce prevention usingCircles, intervention

line, optional link;

iolenues.ashed

erland, Taiwan (China), United Arab Emirates, United Kingdom,nited States, Virgin Islands (U.S.).


http://www.thecommunityguide.org/methods/abstractionform.pdf

http://www.thecommunityguide.org/methods/abstractionform.pdf

ymdfc

1

wgtric

2

3

trpaBrpiE

u(7Q5asmaiw

vtd(pdbmcos

si

scrobwsulcotrw

patipe

A

Iiteamia

Ot

Aleamivwt

stefioCp

S

Siir

A

The primary outcomes assessed in this review—violence byouths and victimization of youths—are referred to as recom-endation outcomes because, if evidence of change in the

esired direction is sufficient or strong, they provide the basisor recommending the intervention. Relative percentagehange was calculated using the following formulas:

. For studies with before-and-after measurements and con-current comparison groups:

Effect size � �Ipost/Ipre�/�Cpost/Cpre� � 1

here Ipost � last reported outcome rate in the interventionroup after the intervention, Ipre � reported outcome rate inhe intervention group before the intervention, Cpost � lasteported outcome rate in the comparison group after thentervention, and Cpre � reported outcome rate in theomparison group before the intervention.

. For studies with post measurements only and concurrentcomparison groups:

Effect size � �Ipost � Cpost�/Cpost

. For studies with before-and-after measurements but noconcurrent comparison:

Effect size � �Ipost � Ipre�/Ipre

The effect of the intervention is reported as desirable whenhe intervention was associated with a decrease in a violence-elated outcome, compared with its effect in the controlopulation, and as undesirable when the intervention wasssociated with an increase in the violence-related outcome.ecause the outcome of interest is violence or aggression, aeduction in the intervention compared with the controlopulation is desirable; thus, negative effect sizes, which

ndicate a decrease in violence or aggression, are desirable.ffects are reported as percentages.To report effect sizes from multiple studies, the median is

sed, and, for seven or more effect sizes, the lower quartileQ1, the 25th percentile) and the upper quartile (Q3, the5th percentile), reported as the interquartile interval (IQI).1 and Q3 provide information on the range of the middle

0% of the study effect sizes, and can therefore be interpreteds reflecting the range of typical effects. In some cases, aelection had to be made among several possible effecteasures for the summary measures of effectiveness. When

vailable, effect measures adjusted for potential confoundersn multivariate analyses, rather than crude effect measures,ere included.The classification and reporting of grades and school levels

ary among U.S. studies, and European and U.S. classifica-ions do not correspond. Programs were classified as beingirected to pre-kindergarten/kindergarten, elementarygrades 1 to 5), middle (6 to 8), and high (9 to 12) as well asossible, using available information on age and programuration. In addition, some programs included students inoth elementary and middle school and others include bothiddle and high school. To limit the number of grade

lassifications and because middle school includes fewer yearsf program exposure (three) than elementary (five) or high
chool (four), programs including elementary and middle l
ugust 2007

chool were classified as elementary programs, and programsncluding middle and high school as high school programs.

Length of follow-up (i.e., time from the conclusion of thechool program to the assessment of the outcome) wasonsidered when deciding which studies to include in theeview. No studies were excluded from the evaluation strictlyn the basis of an insufficient follow-up period. However,ecause enduring effects were of the most interest, studiesith follow-up periods of �1 year were penalized. For the

ame reason, if the intervention program had multiple eval-ations at different follow-up points, the evaluation at the

ongest follow-up period that had an attrition rate �30% washosen. The strength of the body of evidence was summarizedn the basis of the number of available studies, the strength ofheir design and execution, and the size and consistency ofeported effects using Community Guide standards.39 Resultsere presented graphically and statistically.The reviewed studies were classified in terms of several

rogram characteristics, and the variability of program effectsssociated with these characteristics was explored. Becausehese comparisons are bivariate and were not tested withnferential statistics, differences among strata are best inter-reted as suggestive rather than definitive evidence of trueffect differences.

pplicability

f an intervention was found to be effective, evidence regard-ng its applicability was assessed in diverse settings, popula-ions, and circumstances. Note is made of whether existingvidence derives from limited conditions, making its gener-lizability uncertain. The goal of this assessment is the deter-ination of known and unknown conditions under which the

ntervention is effective, and thus, the known limits of itspplication.

ther Effects, Barriers to Implementation of Interven-ions, and Economic Evaluations

s noted, the Community Guide review of school-based vio-ence prevention programs did not systematically assess theffects of this intervention on other outcomes (e.g., drugbuse, school achievement, truancy, psychological adjust-ent). However, some of the noted benefits are mentioned

n the reviewed studies. The potential harms of school-basediolence prevention programs are also noted if these harmsere mentioned in the effectiveness literature or were

hought to be of importance by the team.In Community Guide reviews, barriers to implementation are

ummarized and economic evaluations are performed only ifhe intervention is shown to be effective. In this review, theconomic evaluations identified assessed the costs and bene-ts of the interventions in terms of the violent and otherutcomes that are the focus of the review. Methods used inommunity Guide economic evaluations have been publishedreviously.43,44

ummarizing Research Gaps

ystematic reviews in the Community Guide identify existingnformation on which to base public health decisions aboutmplementing interventions. An additional benefit of theseeviews is identification of areas in which information is
acking or of poor quality. To summarize these gaps in the

lt

RS

Fsco3wfifglssdsltiy

S

Tsahsabl

ibmsviaclgtctapotatt

ble

2.Pr

edom

inan

tch

arac

teri

stic

sof

sch

ool

prog

ram

sby

sch

ool

leve

l

ade

leve

lst

udie

s)Fo

cus

(%)

Inte

rven

tion

stra

tegy

(%)

Per

sonn

el(%

)Se

ttin

g(%

)

5da

ys/w

eek,

%Fr

eque

ncy

(day

s/w

eek)

/du

rati

on(w

eeks

)

Scho

ols

inlo

wSE

San

d/or

high

crim

era

te%

Maj

orit

yra

ce/

ethn

icit

yof

stud

ypo

pula

tion

,by

%of

stud

ies

K-K

(7)

Dis

rupt

ive/

anti

soci

albe

hav

ior

(86)

Cog

nit

ive/

affe

ctiv

e(5

0)So

cial

skill

s(3

3)T

each

ers

(100

)C

lass

room

(100

)33 3.

2/16

.450

Bla

ck50

%W

hit

e33

%m

enta

ry(2

8)D

isru

ptiv

e/an

tiso

cial

beh

avio

r(5

7)G

ener

alvi

olen

ce(2

9)

Soci

alsk

ills

(48)

En

viro

nm

enta

lch

ange

(14)

Tea

cher

s(7

9)C

lass

room

(93)

40 3.3/

30.3

35B

lack

27%

Wh

ite

60%

ddle

(30)

Dis

rupt

ive/

anti

soci

albe

hav

ior

(30)

Gen

eral

viol

ence

(43)

Bul

lyin

g(1

7)

Kn

owle

dge/

info

rmat

ion

(21)

Soci

alsk

ills

(52)

Tea

cher

s(6

6)R

esea

rch

ers

(14)

Non

-sch

ool

pers

onn

el(1

0)

Cla

ssro

om(9

3)35 2.

9/24

29B

lack

37%

Wh

ite

53%

His

pan

ic11

%

gh(4

)G

ener

alvi

olen

ce(5

0)D

atin

g(2

5)B

ully

ing

(25)

Kn

owle

dge/

info

rmat

ion

(33)

Soci

alsk

ills

(33)

Sch

ool

envi

ron

men

tch

ange

(33)

Tea

cher

s(6

6)St

uden

tpe

ers

(33)

Cla

ssro

om(1

00)

0 2.0/

2450

Wh

ite

100%

e:In

this

anal

ysis

,pro

gram

ssp

ann

ing

sch

ool

leve

ls(e

.g.,

elem

enta

ryan

dm

iddl

e)ar

eco

unte

dat

each

span

ned

leve

l(i

.e.,

elem

enta

ryan

dm

iddl

e).B

ecau

seof

smal

ln

umbe

rsof

stud

ies,

only

jor

prop

orti

ons

ofpr

ogra

mch

arac

teri

stic

sar

epr

esen

ted;

perc

enta

ges

may

not

tota

l10

0.,

soci

oeco

nom

icst

atus

.

S

iterature, remaining research questions for each interven-ion evaluated are identified by the team.

esults, Part I. Intervention Effectivenesstudy Characteristics

ifty-three studies20,25,26,32,33,35,36,45– 62,64 –91 of univer-al school-based programs that met our inclusionriteria were found. Of these, seven35,51,55,59,60,62,76 weref greatest design suitability and good execution,225,26,32,36,45–50,53,56–58,61,64–68,71,72,75,77,78,82,83,86,87,89–91

ere of greatest design suitability and fair execution,ve20,33,54,80,88 were of moderate design suitability and

air execution, one52 was of least suitable design andood execution, and eight69,70,73,74,79,81,84,85 were ofeast suitable design and fair execution. Study sampleizes ranged from 21 to 39,168 students, with a medianample size of 563. Forty-one studies (77.4%) employedirect measures of violence or aggression, and 12tudies (23.6%) used proxy outcome measures. Fol-ow-up time from the conclusion of the intervention tohe final assessment ranged from none (assessmentmmediately following the end of the intervention) to 6ears; the median follow-up time was 6 months.

chool Program Characteristics

he characteristics of school programs were first as-essed at four different grade levels: pre-kindergartennd kindergarten (pre-K/K), elementary, middle, andigh school. In this analysis, programs spanning twochool levels (e.g., elementary and middle) are countedt each spanned level (i.e., elementary AND middle)ecause the program characteristics apply to both

evels.Several trends are associated with grade level group-

ngs (Table 2). Focus on disruptive and antisocialehavior is common in elementary school and iniddle school. In middle and high school, the focus

hifts to general violence and to specific forms ofiolence, including bullying and dating violence. Thentervention approach shifts from a cognitive/affectivepproach, which focuses on modifying behavior byhanging the cognitive and affective mechanismsinked with such behavior, to an approach that makesreater use of social skills training, which emphasizeshe development of behavioral skills rather thanhanges in cognition, consequential thinking, or affec-ive processes. With increasing grade level, there maylso be a decreased focus on the teacher as the primaryrogram implementer and an increase in the use ofther personnel such as student peers or members ofhe team conducting the research study. Understand-bly, because of our definition of the intervention inhis review, the classroom is the principal setting of
hese school programs at all grade levels. Finally, no

Ta

Gr

(n Pre

Ele

Mi

Hi

Not

ma

SES


cp

vppp

E

CttcfswB

eoa

G1d�sbptbiuri

T

C

G

I

P

P

C

M

N interS

A

lear trends are apparent in frequency and duration ofrograms by school level.Among studies for which the information is pro-

ided, many are offered in settings characterized by theresence of low SES, high crime rates, or both. Manyrograms are conducted in communities with largeroportions of minority students.

ffectiveness

omparison of program characteristics and popula-ions served at different school levels indicates substan-ial heterogeneity by level and intercorrelation amongharacteristics. Thus, bivariate analysis of program ef-ects by program characteristics might misleadinglyuggest the causal association of these characteristics,hich are actually confounded by other associations.

able 3. Bivariate analysis of effect size by demographic and

haracteristic (n studies)

Medin inpop

radeAll grades combined (65) �15Pre-K/K (6) �32Elementary (34) �18Middle (21) �7High (4) �29

nterventionInformation conveyed (10) �8Cognitive/affective (6) �14Social skills (30) �19Environmental change - classroom (3) �15Environmental change - school (12) �11Peer mediation (2) �61Behavior modification (0) —

rogram focusGeneral violence (19) �10Disruptive or antisocial behavior (33) �19Bullying (10) �6Gang activity (2) �5Dating violence (1) �29

rimary program personnelStudents/peers (4) �41Teachers (49) �17Administrators/counselors (3) 34Non-school personnel (2) �5Researchers (7) �7

ommunity environmentNot stated (14) �1Not low SES/not high crime (24) �21Not stated plus not high crime/low SES (38) �11High crime/low SES (15) �29ajority ethnicityBlack (15) �16White (22) �20Hispanic (6) �0No information provided (13) �30No clear majority (8) �10

ote: Effect sizes calculated as relative change in violent outcomes inES, socioeconomic status; K, kindergarten.

ivariate associations of program characteristics with I

ugust 2007

ffect sizes are provided, recognizing the potential forther program characteristics to confound apparentssociations (Table 3).

rade. For all grades combined, the median effect was a5.0% relative reduction in violent behavior among stu-ents who received the program (IQI, �44.2% to2.3%). The effects of school programs were found at all

chool levels, from a 7.3% relative reduction in violentehavior among middle school students who received therogram (15 studies,49,52,53,55–59,76–78,81,84,88,90 IQI, �35.2%o 2.3%) to a 29.2% relative reduction in violentehavior among high school students (four stud-

es,66,71,82,87 percentiles not calculated) (Table 3, Fig-re 2). In elementary school programs, the medianeduction of violent behavior was 18.0% (26 stud-es,20,25,32,33,35,36,45,46,48,51,54,64,65,67–70,73,75,79,80,84 – 87,89

ram variations

ffect (relative % changention versus controlns) 25th and 75th percentiles

�44.1, �2.3—�44.8, �2.5�35.2, 2.3—

�22.9, 18.3—�35.2, �2.1—�63.6, �1.7——

�50.0, �1.7�44.3, �2.8�64.8, 17.2——

—�44.3, �2.3——�42.5, 2.3

�10.3, 3.3�50.0, �5.2�44.4, �1.4�42.5, �6.7

�44.3, �5.2�40.2, �5.0—�44.4, 8.0�87.5, �1.4

vention compared with control populations.

prog

ian eterve

ulatio

.0

.4

.0

.3

.2

.6

.0

.1

.0

.7

.2

.3

.1

.7

.3

.2

.6

.5

.4

.3

.3

.6

.0

.2

.2

.8

.4

.5

.9

.3

QI, �44.8% to �2.5%). In pre-K/K school programs,


ti

idpt4ipfsKti�sacwoi

arhcbrs

Itatts

Wadeso

SelwFcwwi�waidancr(i(�w

PspIt(tms

Fcipa

Fc

S

he median effect size was �32.4% (six stud-es,26,47,72,74,83,91 percentiles not calculated).

A sensitivity analysis was conducted to examinentervention effects when studies with least suitableesigns (i.e., those without concurrent comparisonopulations who were unexposed or less exposed to

he intervention) were not included in the analysis. The0 studies of greatest or moderate design suitabil-ty20,25,26,32,33,35,36,45–49,51,53–59,64–68,71,72,75–78,80,82,83,86–91

rovided 53 data points. Overall, the median ef-ect size was �13.9% (IQI, �34.5% to �2.0%). Bychool level, the median was �30.9% for pre-K/

(five studies,26,47,72,83,91 percentiles not calcula-ed), �17.5% for elementary school (20 stud-es,20,25,32–36,45,46,48,54,64,65,67,68,75,80,86,87,89 IQI,45.3% to �2.3%), �7.3% for middle school (12

tudies,49,53,55–59,76 –78,88,90 IQI, �31.5% to 2.3%),nd �22.0% for high school (four studies,66,71,82,87 per-entiles not calculated). Thus, results were consistenthether studies with least suitable design were includedr not. The remaining analyses were thus conducted

ncluding studies with least suitable design.Associations between several program characteristics

nd effect size were explored. The goal of this explo-ation was to develop hypotheses that might explain theeterogeneity of program effects. Because of the inter-orrelation of program characteristics noted above, thisivariate presentation should be regarded as simplyeporting empirical associations rather than as an as-essment of causal explanations for effect variability.

ntervention strategies. All school program interven-ion strategies (e.g., informational, cognitive/affective,nd social skills building) were associated with a reduc-ion in violent behavior. All program foci (e.g., disrup-ive or antisocial behavior, bullying, dating violence)

igure 2. Relative change in violence-related outcome asso-iated with school program by school level. Vertical linesnside boxes show medians, ends of boxes show 25th and 75thercentiles, ends of whiskers show 10th and 90th percentiles,nd circle shows an outlier.

imilarly were associated with reduced violent behavior. e


ith the exception of programs administered by schooldministrators or counselors, violent behavior was re-uced in programs administered by all personnel typesxamined, including students and peers, althoughome of these effect sizes were based on small numbersf studies.

chool environments. The effects of programs deliv-red in school environments defined by the presence ofower SES, high rates of crime, or both, were comparedith school environments without these characteristics.or 14 studies,26,35,54,56,57,67,69,75,81,82,85,86,89,91 theseharacteristics were not described. In environmentsith lower SES, high crime rates, or both, effectivenessas consistent with overall study results (15 stud-

es,32,34,36,52,53,58–60,64,71–74,76,79 median �29.2%, IQI,42.5% to �6.7%). It appeared that the programsere similarly effective in settings in which lower SESnd high crime rates were absent (24 stud-

es,20,25,27,43,46–50,55,61,62,65,66,68,70,77,78,80,83,84,87,90,92 me-ian �21.0%, IQI, �50.0% to �5.2%). If it could bessumed that data points for which this information wasot provided actually represented settings without theseharacteristics (which may thus not have been regarded asemarkable or noteworthy), the combined data pointsno information and the absence of these characteristics)ndicated an apparently smaller relative reduction of 11.2%38 studies,20,25–27,33,46–50,54–57,61,62,65–70,75,77,78,80–92 IQI,44.4% to �1.4%), which was still consistent in directionith overall study results.

redominant ethnicity. Finally, the effects of universalchool-based programs were explored in terms of theredominant ethnicity of the study school population.n schools in which the population was �50% black,he median reduction in violent behavior was 14.9%13 studies,26,35,52,53,56,57,60,73,74,76,79,83,86 IQI, �29.2%o 0.50%). Where the population was �50% white, the

edian reduction in violent behavior was 17.7% (16tudies,25,33,36,45,49,54,58,59,66,67,72,75,82,87,88,91 IQI, �34.1%

–100

–80

–60

–40

–20

0

20

40

60

0 10 20 30 40 50 60 70 80Follow-up (months)

Effe

ct s

ize

igure 3. Relative change in violence-related outcome asso-iated with school program by months following program
nd.

tt(ttms�ts

Lmeg(eaTeaer(

TrfdlgoflHOyrltgc

tef

A

UtTchcPb

sp

O

Ohsprih

papdwrw

E

SovqdptacstSfe

cbfismP$actctto

fiPC

A

o �6.3%). Where the population was �50% Hispanic,he median reduction in violent behavior was 18.5%three studies,64,77,78 percentiles not calculated; based onhree studies, this estimate may be unreliable). In studieshat did not provide information on ethnicity, the

edian reduction in violent behavior was 10.3% (15tudies,20,34,47,50,55,62,68–71,81,85,89,90,93 IQI, �47.1% to2.5%); and in studies with no clear majority population,

he median reduction in violent behavior was 23.0% (fivetudies,32,46,48,61,80 IQI �50.0% to �7.3%).

ength of program exposure. To explore whetherore program exposure was associated with larger

ffects, the associations between effect size and pro-ram frequency (sessions per week), as well as durationin months), were examined graphically. An outlierffect size was omitted from this analysis. No clearssociation emerged for either frequency or duration.he association between effect size and total programxposure (the product of the number of days/weeknd the number of weeks, that is, total number ofxposures to the program) was also assessed. No clearelationship was discerned. The correlation coefficientsr) were less than 0.17 for all of these associations.

ime since program end. To determine whether theeduction in violent behavior diminished as the timeollowing the end of the intervention increased (i.e., aecay in program benefit), the association between

ength of follow-up time and effect size was assessedraphically (Figure 3). Again, an outlier effect size wasmitted from this analysis. Although the review includedew studies with long-term follow-up, it appeared thatonger follow-up was associated with smaller effect size.

owever, the R2 for this association was less than 0.03.mitting the longest follow-up times (48 and 72 months)

ielded similar results, that is, an apparent decrease in theeduction in violent behavior was still associated withonger follow-up (graph not presented). Thus, within theime periods assessed, the effectiveness of school pro-rams appears to decrease slightly as time since theonclusion of the program increases.

Overall, according to Community Guide methods,39

he results of this systematic review provide strongvidence that universal school-based programs are ef-ective in reducing violence.

pplicability

niversal school-based programs were found to be effec-ive at all school levels and across different populations.he reviewed studies assessed the effects of programs inommunities characterized by the presence of lower SES,igh rates of crime, or both, as well as in communitiesharacterized by the absence of both of these factors.rogram effectiveness was also assessed in predominantly
lack and predominantly white school populations; few n
ugust 2007

tudies were conducted in predominantly Hispanic schoolopulations.

ther Positive and Negative Effects

ther benefits of universal school-based programsave been noted, and supporting evidence exists forome of these effects. Improvements have been re-orted for social behavior more broadly,64 includingeduced drug abuse56 and delinquency.25 Substantialmprovements in school attendance and achievementave also been reported.20,34,65,94

Potential harms are difficult to imagine, but it isossible that, during exposure to ideas and expressionsbout violence, students’ unaddressed or unresolvedersonal issues with violence may be aroused. Noescriptions or evidence of this situation occurringere found in the reviewed literature. Such occur-ences might call for the linking of affected studentsith counseling services.

conomic Efficiency

chool-based programs to prevent violent behavior areften associated with benefits not directly related toiolence or aggression, such as reduced rates of delin-uency and substance abuse, improved school atten-ance, higher educational attainment, and better jobrospects. However, many benefits from such preven-ion programs would be future benefits to taxpayersnd potential crime victims from reduced crime.95 Aomprehensive economic evaluation of these programshould consider all current and future benefits both tohe participants and to members of society at large.uch a task is made difficult because of the longollow-up period required to properly capture all ben-fits throughout the lifetime of program participants.Most studies identified by our search reported the

osts of programs but no economic summary measuresased on both costs and benefits. Vazsonyi et al.96

ound cost per child to be �$200 for the PeaceBuildersntervention implemented in nine schools in the Tuc-on metropolitan area. Aber et al.97 found the imple-entation costs for the Resolving Conflict Creatively

rogram in 15 New York City elementary schools to be2449 per teacher and $98 per child. Greenberg etl.27,45 estimated that the program costs for the PATHSurriculum over a 3-year period would range from $15o $45 per student per year, depending on whether theurrent staff was redeployed or a new on-site coordina-or was hired. These costs appear minimal compared tohe potential benefits of reduced violence, as well as thether benefits of such programs.The only study that estimated both costs and bene-

ts95 was based on the Seattle Social Developmentroject.32 This study was rated as good according to theommunity Guide’s quality assessment criteria for eco-
omic studies.44 The average effect size for this pro-

gcotjfUatdsoctTfm

psMap

B

SmateHtsstmddd

ficttippwtpe

C

Aon

qsace

R

Ast

●

●

●

●

●

●

D

Tl

S

ram, which focused on elementary schools in a high-rime urban area, was a decrease of 13% in basic crimeutcomes. The total benefits, including cost savings toaxpayers because of reduced expenses for the criminalustice system and reduced personal and property lossesor crime victims, were estimated to be $14,426 in 2003.S. dollars per participant. Net saving per participant

mounted to $9837. In terms of cost– benefit ratio,his program showed a benefit of $3.14 for everyollar invested in the program. Although the effectize found in the study is near the median effect sizef other school programs, this program is moreomprehensive, more intensive and of longer dura-ion than many programs, as well as more costly.hus, the cost benefit of many programs may differ

rom that determined for the Seattle Social Develop-ent Project.In summary, investment in universal school-based

rograms to prevent violence has the potential forignificant positive economic returns in the future.

ore studies are needed, however, to establish firmlynd conclusively the economic benefits from suchrograms.

arriers to Intervention Implementation

chools and their curricula are subject to many require-ents and demands. Because the reduction of violence

nd aggression may not be seen as necessary or centralo instructional goals, it may be difficult to introduceffective programs into school curricula and schedules.owever, the beneficial effects of many programs on

raditional academic outcomes, such as attendance andchool performance, may enhance the interest ofchool policy makers, administrators, and teachers inhese programs. The need for teacher training also may

ake acceptance and implementation of the programsifficult. Finally, some programs may make additionalemands on parents and the community, and theseemands may be resisted.Few of the studies reviewed have measured the

delity of program implementation.25,56,98 Where out-ome information was reported by level of implemen-ation, findings from the highest level of implementa-ion available were used, on the assumption that thisnformation best represented the potential effect of therogram.99 Fidelity can be a substantial obstacle torogram success, and may be particularly problematichen implemented by communities without investiga-

or scrutiny. It may be important for programs torovide ways in which school or community implement-rs can monitor fidelity.100

onclusion

ccording to Community Guide methods,39 the numberf studies in this systematic review overall and the
umber of studies at each grade level, of adequate c

uality, consistency of effect, and effect size, providetrong evidence that universal school-based programsre associated with decreases in violence-related out-omes. Beneficial results were found at all school levelsxamined, from pre-kindergarten through high school.

esults, Part II. Research Issues

lthough strong evidence was found that universalchool-based programs to prevent violence are effec-ive, important research issues remain.

Some school programs are more effective than oth-ers. What characteristics of the programs, or perhapsof the settings in which they are implemented, makesome programs or settings more or less effective?There appears to be a decrease in program effective-ness as time after program completion passes. It willbe important to explore ways to extend the benefitof programs, either within the programs themselvesor with booster programs.Are school programs equally effective for high-risk andlow-risk children, and in high-risk and low-risk environ-ments? Are programs targeted to high-risk childrenoverall more effective, and, if so, more cost effective,than universal programs?Many programs assessed in the review were notongoing, standing programs, but instead were con-ducted for purposes of research. Because researchprograms are often more effective than ongoingprograms—perhaps because of the intensity of mon-itoring and implementation—it will be important tounderstand what maximizes the effectiveness andsustainability of ongoing programs.In what ways is the effectiveness of universal school-based programs to prevent violence moderated bythe predominant ethnicity of the student popula-tion? How might addressing cultural and socialdifferences in diverse populations improve the effec-tiveness of school programs?Studies of the economic efficiency of school pro-grams, measured, for example, as net benefits orcost-benefit ratio, should assess not only violent orcriminal behavior averted, but all current and futuresocial, health, academic, and labor market outcomesassociated with school violence prevention pro-grams. It will be interesting to assess what proportionof the total benefits is crime-related. It will beimportant to assess the extent to which the data usedin the derivation of the summary measure are na-tionally representative.

iscussion

his review is subject to several limitations common toiterature syntheses. Classifying interventions, their
omponents, and the outcomes assessed in evaluation

snuccataattcppp

upsdTorl

bspcasbawv

stcma

LbgpaeidTLtt

aa

ssottaGaGas

y5tisrtrmpcsrnbisLdsh1vbrGi

iGaerpaldtWgrawi

A

tudies is challenging, because interventions often areot described in sufficient detail, study reports do notse common vocabulary, common terms do not alwaysarry the same meaning, and descriptions may notonform to what was actually carried out. Interventionsnd outcomes were classified as well as possible, givenhe available descriptions in study reports; and anttempt has been made to be as transparent as possiblebout the procedures. Problems of program implemen-ation may also affect the evaluation of programs, sohat when a program is reported not to work, it is notlear whether a program truly does not work even whenerfectly implemented or simply that an efficaciousrogram was not adequately executed. The quality ofrogram implementation is not routinely reported.It should be recalled that this review addresses only

niversal school violence prevention programs, that is,rograms delivered to all children in a given schooletting, and not programs targeted exclusively to chil-ren considered to be at high risk of violent behavior.hus, conclusions can be drawn only about the effectsf the universal programs reviewed, and not about theelative effects of universal versus more targeted vio-ence prevention programs.

School violence prevention programs are unlikely toe a panacea for societal violence. Childhood aggres-ive and violent behavior is an important risk factor andredictor of later violent behavior and violent criminalareers. However, school programs do not eliminateggressive and violent behavior, and may not affectome children who are at highest risk of later violentehavior. Nevertheless, it is likely that the programsssessed here, if widely and thoroughly implemented,ould result in a substantial reduction of societaliolence.

A recently updated meta-analysis by Wilson and Lip-ey101 used roughly the same definition of the interven-ion as used in this report, and assessed similar out-omes. Commonalities and differences in the scope,ethods, and results of both reviews were explored to

ssess and extend the findings of both.As in the present Community Guide review, Wilson and

ipsey were interested in the effectiveness of school-ased programs intended to reduce violence and ag-ressive behaviors. In addition to reviewing “universal”rograms, in which all children in a classroom receiveprogram, they also reviewed programs addressed

xclusively to children who either had previously man-fested substantial violent or aggressive behavior or whoemonstrated risk factors for such behavior (Table 4).hese latter programs were not included. Wilson andipsey, however, stratified their results by program

ype; results could, therefore, be compared, excludinghese selected or indicated programs.

Detailed comparison of the frameworks, definitions,nd methods of the Community Guide and the Wilson
nd Lipsey reviews indicates substantial overlap, but i
ugust 2007

ome differences as well (Table 4). The two reviews areimilar in intervention types and typologies and inutcomes assessed; however, the reviews differ in theiming of preferred outcomes—immediately followinghe intervention (Wilson and Lipsey) versus as longfter the intervention as data are available (Communityuide review). Wilson and Lipsey included a broaderrray of literature sources than did the Communityuide, whereas the Community Guide allowed a broaderrray of methods in the studies reviewed and includedtudies in a broader group of national settings.

The present review and the Wilson and Lipsey anal-sis reviewed 27 studies in common, approximately0.9% of the total number (53) of studies included inhe present review and 43.5% of the total (62) includedn the Wilson and Lipsey review. (The numbers oftudies reported here do not match the numberseported by Wilson and Lipsey because the counts inheir analysis included studies reporting results sepa-ately for distinct populations, such as males and fe-ales. Just as we did in counting studies in the

resent review, we also collapsed such separatelyounted results in counting their studies.) Of thetudies included in Wilson and Lipsey and not in thiseview, 15102–116 were dissertations, four117–120 wereot published, and seven studies could not be usedecause of limited quality of execution121–125 or

ncommensurate outcome assessments.126,127 Of thetudies included in this review and not in Wilson andipsey, nine52,69,70,73,74,79,81,84,85 were of least suitableesign, four33,47,84,87 were conducted in non-English-peaking countries, two45,68 were conducted amongandicapped children, and the remaining120,32,36,48,54,65,66,76,80,88,90 were not included for aariety of methodologic reasons. The lack of overlapetween these reviews is substantial, allowing eacheview to serve as a sensitivity analysis for the other.iven the lack of overlap, the consistency of findings

s remarkable.Basic findings of Wilson and Lipsey’s meta-analysis

ndicate that the associations reported in the Communityuide review are not greatly confounded; the meta-nalysis, controlling for measured potential confound-rs, reaches conclusions similar to those of the presenteview. Both reviews found evidence that school-basedrograms for the prevention of violence are effective atll school levels. Although effect sizes differ by schoolevel, Wilson and Lipsey do not draw a conclusion ofifferences in effectiveness, because the program con-ent and outcome measures also differ by school level.

ilson and Lipsey found that the programs havereater effect on elementary school children in high-isk than in non high-risk settings. Wilson and Lipseylso found that program duration is inversely associatedith effect size; they speculate that programs of differ-

ng duration may also differ in intensity. The (standard-
zed) effect size (reported as positive when the outcome

wwnWog

tbia

Ta

T

L

S

G

I

O

O

M

P

I

S

as relatively improved in the intervention comparedith the control populations) of 0.06 for multicompo-ent programs (referred to as “comprehensive” byilson and Lipsey), although statistically significant, is

ne third that for single component, classroom pro-

able 4. Universal school-based programs to prevent violencnd Lipsey101 meta-analysis

Community(53 studie

arget populationsUniversal programs IncludedSelected/indicated Not includSpecial schools/classes Included i

studen

iterature reviewedPublished studies YesGovernment publications YesUnpublished studies NoDissertations No

tudy designProspective studies, with

comparison populationYes

Retrospective studies or studies withmultiple pre/post-assessmentswithout comparison

Yes

Pre-post studies without comparisonpopulation

Yes (sensitcompa

eographic inclusion Studies cohigh-ipublis

ntervention focusBehavioral strategies IncludedCognitively oriented strategies IncludedSocial skills programs IncludedCounseling, talk therapy Included (

classroMulticomponent (involving

interventions beyond classroom)Included i

studen

utcomesAggressive or violent behavior (e.g.,

fighting, bullying, crimes againstpersons)

Included

Disruptive behavior (e.g., behaviorproblems, conduct disorder,acting out)

Included

Problem behavior (e.g., includeboth internalizing andexternalizing behavior)

Included

utcome assessmentFollow-up Latest outAdjustment for small sample sizes

and for extreme valuesNo

Effect size measure Relative ch

ethod of handling extreme outliers Deleted fobut inIQIs

ercentage of studies in common 50.9%

QI, interquartile interval.

rams (0.18). l


Wilson and Lipsey also found additional benefits ofhese programs beyond those on violent or aggressiveehavior, including reduced truancy and improvements

n school achievement, “problem behavior,” activity levels,ttention problems, social skills, and internalizing prob-

mparison of Community Guide systematic review and Wilson

e systematic review Wilson and Lipsey meta-analysis (62 studies)

IncludedIncluded

inistered to allclass

Included (excludinghandicappedchildren)

YesYesYesYes

Yes (if same cohortmeasured over time)

Yes

nalyses suggestresults)

No

ted in countries withe economies,a

n English language

Studies conducted amongEnglish-speakingpopulations only

IncludedIncludedIncluded

vided to all inetting)

Included

inistered to allclass

Included

Included

Included

Included

assessed Earliest outcome assessedYes

Standardized meandifference

phic presentation,d in medians and

Restricted the magnitudeof outliers

43.5%

e. Co

Guids)

edf admts in

ivity arable

nducncomhed i

if proom sf admts in

come

ange

r graclude

ems (e.g., anxiety and depression). The effect sizes for


tsdwboa

stfasgtfibqtga

Waia

LCJJCCPHS

p

aDHm

o

R

A

hree outcomes—association with antisocial peers, sub-tance abuse, and family adjustment—are not statisticallyifferent from zero. The studies included in their reviewere selected because they addressed violent or aggressiveehavior rather than these other outcomes; their findingsn these questions, therefore, may not represent allvailable evidence on these topics.

Schools present a critical opportunity for changingocietal behavior. Almost the entire population passeshrough these institutions, starting during an early andormative period and continuing over many years. Withpproximately 71 million children in U.S. primary andecondary schools in 2003 and an overall high schoolraduation rate of 85%,63 it is difficult to overestimatehis opportunity to effect change. The potential bene-ts of improved school function alone are notable. Theroader and longer-term benefits in reduced delin-uency and anti-social behavior are yet more substan-ial. Universal school-based violence prevention pro-rams are an important means of reducing violent andggressive behavior in our society.

e are grateful for the collaboration of Sandra Wilson, PhD,nd Mark Lipsey, PhD, (Vanderbilt University, Nashville TN)n the final stages of both converging projects for sharing datand comparing results.

We would also like to thank our consultants: DanielleaRaque, MD, Department of Pediatrics, Harlem Hospitalenter, New York; Colin Loftin, PhD, School of Criminal

ustice, University at Albany, State University of New York;ames Mercy, PhD, Division of Violence Prevention, Nationalenter for Injury Prevention and Control, Centers for Diseaseontrol and Prevention, Atlanta GA; Patricia Smith, Violencerevention Section, Michigan Department of Communityealth, Lansing, MI; and Suzanne Salzinger, PhD, New York

tate Psychiatric Institute, New York.The Oak Ridge Institute for Scientific Education (ORISE)

rovided funding for the work of DF-W, JL, and HW.The findings and conclusions in this report are those of the

uthors and do not necessarily reflect those of the Centers forisease Control and Prevention, the National Institutes ofealth, the National Institute of Justice, or the U.S. Depart-ent of Justice.No financial conflict of interest was reported by the authors

f this paper.

eferences1. Maguire K, Pastore AL. Sourcebook of criminal justice statistics 2003. In:

Pastore AL, ed. Washington DC; U.S. Department of Justice, Bureau ofJustice Statistics: 2005.

2. U.S. Census Bureau. Monthly postcensal resident population, by singleyear of age, sex, race, and Hispanic origin (1/1/2003-6/1/2003). Wash-ington DC: U.S. Census Bureau, 2006.

3. Snyder HN, Sickmund M. Juvenile offenders and victims: 1999 nationalreport. Washington DC: Office of Juvenile Justice and DelinquencyPrevention, 1999.

4. National Center for Health Statistics. Health, United States, 2005. Hyatts-ville MD: U.S. Department of Health and Human Services, 2005.

5. Wasserman G, Keenan K, Tremblay R, et al. Risk and protective factors ofchild delinquency. Washington DC: U.S. Department of Justice, 2003.

ugust 2007

6. Office of Juvenile Justice and Delinquency Prevention. Co-occurrence ofdelinquency and other problem behaviors. Washington DC: U.S. Depart-ment of Justice, 2000.

7. Achenbach TM, Edelbrock C. Manual for the child behavior checklist andrevised child behavior profile. Burlington: Department of Psychiatry,University of Vermont, 1983.

8. Patterson G, Dishion T, Yoerger K. Adolescent growth in new forms ofproblem behavior: macro and micro peer dynamics. Prev Sci 2000;1:3–13.

9. Office of Juvenile Justice and Delinquency Prevention. Serious and violentjuvenile offenders. Washington DC: U.S. Department of Justice, Office ofJustice Programs, 1998.

10. U.S. Department of Health and Human Services. Youth violence: a reportof the Surgeon General. Washington DC: U.S. Department of Health andHuman Services, 2001.

11. Pastor P, Makuc DM, Reuben C, Xia H. Health, United States, 2002, withchartbook on trends in the health of Americans. Hyattsville MD: NationalCenter for Health Statistics, 2002.

12. Krug EG, Dahlberg LL, Mercy JA, Zwi AB, Lozano R. World report onviolence and health. Geneva: World Health Organization, 2002.

13. DeVoe J, Peter K, Noonan M, Snyder T, Baum K. Indicators of schoolcrime and safety: 2005. Washington DC: U.S. Department of Educationand U.S. Department of Justice, 2005.

14. Eaton DK, Kann L, Kinchen S, et al. Youth risk behavior surveillance-United States, 2005. MMWR Surveill Summ 2006;55(SS-5):1–108.

15. Grunbaum J, Kann L, Kinchen S, et al. Youth risk behavior surveillance,United States 2003. MMWR Surveill Summ 2004;53(SS-2):1–96.

16. Centers for Disease Control and Prevention. School Health Policy andPrograms Study. Available at: www.cdc.gov/healthYouth/shpps/factsheets/pdf/violence.pdf.

17. Dodge K, Dishion TJ, Lansford J. Deviant peer influences in interventionand public policy for youth. SRCD Social Policy Rep 2006;20:1–19.

18. Goleman D. Emotional intelligence. New York: Bantam Books, 1994.19. Glanz K, Rimer B, Lewis F, eds. Health behavior and health education:

theory, research, and practice. San Francisco: Jossey-Bass, 2002.20. Flay BR, Allred C. Long-term effects of the Positive Action program. Am J

Health Behav 2003;27(suppl 1):S6–21.21. Meyer A, Farrell A, Northup W, Kung E, Plybon L. Promoting nonviolence

in early adolescence: responding in peaceful and positive ways. New York:Kluwer Academic Publishers, 2000.

22. Kelder SH, Orpinas P, McAlister A, Frankowski R, Parcel GS, Friday J. TheStudents for Peace Project: a comprehensive violence-prevention programfor middle school students. Am J Prev Med 1996;12:22–30.

23. Bandura A. Social foundations of thought and action: a social cognitivetheory. Englewood Cliffs, NJ: Prentice-Hall, 1986.

24. Frey KS, Hirschstein MK, Guzzo BA. Second Step: preventing aggressionby promoting social competence. J Emotion Behav Disord 2000;8:102–12.

25. Grossman DC, Neckerman HJ, Koepsell TD, et al. Effectiveness of aviolence prevention curriculum among children in elementary school: arandomized controlled trial. JAMA 1997;277:1605–11.

26. Shure MB, Spivack G. Interpersonal problem-solving in young children:a cognitive approach to prevention. Am J Community Psychol 1982;10:341–55.

27. Greenberg MT, Kusche C, Mihalic SF. Blueprints for violence prevention.Book 10. Promoting Alternative Thinking Strategies (PATHS). BoulderCO: Center for the Study and Prevention of Violence, 1998.

28. Embry D, Flannery DJ, Vazsonyi AT, Powell KE, Atha H. PeaceBuilders: atheoretically driven, school-based model for early violence prevention.Am J Prev Med 1996;12(suppl 5):91–100.

29. Task Force on Community Preventive Services. The Guide to CommunityPreventive Services. What works to promote health? Zaza S, Briss PA,Harris KW, eds. New York: Oxford University Press, 2005.

30. U.S. Department of Health and Human Services. Healthy people 2010.Washington DC: U.S. Department of Health and Human Services, 2000.

31. Chan LS, Kipke MD, Schneir A, et al. Prevening violence and relatedhealth-risking social behaviors in adolescents. Evidence report/technol-ogy assessment no. 107. Rockville MD: Agency for Healthcare Researchand Quality, October 2004. Available at: www.ahrq.gov/downloads/pub/evidence/pdf/adolviol/adolviol.pdf.

32. Hawkins JD, Catalano RF, Kosterman R, Abbott R, Hill KG. Preventingadolescent health-risk behaviors by strengthening protection during child-hood. Arch Pediatr Adolesc Med 1999;153:226–34.

33. Olweus D. Bully/victim problems in school: facts and intervention. Eur
J Psychol Educ 1997;12:495–510.

S

34. Conduct Problems Prevention Research Group. Initial impact of the fasttrack prevention trial for conduct problems. II. Classroom effects. JConsult Clin Psychol 1999;67:648-57.

35. Ialongo N, Werthamer L, Kellam SG. Proximal impact of two first-gradepreventive interventions on the early risk behaviors for later substanceabuse, depression and antisocial behavior. Am J Community Psychol1999;27:599–641.

36. Reid JB, Eddy JM, Fetrow RA, Stoolmiller M. Description and immediateimpacts of a preventive intervention for conduct problems. Am J Com-munity Psychol 1999;27:483–517.

37. Olweus D, Limber S, Mihalic SF. Bullying prevention program. BoulderCO: Center for the Study and Prevention of Violence, 1999.

38. Center for the Study and Prevention of Violence. Blueprints promisingprograms overview. Boulder CO: Center for the Study and Prevention ofViolence, 2002. Available at: www.colorado.edu/cspv/blueprints/promis-ing/overview.html.

39. Briss PA, Zaza S, Pappaioanou M, et al. Developing an evidence-basedGuide to Community Preventive Services-methods. Am J Prev Med 2000;18(suppl 1):35–43.

40. Zaza S, Wright-de Aguero L, Briss PA, et al. Data collection instrumentand procedure for systematic reviews in the Guide to Community Preven-tive Services. Am J Prev Med 2000;18(suppl 1):44–74.

41. American Psychiatric Association. Diagnostic and statistical manual ofmental disorders (DSM-IV). Washington DC: American Psychiatric Pub-lishing, 2002.

42. Hann DM, Borek N, eds. Taking stock of risk factors for child/youthexternalizing behavior problems. Bethesda MD: Department of Healthand Human Services, Public Health Service, National Institute of MentalHealth, 2001.

43. Carande-Kulis VG, Maciosek MV, Briss PA, et al. Methods for systematicreviews of economic evaluations for the Guide to Community PreventiveServices. Am J Prev Med 2000;18(suppl 1):75–91.

44. Community Guide Economic Review Team. Economic evaluation abstractionform, version 3.0. Atlanta GA, 2001. Available at: www.thecommunityguide.org/PDFS/ECONEVAL_version3.pdf.

45. Greenberg MT, Kusche C. Preventive intervention for school-aged deafchildren: the PATHS curriculum. J Deaf Stud Deaf Educ 1998;3:49–63.

46. Aber JL, Jones SM, Brown JL, Chaudry N, Samples F. Resolving conflictcreatively: evaluating the developmental effects of a school-based violenceprevention program in neighborhood and classroom context. Dev Psy-chopathol 1998;10:187–213.

47. Alsaker FD, Valkanover S. Early diagnosis and prevention of victimizationin kindergarten. In: Juvonen J, Graham S, eds. Peer harassment in school:the plight of the vulnerable and victimized. New York: Guilford Press,2001:175-95.

48. Battistich V, Schaps E, Watson M, Solomon D, Lewis C. Effects of the ChildDevelopment Project on students’ drug use and other problem behaviors.J Prim Prev 2000;21:75–99.

49. Bosworth K, Espelage D, DuBay T, Daytner G, Karageorge K. Preliminaryevaluation of a multimedia violence prevention program for adolescents.Am J Health Behav 2000;24:268–80.

50. Boulton MJ, Flemington I. The effects of a short video intervention onsecondary school pupils’ involvement in definitions of and attitudestowards bullying. Sch Psychol Int 1996;17:331–45.

51. Conduct Problems Prevention Research Group. Evaluation of the first 3years of the Fast Track prevention trial with children at high risk foradolescent conduct problems. J Abnormal Child Psychol 2002;30:19-35.

52. DuRant RH. Comparison of two violence prevention curricula for middleschool adolescents. J Adolesc Health 1996;19:111–7.

53. DuRant RH, Barkin S, Krowchuck DP. Evaluation of a peaceful conflictresolution and violence prevention curriculum for sixth-grade students. JAdolesc Health 2001;28:386–93.

54. Elias MJ, Gara MA, Schuyler T, Branden-Muller LR, Sayette MA. Thepromotion of social competence: longitudinal study of a preventiveschool-based program. Am J Orthopsychiatry 1991;61:409–17.

55. Esbensen FA, Osgood DW, Taylor TJ, Peterson D, Freng A. How great isG. R. E. A. T.? Results from a longitudinal quasi-experimental design.Criminol Public Policy 2001;1:87–118.

56. Farrell AD, Meyer AL. The effectiveness of a school-based curriculum forreducing violence among urban sixth-grade students. Am J Public Health1997;87:979–84.

57. Farrell AD, Meyer AL, White KS. Evaluation of Responding in Peacefuland Positive Ways (RIPP): a school-based prevention program for reduc-
ing violence among urban adolescents. J Clin Child Psychol2001;30:451–63.

58. Farrell AD, Valois RF, Meyer AL. Evaluation of the RIPP-6 violenceprevention program at a rural middle school. Am J Health Educ2002;33:167–72.

59. Farrell AD, Valois RF, Meyer AL, Tidwell R.P. Impact of the RIPP violenceprevention program on rural middle school students. J Prim Prev2003;24:143–67.

60. Flay BR, Graumlich S, Segawa E, Burns JL, Holliday MY. Effects of 2prevention programs on high-risk behaviors among African-Americanyouth: a randomized trial. Arch Pediatr Adolesc Med 2004;158:377–84.

61. Sandy SV, Boardman SK. The peaceful kids conflict resolution program.Int J Conflict Manag 2000;11:337–57.

62. Sawyer MG, MacMullin C, Graetz B, Said JA, Clark JJ, Baghurst P. Socialskills training for primary school children: a one-year follow-up study. JPaediatr Child Health 1997;33:378–83.

63. U.S. Census Bureau. Statistical Abstract of the United States, 2003. TheNational data book. Washington DC: U.S. Department of Commerce,2004.

64. Flannery DJ, Liau AK, Powell KE, et al. Initial behavior outcomes for thePeaceBuilders universal school-based violence prevention program. DevPsychol 2003;39:292–308.

65. Flay BR, Allred CG, Ordway N. Effects of the Positive Action program onachievement and discipline: two matched-control comparisons. Prev Sci2001;2:71–89.

66. Foshee VA, Bauman KE, Greene WF, Koch GG, Linder GF, MacDougallJE. The Safe Dates program: 1-year follow-up results. Am J Public Health2000;90:1619–22.

67. Gesten EL, Rains MH, Rapkin BD, et al. Training children in socialproblem-solving competencies: a first and second look. Am J CommunityPsychol 1982;10:95–115.

68. Hindley P, Reed H. Promoting alternative thinking strategies (PATHS):mental health promotion with deaf children in school. In: Decker S, KirbyS, Greenwood A, Moore D, eds. Taking children seriously: applications ofcounseling and therapy in education. London: Continuum InternationalPublishing Group, 1999:113-30.

69. Johnson DW, Johnson RT, Dudley BS. Effects of peer mediation trainingon elementary school students. Mediation Q 1992;10:89–99.

70. Johnson DW, Johnson RT, Dudley BS, Ward M, et al. The impact of peermediation training on the management of school and home conflicts. AmEduc Res J 1995;32:829–44.

71. Kenney DJ, Watson TS. Reducing fear in the schools: managing conflictthrough student problem solving. Educ Urban Soc 1996;28:436–55.

72. Lynch KB, Geller SR, Schmidt MG. Multi-year evaluation of the effective-ness of a resilience-based prevention program for young children. J PrimPrev 2004;24:335–53.

73. McMahon SD, Washburn JJ. Violence prevention: an evaluation of pro-gram effects with urban African-American students. J Prim Prev2003;24:43–62.

74. McMahon SD, Washburn J, Felix ED, Yakin J, Childrey G. Violenceprevention: program effects on urban preschool and kindergarten chil-dren. Appl Prev Psychol 2000;9:271–81.

75. Nelson G, Carson P. Evaluation of a social problem-solving skills programfor third- and fourth-grade students. Am J Community Psychol1988;16:79–99.

76. O’Donnell L, Stueve A, Doval AS, et al. Violence prevention and youngadolescents’ participation in community youth service. J Adolesc Health1999;24:28–37.

77. Orpinas P, Parcel GS, McAlister A, Frankowski R. Violence prevention inmiddle schools: a pilot evaluation. J Adolesc Health 1995;17:360–71.

78. Orpinas P, Kelder S, Frankowski R, Murray N, Zhang Q, McAlister A.Outcome evaluation of a multi-component violence-prevention programfor middle schools: the Students for Peace project. Health Educ Res2000;15:45–58.

79. Orpinas P, Horne A. School bullying: changing the problem by changingthe school. Sch Psychol Rev 2003;32:431–44.

80. Pepler DJ, Craig WM, Ziegler S, Charach A. An evaluation of ananti-bullying intervention in Toronto schools. Can J Commun MentHealth 1994;13:95–110.

81. Palumbo DJ, Ferguson JL. Evaluating Gang Resistance Education andTraining (GREAT): is the impact the same at that of Drug AbuseResistance Education (DARE)? Eval Rev 1995;19:597–619.

82. Renfro J, Huebner R, Ritchey B. School violence prevention: the effects ofa university and high school partnership. J Sch Violence 2003;2:81–99.

83. Rickel AU, Eshelman AK, Loigman GA. Social problem solving training: afollow-up study of cognitive and behavioral effects. J Abnormal Child
Psychol 1983;11:15–28.

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

A

84. Roland E. Bullying: a developing tradition of research and management.In: Tattum D, ed. Understanding and managing bullying. Oxford: Hei-nemann Education, 1993:15–30.

85. Roush G, Hall E. Teaching peaceful conflict resolution. Mediation Q1993;11:185–91.

86. Shapiro J, Burgoon JD, Welker CJ, Clough JB. Evaluation of The Peace-makers Program: school-based violence prevention for students in gradesfour through eight. Psychol Sch 2002;39:87–100.

87. Stevens V, De Bourdeaudhuij I, Van Oost P. Bullying in Flemish schools:an evaluation of anti-bullying intervention in primary and secondaryschools. Br J Educ Psychol 2000;70:195–210.

88. Taylor CA, Liang B, Tracy AJ, Williams LM, Seigle P. Gender differencesin middle school adjustment, physical fighting, and social skills: evaluationof a social competency program. J Prim Prev 2002;23:261–73.

89. Weissberg RP, Gesten EL, Carnrike CL, et al. Social problem-solving skillstraining: a competence building intervention with 2nd-4th grade chil-dren. Am J Community Psychol 1981;9:411–24.

90. Whitney I, Rivers I, Smith P, Sharp S. The Sheffield project: methodologyand findings. In: Smith P, Sharp S, eds.School bullying: insights andperspectives. London: Routledge, 1994;20–56.

91. Winer JI, Hilpert PL, Gesten EL, Cowen EL, Schubin WE. The evaluationof a kindergarten social problem solving program. J Prim Prev1982;2:205–16.

92. Olweus D. Bullying among schoolchildren: intervention and prevention.In: Peters RD, McMahon PM, Quinsey VL, eds. Aggression and violencethroughout the life span. Newbury Park CA: Sage, 1992:100-25.

93. Roland E. Bullying: The Scandinavian Research Tradition. In: Tattum D,Lane D, eds. Bullying in schools. Trentham: Stoke-on-Trent, 1989:21–32.

94. Shure MB. I can problem solve (ICPS): interpersonal cognitive problemsolving for young children. Early Child Dev Care 1993;96:49–64.

95. Aos S, Lieb R, Mayfield J, Miller M, Pennucci A. Benefits and costs ofprevention and early intervention programs for youth. Olympia: Wash-ington State Institute for Public Policy, 2004. Available at: www.wsipp.wa.gov/rptfiles/04-07-3901.pdf.

96. Vazsonyi A, Belliston L, Flannery D. Evaluation of a school-based, univer-sal violence prevention program: low-, medium-, and high-risk children.Youth Violence Juvenile Justice 2004;2:185–206.

97. Aber JL, Pedersen S, Brown JL, Jones SM, Gershoff ET. Changingchildren’s trajectories of development: two-year evidence for the effective-ness of a school-based approach to violence prevention. New York:Mailman School of Public Health, Columbia University: National Centerfor Children in Poverty, 2003.

98. Foshee VA, Bauman KE, Arriaga XB, Helms RW, Koch GG, Linder GF. Anevaluation of Safe Dates, an adolescent dating violence prevention pro-gram. Am J Public Health 1998;88:45–50.

99. Wilson S, Lipsey MJ, Derzon JH. The effects of school-based interventionprograms on aggressive behavior: a meta-analysis. J Consult Clin Psychol2003;71:136-49.

00. Mihalic S, Irwin K, Fagan A, Ballard D, Elliott D. Successful programimplementation: lessons from Blueprints. Available at: www.ncjrs.gov/pdffiles1/ojjdp/204273.pdf. 2004.

01. Wilson SJ, Lipsey MJ. Effectiveness of school-based intervention programson aggressive behavior: update of a meta-analysis. Am J Prev Med2007;33(suppl 2):S130–S143.

02. Roseberry LL. An applied experimental evaluation of conflict resolutioncurriculum and social skills development. Doctoral diss. Loyola Universityof Chicago, 1997.

03. Allen RJ. An investigatory study of the effects of a cognitive approach tointerpersonal problem solving on the behavior of emotionally upsetpsychosocially deprived preschool children. Doctoral diss. BrookingsInstitute, Union Graduate School, 1978.

04. Botzer EA. An evaluation of the effectiveness of the Second Step violenceprevention curriculum for third-grade students. Doctoral diss. StateUniversity of New York at Buffalo, 2003.

05. Cisek JV. The effects of teaching conflict management and interpersonalproblem solving skills to junior high school students. Doctoral diss.
Northern Illinois University, 1979.
ugust 2007

06. Dodsworth PR. Evaluation of a social problem-solving curriculum forpreventing early adolescent aggression. Doctoral diss. University of NewBrunswick, Canada, 2002.

07. Henderson P. The effects of affective education on nonacademic, aca-demic, and behavioral characteristics of fifth and sixth graders. Doctoraldiss. University of Cincinnati, 1982.

08. Hughes BG. The effects of a preventative intervention program on urbanand suburban middle school students in the areas of risk and resiliency.Doctoral diss. University of Wisconsin-Milwaukee, 2003.

09. Johnson JL. Preventing conduct problems and increasing social compe-tence in high-risk preschoolers. Doctoral diss. Regent University, 2001.

10. Kaiser-Ulrey CL. Bullying in middle school: A study of B.E.S.T.-BullyingEliminated from Schools Together-an anti-bullying program for seventh-grade students. Doctoral diss. Florida State University, 2004.

11. Korsh NB. Effects of interventions designed to improve cooperative socialinteraction and performance in small groups of third grade students inthe open classroom. Doctoral diss. Xerox University, 1977.

12. Matthews DB. The effects of reality therapy on reported self-concept,social adjustment, reading achievement, and discipline of fourth and fifthgraders in two elementary schools. Doctoral diss. University of SouthCarolina, 1973.

13. McCabe LA. Violence prevention in early childhood: Implementing theSecond Step curriculum in child care and Head Start classrooms. Doctoraldiss. Ithaca: Cornell University, 2000.

14. McNeese RMF. Reducing violent behavior in the classroom: A comparisonof two middle schools. Doctoral diss. Atlanta: Georgia State University,2000.

15. McPhee MA. The efficacy of interpersonal cognitive problem-solving(ICPS) and young children: prevention in a naturalistic preschool envi-ronment. Doctoral diss. Boston College, 1994.

16. Pride-Lawson TA. The effects of a classroom anger management programon levels of anger and depression in fourth-grade students. Doctoral diss.Norman: University of Oklahoma, 1998.

17. Catalano R, Harachi TW, Abbott RD, Haggerty KP, Fleming CB. Raisinghealthy children through enhancing social development in elementaryschool: results after 1.5 years. Seattle: University of Washington School ofSocial Work. Unpublished.

18. Gottfredson DC. An assessment of a delinquency prevention demonstra-tion with both individual and environmental interventions. BaltimoreMD: Center for Social Organization of Schools, Johns Hopkins University,1986 (research rept 361).

19. Manos MJ. Youth Development Project: Preventive intervention in delin-quency. Three year evalution report 1984-1987. Manoa: Center for YouthResearch, Hawaii University, 1988.

20. Weddle KD, Williams F. Implementing and assessing the effectiveness ofthe Interpersonal Cognitive Problem-Solving (ICPS) curriculum in fourexperimental and four control classrooms. Report to the Faculty SmallResearch Grant Program. Memphis TN: Memphis State University, 1993.

21. Hausman A, Pierce G, Briggs L. Evaluation of comprehensive violenceprevention education: effects on student behavior. J Adolesc Health1996;19:104–10.

22. Hunter L, Elias MJ, Norris J. School-based violence prevention: challengesand lessons learned from an action research project. J Sch Psychol2001;39:161–75.

23. Komro KA, Perry CL, Veblen-Mortenson S, et al. Violence-related out-comes of the D.A.R.E. Plus Project. Health Educ Behav 2004;31:335–54.

24. Metropolitan Area Research Group. A cognitive-ecological approach topreventing aggression in urban settings: initial outcomes for high-riskchildren. J Consult Clin Psychol 2002;70:179–94.

25. Twemlow SW, Fonagy P, Sacco FC, Gies ML, Evans R, Ewbank R. Creatinga peaceful school learning environment: a controlled study of an elemen-tary school intervention to reduce violence. Am J Psychiatry2001;158:808–10.

26. Serna L, Lamros K, Nielson E, Forness S. Head Start children at risk foremotional or behavioral disorders: behavior profiles and clinical implica-tions of a primary prevention program. Behav Disord 2002;27:137–41.

27. Serna L, Nielson E, Mattern N. Primary prevention in mental health for
Head Start classrooms: partial replication with teachers as intervenors.Behav Disord 2003;28:124–9.

http://www.ncjrs.gov/pdffiles1/ojjdp/204273.pdf

http://www.ncjrs.gov/pdffiles1/ojjdp/204273.pdf

Effectiveness of Universal School-Based Programs to Prevent Violent and Aggressive Behavior A...

Documents

Transcript of Effectiveness of Universal School-Based Programs to Prevent Violent and Aggressive Behavior A...