What do we know about preventing school violence? A systematic review of systematic reviews

Abstract

Many children across the world are exposed to school violence, which undermines their right to education and adversely affects their development. Studies of interventions for school violence suggest that it can be prevented. However, this evidence base is challenging to navigate. We completed a systematic review of interventions to reduce four types of school violence: (a) peer violence; (b) corporal punishment; (c) student-on-teacher violence and (d) teacher-on-student violence. Reviewers independently searched databases and journals. Included studies were published between 2005 and 2015; in English; considered school-based interventions for children and measured violence as an outcome. Many systematic reviews were found, thus we completed a systematic review of systematic reviews. Only systematic reviews on interventions for intimate partner violence (IPV) and peer aggression were found. These reviews were generally of moderate quality. Research on both types of violence was largely completed in North America. Only a handful of programmes demonstrate promise in preventing IPV. Cognitive behavioral, social-emotional and peer mentoring/mediation programmes showed promise in reducing the levels of perpetration of peer aggression. Further research needs to determine the long-term effects of interventions, potential moderators and mediators of program effects, program effects across different contexts and key intervention components.

Keywords:

• Systematic review
• school violence
• school-based
• peer aggression
• intimate partner violence

Introduction

School violence undermines children’s right to education and adversely affects their development. The long term consequences are also costly for broader society (Burton & Leoschut, 2013). Worryingly, children across the world report exposure to violence at school (Due, Holstein, & Soc, 2008).

Although bullying is a major focus of school violence research, violence in schools encompasses much more. Bullying is defined as repeated aggressive episodes where there is a power imbalance between the bully and his/her victim (Menesini & Salmivalli, in press). Bullying is thus a subset of peer violence, a broader group of behaviors that include ‘the intentional use of physical force or power, threatened or actual, …. that either results in or has a high likelihood of resulting in injury, death, psychological harm, maldevelopment, or deprivation’ (Krug, Dahlberg, Mercy, Zwi, & Lozano, 2002, p. 4). School violence thus includes any violence between students, corporal punishment of students by teachers (Burton & Leoschut, 2013), other forms of violence directed at students by teachers such as verbal aggression or rape (Lee, 2015), and violence directed by students at teachers (Dzuka & Dalbert, 2007; Wilson, Douglas, & Lyon, 2011). Furthermore, school violence is specifically defined as violence occurring on school premises, while traveling to or from school, or during a school-sponsored event (http://www.cdc.gov/violenceprevention/youthviolence/schoolviolence/).

A number of interventions have been tested for their potential to prevent school violence. These may be universal (all students participate; Mrazek & Haggerty, 1994). Others may target students at increased risk for violence or those already demonstrating violent behaviors, known respectively as selected and indicated interventions (Mrazek & Haggerty, 1994). Additionally, interventions using a whole-school approach intervene at multiple levels within a school (Gevers & Flisher, 2012), whereas discrete interventions work only with a particular aspect of the school, for example just the students (Gevers & Flisher, 2012). Comprehensive programmes address a range of risk behaviors, whereas specific programmes address a particular problem (Gevers & Flisher, 2012). Such complexity can make it challenging to determine exactly which interventions are the most effective for different types of school violence.

A number of reviews of school violence interventions have synthesized the literature and so addressed a variety of these issues; thus, following Mikton and Butchart’s (2009) approach to understanding interventions to prevent child maltreatment, we aimed to complete a systematic review of systematic reviews that addressed the question: What do we know about preventing school violence?

Methods

Search strategy

Pairs of research assistants each independently searched 49 electronic databases, 3 clinical trial registries and 10 online journals for articles on school violence (see Appendix A). Searches were limited to papers in English and in publication years 2005–2015, except for those addressing corporal punishment. Two searches of abstracts were conducted. The first used search terms: school AND (violen* OR aggress* OR bully* OR bulli*), while the second used the search terms school AND ‘corporal punishment’. Literature on corporal punishment was sought from 1980 to 2015, because of the small body of work completed on this type of violence in schools (there is a large body of work on parental corporal punishment; Gershoff & Grogan-Kaylor, 2016). Experts in the field who were part of the kNOw Violence in Childhood Project School’s Learning Group were also consulted about relevant studies.

Studies were considered relevant if they:

(1)	Were in English;
(2)	Included change in violent behavior or one of its synonyms (such as aggression, externalizing behavior/problems, conduct behavior/problems or intimate partner violence [IPV]) as an outcome;
(3)	Addressed an intervention for violent behavior that was implemented at, or recruited participants from, school; and
(4)	Included pre-primary, primary or secondary school students.

We focused on change in behavior because changes in knowledge and attitudes alone are not sufficient to change behavior (De La Rue, Polanin, Espelage, & Pigott, 2014; Whitaker, Murphy, Eckhardt, Hodges, & Cowart, 2013). In addition, articles with (a) suicide, (b) school shootings and (c) teacher-on-teacher violence as an outcome were excluded. Information and communication technology interventions (which relate more to cyberbullying), psychopharmacological interventions, and interventions which extended across multiple domains like multisystemic therapy (Henggeler, Melton, Brondino, Scherer, & Hanley, 1997), were also excluded.

The initial search identified a large number of relevant systematic reviews, and we therefore decided to do a systematic review of systematic reviews, rather than a systematic review of primary studies (Mikton & Butchart, 2009). Research assistants then screened the full text of reviews to determine whether they met an additional inclusion criterion: the review included at least three primary studies about interventions which were implemented at school or recruited participants from school (see Appendix B and C respectively, for included and excluded reviews).

Data extraction

The quality of the relevant reviews was assessed, and descriptive information captured (see Appendix D for extraction document).

We used the AMSTAR tool to assess methodological quality of each review (Shea et al., 2009). AMSTAR scores between 0 and 4 indicate that a review is of poor quality, scores between 5 and 8 indicate moderate quality, and scores of 9–11 indicate high quality (Mikton & Butchart, 2009). A second reviewer checked 42% of the AMSTAR scores. An intra-class correlation coefficient of above .80 was achieved, indicating a good level of coding consistency (Aspland & Gardner, 2003).

Results

Our initial screening identified over 400 systematic reviews. A second round of screening found 36 that were eligible for inclusion (see Figure 1). These only addressed interventions for IPV and peer aggression.

Figure 1. PRISMA flowchart.

A small number of narrative reviews and primary studies (which were excluded) were identified on student-on-teacher violence, teacher-on-student violence and corporal punishment in schools.

IPV

Five reviews of interventions for IPV were identified. On average, these were of moderate quality (see Table 1).

Table 1. Quality of reviews on intimate partner violence.

Review	AMSTAR score	No. of studies included in review	No. of school-based interventions included	No. of school-based interventions studied in RCTs
DeGue et al. (2014)	4	35	6	6
De Koker et al. (2014)	6	8	8	8
De La Rue et al. (2014)	10	23	23	10
Whitaker et al. (2006)	6	11	10	4
Whitaker et al. (2013)	4	9	4	4

Descriptions of programmes to prevent IPV

Since a number of school-based IPV prevention programmes have been studied using randomised controlled trials (RCTs) – the strongest evaluation design – we report only on these 11 programmes (see Table 2).

Table 2. Intimate partner violence programmes assessed in RCTs with behavioral outcomes.

Programme	Target population, type of program and country of study	Implementer	No. of sessions and duration	Delivery mechanism
1. Dating Violence Prevention Program (Avery-leaf, Cascardi, O’Leary, & Cano, 1997)	11th and 12th grade students	Teachers	1 week	Psychoeducation on ‘courtship’ aggression
	Universal, discrete, specific
	Study conducted in New York
2. Safe Dates (Foshee et al., 1998, 2005, 2000, 1996)	8th and 9th grade students	Teachers	10 45-min sessions	Lecture, poster contest, peer theatre production;
	Universal, whole school, specific			Also includes a community component (crisis line, support groups, material for parents, training of service providers)
	Study conducted in North Carolina
3. Safe Dates with booster (Foshee et al., 2004)	This is a trial within the original trial, provided to randomly selected participants after the 2-year follow-up	Health educator	–	Newsletter containing information drawing on the Safe Dates curriculum; personal telephone call
	Universal, whole school, specific
4. Ending Violence (Jaycox et al., 2006)	9th grade students	Attorneys	3 days	Lecture and discussion of legal issues
	Universal, discrete, specific
	Study conducted in California
5. Stepping Stones (Jewkes et al., 2008)	Secondary school students	Project staff	50  for 6–8 weeks outside of school hours	Discussion, role-plays, drama
	Universal, discrete, comprehensive
	Study conducted in rural South Africa
6. Fourth R: Skills for Youth Relationships (Wolfe et al., 2009)	9th grade students	Teacher	21 sessions over 7 weeks	Videos, handouts, role-play
	Universal, discrete, comprehensive
	Study conducted in Canada
7. Law and Justice Curriculum (Taylor et al., 2010a, 2010b)	6th and 7th grade students	–	5 sessions	Knowledge-based curriculum
	Universal, discrete, specific
	Study conducted in Ohio
8. Interaction-based Treatment (Taylor et al., 2010a, 2010b)	6th and 7th grade students	–	5 sessions	Identifying unwanted behavior, setting boundaries, bystander intervention
	Universal, discrete, specific
	Study conducted in Ohio
9. Shifting Boundaries classroom-level (Taylor, Stein, Woods, & Mumford, 2011; Taylor et al., 2013)	6th and 7th grade students	Teachers	8 weeks	Lecture and discussion about identifying unwanted behavior and setting boundaries
	Universal, discrete, specific
	Study conducted in New York
10. Shifting Boundaries school-level (Taylor et al., 2011, 2013)	6th and 7th grade students	–	8 weeks	‘Building-based restraining orders’; school violence protocols with emphasis on reporting to teachers; awareness posters; student-created ‘hotspot’ map
	Universal, whole-school, comprehensive
	Study conducted in New York
11. Coaching Boys Into Men (Miller et al., 2013, 2012)	Male middle-school students in sports teams	Sports coaches	11 brief discussions (10–15 min) during sports season	Discussion
	Universal, discrete, specific
	Study conducted in the USA

All programmes were universal and largely specific to IPV, and barring two (the building-based version of Shifting Boundaries, which targeted the whole school; Taylor, Stein, Mumford, & Woods, 2013; and the Safe Dates poster and theatre elements; Foshee et al., 2005) were discrete. Interventions were aimed at high school students of both genders, with the lone exception of Coaching Boys into Men, which focused only on boys (Miller et al., 2013).

All but one of the primary studies included in the reviews were completed on the North American continent (10 studies), and largely in the USA. One study by Jewkes et al. (2008) was conducted in the African region (South Africa), and none in any other region. Yet rates of IPV are highest in Africa, the Eastern Mediterranean and South East Asia, followed by the Americas (Stöckl, Devries, & Watts, 2015). Most programmes have thus been tested in contexts that need them least.

Evidence for programmes to prevent IPV

Safe Dates (Foshee et al., 2005), the Fourth R (Wolfe et al., 2009), Stepping Stones (Jewkes et al., 2008) and the building-level version of Shifting Boundaries (Taylor et al., 2013) stand out as the only programmes that achieved positive effects (see Table 3). Teachers, project staff and health educators implemented these programmes. The duration of the latter three programmes seemed to average around 7 weeks. However, number of sessions ranged from 10 to 21. Safe Dates (Foshee et al., 2005), the Fourth R (Wolfe et al., 2009) and Stepping Stones (Jewkes et al., 2008) are also conspicuous as having been studied in trials with the strongest methods for determining evidence of effect in that they have the longest follow-up periods (3, 2.5 and 2 years, respectively). The Safe Dates trial was also strong in that it measured the widest range of forms of dating violence, and was able to show that effects for several forms of violence persisted over time (Foshee et al., 1998, 2004, 2005, 2000, 1996). Two programmes – the Law and Justice Curriculum (Taylor, Stein, & Burden, 2010a) and Interaction-Based Treatment (Taylor et al., 2010a) – were identified as possibly doing harm, in that they led to increased reporting of perpetration.

Table 3. Effectiveness of intimate partner violence prevention programmes.

Programme	Follow-up period^a	Types of perpetration assessed	Types of victimisation assessed	Perpetration prevention	Victimisation prevention
1. Dating Violence Prevention Program (Avery-leaf et al., 1997)	Immediate post-test	Moderate physical aggression	–	No effect	–
2. Safe Dates (Foshee et al., 1998, 2005, 2000, 1996)	3 years	Moderate & severe physical; sexual; psychological violence	Moderate & severe physical; sexual; psychological violence	Effective at 2 years for all outcomes; at 3 years only for psychological and severe physical perpetration	Effective at 2 years for moderate physical and sexual victimisation; no effect at 3 years
3. Safe Dates with booster (tested against Safe Dates; Foshee et al., 2004)	3 years	Moderate & severe physical; sexual; psychological violence	Moderate & severe physical; sexual; psychological violence	Significant effect only for psychological perpetration	No effect
4. Ending Violence (Jaycox et al., 2006)	6 months	IPV	IPV	No effect	No effect
5. Stepping Stones (Jewkes et al., 2008)	1 year	Physical and sexual IPV, males	Physical and sexual IPV, females	Effective for perpetration by men at 2-year follow-up	No effect
6. Fourth R: Skills for Youth Relationships (Wolfe et al., 2009)	2.5 years	Moderate physical perpetration	–	Effective	–
7. Law and Justice Curriculum (Taylor et al., 2010a, 2010b)	6 months	Sexual violence	–	Increased reporting (possibly because of increased awareness)	–
8. Interaction-based Treatment (Taylor et al., 2010a, 2010b)	6 months	Sexual violence with dating partner	Sexual violence with dating partner	Increased reporting (possibly because of increased awareness)	No effect
9. Shifting Boundaries classroom-level (Taylor et al., 2011, 2013)	6 months	Prevalence and frequency of IPV	Prevalence and frequency of IPV	No effect	No effect
10. Shifting Boundaries school-level (Taylor et al., 2011, 2013)	6 months	Prevalence and frequency of IPV	Prevalence and frequency of IPV	No effect	Reductions in prevalence and frequency
11. Coaching Boys Into Men (Miller et al., 2013, 2012)	1 year	IPV perpetration	–	No effect	–

^a We report only the results from the longest possible follow-up period.

No program had been studied in more than one RCT, and so the evidence for any program can at best only be considered promising by two of the current standards for prevention science: Blueprints for Healthy Youth Development (http://www.blueprintsprograms.com), and those of the Society for Prevention Research (Gottfredson et al., 2015). Many of the trials reviewed also had some risk of bias (De Koker, Mathews, Zuch, Bastien, & Mason-Jones, 2014; Whitaker et al., 2006).

Moderation effects are also key in understanding programmes (Gottfredson et al., 2015): Safe Dates has produced evidence that there is no difference in effectiveness by gender, by white vs. other ethnicity, or by whether students had previous experience of dating violence; but the trial of the Fourth R showed that the effect was present only for boys (Whitaker et al., 2013).

Safe Dates thus appears to be the most effective school-based program for preventing dating violence, but the evidence base in general needs much more development.

Peer aggression

We identified a total of 31 reviews addressing effectiveness of interventions to prevent peer aggression. AMSTAR scores (see Table 4) had a mean of 6, indicating that on average the reviews were of moderate quality. Nearly 40% (387) of the primary studies on school-based interventions evaluated the interventions in RCTs, and 213 (22%) utilized quasi-experimental designs. However, many reviews did not provide information on study design.

Table 4. Quality of reviews on peer aggression.

Review	AMSTAR score	No. of studies included in review	No. of studies with school-based intervention and effects for violence^a	No. of school-based studies using randomised controlled trials	No. of school-based studies using quasi-experimental designs
Allen-Meares, Montgomery, and Kim (2013)	3	18	3	1	2
Barnes, Smith, and Miller (2014)	4	25	20	13	7
Blank et al. (2010)	3	37	6	4	–
Bond, Woods, Humphrey, Symes, and Green (2013)	6	38	5	0	5
Bonell, Wells, et al. (2013)	7	10	4	3	1
Durlak, Weissberg, Dymnicki, Taylor, and Schellinger (2011)	5	213	112	–	–
Dymnicki, Weissberg, and Henry (2011)	4	26	26	–	–
Fagan and Catalano (2013)	4	18	9	7	2
Farahmand, Grant, Polo, Duffy, and DuBois (2011)	8	21	5	5	0
Gansle (2005)^b	4	27	22	–	–
Gavine, Donnelly, and Williams (2016)	7	16	12	7	5
Hahn et al. (2007)	7	65	65	–	14
Hale, Fitzgerald-Yau, and Mark Viner (2014)	6	50	8	8	–
Leff, Waasdorp, and Crick (2010)	4	10	9	7	2
Limbos et al. (2007)	5	41	22	–	–
Moestue, Moestue, and Muggah (2013)	5	18	4	3	1
Mytton, DiGuiseppi, Gough, Taylor, and Logan (2006)	8	51	34	34	0
Oliver, Reschly, and Wehby (2011)	4	12	4	4	0
Park-Higgerson, Perumean-Chaney, Bartolucci, Grimley, and Singh (2008)	5	26	26	26	0
Reddy, Newman, De Thomas, and Chun (2009)	9	29	22	4	18
Reese, Prout, Zirkelback, and Anderson (2010)^c	4	188	59	–	–
Sancassiani et al. (2015)	8	22	3	3	0
Schindler et al. (2015)	6	31	31	–	–
Sklad et al. (2012)	6	75	35	–	–
Stoltz et al. (2012)	6	24	24	18	6
Tolan et al. (2013)	9	46	3	–	–
Vidrine (n.d.)	6	10	10	8	2
Vreeman and Carroll (2007)	4	26	11	2	9
Wilson and Lipsey (2006a)	9	47	47	40	7
Wilson and Lipsey (2006b)	9	73	73	32	41
Wilson and Lipsey (2007)	9	399	249	158	91
Total	-	1692	963	387	213
Mean	5.93
Percentage^d				39.77%	21.89%

^a If the number of studies utilizing a randomised controlled trial design and quasi-experimental design do not equal the number of studies on school-based interventions for the same reviews, this study design information was not specified.

^b Figures based on the number of comparisons instead of the number of studies.

^c Figures based on the number of outcome measures instead of studies.

^d Total of column/total number of studies with school-based interventions.

Descriptions of programmes to prevent peer aggression

Universal interventions were much more commonly included in the reviews than selected and indicated interventions, as were discrete rather than multi-level or whole-school interventions (see Table 5). There were also more specific than comprehensive programmes. Nearly half of all the interventions targeted children of primary school age. Interventions were also generally delivered to both genders.

Table 5. Characteristics of school-based programmes with effects on peer aggression.

Review	Prevention target (n; %)^a–c	Intervention approach (n; %)^a^,^b^,^d	Intervention content (n; %)^a^,^b^,^e	School level (n; %)^a^,^b^,^f	Participant gender (n; %)^a^,^b^,^g
Allen-Meares et al. (2013)	U (2; 67%)	W (0)	C (0)	PP (0)	M (0)
	S (1; 33%)	M(0)	S (0)	P (2; 67%)	F (0)
	I (0)	D (0)	NS (3; 100%)	H (0)	B (3; 100%)
	U & S (0)	NS (3; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (1; 33%)
	S & I (0)			NS (0)
	NS (0)
Barnes et al. (2014)	U (14; 70%)	W (0)	C (0)	PP (0)	M (1; 5%)
	S (5; 25%)	M (0)	S (11; 55%)	P (19; 95%)	F (0)
	I (1; 5%)	D (20; 100%)	NS (9; 45%)	H (0)	B (19; 95%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (1; 5%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Blank et al. (2010)	U (6; 100%)	W (6; 100%)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (1; 17%)	P (0)	F (0)
	I (0)	D (0)	NS (5; 83%)	H (1; 17%)	B (5; 83%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (1; 17%)
	U & I (0)			C – P & H (2; 33%)
	S & I (0)			C – (0)
	NS (0)			NS (3; 50%)
Bond et al. (2013)	U (0)	W (0)	C (3; 60%)	PP (0)	M (0)
	S (0)	M (3; 60%)	S (2; 40%)	P (3; 60%)	F (0)
	I (5; 100%)	D (2; 40%)	NS (0)	H (0)	B (5; 100%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (1; 20%)
	S & I (0)			C – (0)
	NS (0)			NS (1; 20%)
Bonell, Wells, et al. (2013)	U (4; 100%)	W (1; 25%)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (3; 89%)	P (2; 50%)	F (0)
	I (0)	D (3; 75%)	NS (1; 11%)	H (0)	B (4; 100%)
	U & S (0)	NS (0)		C – PP & P (1; 25%)	NS (0)
	U & I (0)
	S & I (0)			C – P & H (1; 25%)
	NS (0)			C – (0)
				NS (0)
Durlak et al. (2011)	U (112; 100%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (112; 100%)	H (0)	B (0)
	U & S (0)	NS (112; 100%)		C – PP & P (0)	NS (112; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (112; 100%)
Dymnicki et al. (2011)	U (26; 100%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (26; 100%)	H (0)	B (0)
	U & S (0)	NS (26; 100%)		C – PP & P (26; 100%)	NS (26; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Fagan and Catalano (2013)	U (2; 22%)	W (0)	C (0)	PP (2; 22%)	M (1; 11%)
	S (3; 33%)	M (7; 78%)	S (2; 22%)	P (6; 67%)	F (0)
	I (1; 12%)	D (2; 22%)	NS (7; 78%)	H (0)	B (8; 89%)
	U & S (3; 33%)	NS (0)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (1; 11%)
	S & I (0)			C –(0)
	NS (0)			NS (0)
Farahmand et al. (2011)	U (2; 40%)	W (0)	C (0)	PP (0)	M (0)
	S (3; 60%)	M (0)	S (5; 100%)	P (4; 80%)	F (0)
	I (0)	D (0)	NS (0)	H (1; 20%)	B (5; 100%)
	U & S (0)	NS (5; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Gansle (2005)	U (0)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (22; 100%)	H (3; 15%)	B (0)
	U & S (0)	NS (22; 100%)		C – PP & P (7; 35%)	NS (20; 100%)
	U & I (0)			C – P & H (10; 50%)
	S & I (0)			C – (0)
	NS (22; 100%)			NS (0)
Gavine et al. (2016)	U (12; 100%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (5; 42%)	S (0)	P (5; 42%)	F (0)
	I (0)	D (7; 58%)	NS (12; 100%)	H (2; 36%)	B (0)
	U & S (0)	NS (0)		C – PP & P (0)	NS (21; 100%)
	U & I (0)			C – P & H (5; 42%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Hale et al. (2014)	U (7; 78%)	W (2; 43%)	C (8; 100%)	PP (0)	M (0)
	S (1; 12%)	M (3;37)	S (0)	P (7; 87%)	F (0)
	I (0)	D (3; 37%)	NS (0)	H (0)	B (8; 100%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (1; 13%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Hahn et al. (2007)	U (65; 100%)	W (1; 2%)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (34; 52%)	F (0)
	I (0)	D (0)	NS (65; 100%)	H (4; 6%)	B (0)
	U & S (0)	NS (64; 98%)		C – PP & P (6; 9%)	NS (65; 100%)
	U & I (0)			C – P & H (21; 33%)
	S & I (0)			C – (0)
	NS (0)			NS (53; 100%)
Leff et al. (2010)	U (6; 67%)	W (1; 11%)	C (0)	PP (1; 11%)	M (6; 67%)
	S (0)	M (2; 22%)	S (9; 100%)	P (5; 56%)	F (3; 33%)
	I (3; 33%)	D (6; 67%)	NS (0)	H (0)	B (0)
	U & S (0)	NS (0)		C – PP & P (2; 22%)	NS (0)
	U & I (0)			C – P & H (1; 11%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Limbos et al. (2007)	U (17; 77%)	W (0)	C (0)	PP (0)	M (0)
	S (5; 23%)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (22; 100%)	H (2; 9%)	B (0)
	U & S (0)	NS (22; 100%)		C – PP & P (3; 14%)	NS (22; 100%)
	U & I(0)			C – P & H (16; 73%)
	S & I (0)			C – (1; 4%)
	NS (0)			NS (0)
Moestue et al. (2013)	U (3; 75%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (1; 25%)	S (3; 75%)	P (1; 25%)	F (0)
	I (1; 25%)	D (3; 75%)	NS (1; 25%)	H (0)	B (4; 100%)
	U & S (0)	NS (0)		C – PP & P (1; 25%)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (2; 50%)
Mytton et al. (2006)h	U (0)	W (0)	C (0)	PP (0)	M (12; 35%)
	S (0)	M (0)	S (0)	P (22; 65%)	F (0)
	I (0)	D (0)	NS (34; 100%)	H (0)	B (22; 65%)
	U & S (0)	NS (34; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (12; 35%)
	S & I (34; 100%)			C – (0)
	NS (0)			NS (0)
Oliver et al. (2011)	U (4; 100%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (1; 25%)	S (4; 100%)	P (4; 100%)	F (0)
	I (0)	D (3; 75%)	NS (0)	H (0)	B (3; 75%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (1; 25%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Park-Higgerson et al. (2008)	U (7; 27%)	W (0)	C (7; 27%)	PP (0)	M (3; 11%)
	S (17; 65%)	M (10; 39%)	S (19; 73%)	P (19; 73%)	F (0)
	I (1; 4%)	D (16; 61%)	NS (0)	H (3; 11%)	B (23; 89%)
	U & S (1; 4%)	NS (0)		C – PP & P (2; 8%)	NS (0)
	U & I (0)			C – P & H (2; 8%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Reese et al. (2010)	U (0)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (59; 100%)	H (0)	B (0)
	U & S (0)	NS (59; 100%)		C – PP & P (0)	NS (59; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (59; 100%)			NS (59; 100%)
Reddy et al. (2009)	U (0)	W (0)	C (0)	PP (0)	M (0)
	S (8; 36%)	M (0)	S (0)	P (0)	F (0)
	I (14; 64%)	D (0)	NS (22; 100%)	H (0)	B (0)
	U & S (0)	NS (22; 100%)		C – PP & P (0)	NS (22; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (22; 100%)
Sancassiani et al. (2015)	U (3; 100%)	W (3; 100%)	C (2; 67%)	PP (0)	M (3; 100%)
	S (0)	M (0)	S (1; 33%)	P (2; 67%)	F (0)
	I (0)	D (0)	NS (0)	H (1; 33%)	B (0)
	U & S (0)	NS (0)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Schindler et al. (2015)	U (0)	W (0)	C (0)	PP (31; 100%)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (31; 100%)	H (0)	B (0)
	U & S (0)	NS (31; 100%)		C – PP & P (0)	NS (31; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (31; 100%)			NS (0)
Sklad et al. (2012)	U (35; 100%)	W (0)	C (0)	PP (0)	M (0)
	S (0)	M (0)	S (0)	P (0)	F (0)
	I (0)	D (0)	NS (35; 100%)	H (0)	B (0)
	U & S (0)	NS (35; 100%)		C – PP & P (0)	NS (35; 100%)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (35; 100%)
Stoltz et al. (2012)	U (0)	W (0)	C (0)	PP (2; 8%)	M (6; 25%)
	S (0)	M (13; 54%)	S (0)	P (15; 63%)	F (0)
	I (24; 100%)	D (11; 46%)	NS (24; 100%)	H (0)	B (18; 25%)
	U & S (0)	NS (0)		C – PP & P (5; 21%)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (2; 8%)
Tolan et al. (2013)	U (0)	W (0)	C (0)	PP (0)	M (0)
	S (2; 67%)	M (2; 67%)	S (0)	P (1; 33.33%)	F (0)
	I (1; 33%)	D (1; 33%)	NS (3; 100%)	H (1; 33.33%)	B (2; 67%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (1; 33%)
	U & I (0)			C – P & H (1; 33.33%)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Vidrine (n.d.)	U (0)	W (0)	C (0)	PP (6; 60%)	M (0)
	S (0)	M (0)	S (10; 100%)	P (4; 40%)	F (0)
	I (0)	D (0)	NS (0)	H (0)	B (10; 100%)
	U & S (0)	NS (10; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (10; 100%)			NS (0)
Vreeman and Carroll (2007)	U (5; 46%)	W (3; 27%)	C (2; 18%)	PP (0)	M (0)
	S (3; 27%)	M (2; 18%)	S (9; 82%)	P (6; 55%)	F (1; 9%)
	I (1; 9%)	D (6; 55%)	NS (0)	H (1; 9%)	B (10; 91%)
	U & S (0)	NS (0)		C – PP & P (0)	NS (0)
	U & I (1; 9%)			C – P & H (4; 36%)
	S & I (0)			C – (0)
	NS (1; 9%)			NS (0)
Wilson and Lipsey (2006b)	U (73; 100%)	W (0)	C (0)	PP (14; 19%)	M (8; 11%)
	S (0)	M (0)	S (0)	P (47; 64%)	F (6; 8%)
	I (0)	D (0)	NS (73; 100%)	H (12; 16%)	B (59; 81%)
	U & S (0)	NS (73; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Wilson and Lipsey (2006a)	U (0)	W (0)	C (0)	PP (0)	M (14; 30%)
	S (17; 36%)	M (0)	S (0)	P (31; 66%)	F (1; 2%)
	I (30; 64%)	D (0)	NS (47; 100%)	H (16; 34%)	B (32; 68%)
	U & S (0)	NS (47; 100%)		C – PP & P (0)	NS (0)
	U & I (0)			C – P & H (0)
	S & I (0)			C – (0)
	NS (0)			NS (0)
Wilson and Lipsey (2007)	U (89; 36%)	W (0)	C (0)	PP (0)	M (43; 17%)
	S (0)	M (21; 8%)	S (0)	P (178; 72%)	F (17; 7%)
	I (0)	D (228; 92%)	NS (249; 100%)	H (50; 20%)	B (179; 72%)
	U & S (0)	NS (0)		C – PP & P (21; 8%)	NS (10; 4%)
	U & I (0)			C – P & H (0)
	S & I (117; 47%)			C – (0)
	NS (43; 17%)			NS (0)
Total across reviews	U (494; 51.30%)	W (17; 1.77%)	C (22; 2.28%)	PP (56; 5.82%)	M (88; 9.14%)
	S (65; 6.75%)	M (70; 7.27%)	S (79; 8.20%)	P (417; 43.30%)	F (28; 2.91%)
	I (81; 8.41%)	D (311; 32.29%)	NS (862; 89.51%)	H (94; 9.76%)	B (440; 45.69%)
	U & S (5; 0.52%)	NS (565; 58.67%)		C – PP & P (68; 7.06%)	NS (407; 42.26%)
	U & I (1; 0.10%)			C – P & H (69; 7.17%)
	S & I (151; 15.68%)			C – (1; 0.10%)
	NS (166; 17.24%)			NS (258; 26.79%)

^a % = number of studies on school-based interventions with effects for peer aggression with characteristic/total number of studies on school-based interventions with effects for peer aggression.

^b NS = not specified.

^c U = universal, S = selected, I = indicated.

^d W = whole-school, M = multilevel, D = discrete.

^e C = comprehensive, S = specific.

^f PP = pre-primary school, P = primary school, H = high school.

^g F = female-only participants, M = male-only participants, B = participants of both genders.

^h All descriptives and effects reflect studies using measures of the level or extent of actual aggressive behavior or physical acts of aggression, either observed or reported only.

Most of the interventions were studied in North America, specifically within the USA (see Table 6 and Figure 2). This is exceptionally problematic as countries outside the USA show comparable, if not higher; levels of peer aggression (for instance, see; Chen & Avi Astor, 2010).

Table 6. No of studies by country, by WHO regions.

WHO regions	No. of studies
North America (total)^a	562
USA	527
Canada	35
Europe (total)	14
UK	6
Italy	2
Norway	2
Israel	1
Netherlands	1
Finland	1
Spain	1
South America (total)	7
Argentina	2
Columbia	2
Brazil	1
Jamaica	1
Mexico	1
Western Pacific (total)	7
Australia	6
China	1
South East Asia (total)	1
India	1
Africa (total)	0
Eastern Mediterranean (total)	0
Not Specified	372
Total Relevant Studies	963

^a We decided to split the Americas region into two: North (USA and Canada) and South (all other countries in the Americas), because of the vastly disproportionate amount of research typically conducted in North America.

Figure 2. WHO regions covered by peer aggression programmes.

Evidence for programmes to prevent peer aggression

We analyzed effectiveness in reducing peer victimization (see Table 7) and perpetration of peer aggression (see Table 8) separately. Less than half the studies used RCTs to examine program effects, thus some caution is required when interpreting findings relating to effectiveness of interventions.

Table 7. Effectiveness of peer aggression programmes to prevent victimisation.

Review	Prevention target				Intervention approach				Harmful effects	Outcome behavior^a	Significant moderators
	Universal: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Selected: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Indicated: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Not specified: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Whole-school: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Multi-level: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Discrete: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Not specified: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h
Barnes et al. (2014)	NS:1 (1)^f						NS:1 (1)^f – 5		No	4	Prevention target (universal vs selected)
Blank et al. (2010)^i	NS: 7 (1)^f				NS: 7 (1)^f – 10				No	1	Not considered
Bonell, Wells, et al.(2013)	PT: 5 (2)^f						PT: 5 (2)^f – 9		No	1; 2; 3; 4	Not considered
	FU: 7 (1)^f						FU: 7 (1)^f – 9
Gavine et al. (2016)^j	FU: 7 (1)^f					FU: 7 (1) – 1	NS: 6 (2)^f – 1		No	1; 2; 3; 4	Not considered
	NS: 5 (3)^f					NS: 4 (1) – 1
Hale et al. (2014)	FU: 4 (1)^f						FU: 4 (1)^f – 9		No	2	Not considered
Leff et al. (2010)	NS: 1 (2)^f				NS: 1 (1)^f – 1		NS: 1 (1)^f – 1		No	2; 4	Not considered
Sancassiani et al. (2015)	PT: 7 (1)^f				PT: 7 (1)^f – 4				No	1	Not considered
Vreeman and Carroll (2007)^k		PT: 7 (1)^f					PT: 7 (1)^f – 4		No	1	Not considered

^a 1 – Victimisation (specific type not specified); 2 – Physical victimsation; 3 – Verbal victimisation; 4 – Relational victimization.

^b Post-test effects.

^c Follow-up effects.

^d Post-test or follow-up measurement was not specified in review.

^e Effectiveness rating based on rubric. We rated the effectiveness of interventions reviewed according to the following rubric: (1) Found overall to be effective (i.e. peer aggression reduced in 100% of the effects, or as reported by a meta-analysis); (2) Found to be mostly effective (i.e. peer aggression reduced in 75% or more of the reviewed studies’ effects as a result of the intervention); (3) Found to be slightly effective (i.e. peer aggression reduced in 56–74% of the reviewed studies’ effects as a result of the intervention); (4) Overall mixed effects (i.e. peer aggression reduced in 45–55% of the reviewed studies’ effects as a result of the intervention); (5) A minority of studies found an effect (i.e. peer aggression reduced in 26–44% of the reviewed studies’ effects as a result of the intervention); (6) Found to be mostly ineffective (i.e. peer aggression reduced in 25% or less of the reviewed studies’ effects as a result of the intervention); (7) Found overall to be ineffective (i.e. no effects/change on peer aggression overall, or as reported by a meta-analysis).

^f Number of primary studies used in calculating effect.

^g Based on effect derived from meta-analysis.

^h Intervention type: 1 – Violence prevention programmes (also includes conflict training, modified discipline, anger management); 2 – Classroom management programmes; 3 – Life skills programmes (also includes problem solving skills training, perspective taking skills training and coping skills training); 4 – Social-emotional programmes (also includes social skills programmes); 5 – Cognitive behavioral interventions (also includes cognitive interventions and interventions for behavior modification) or social cognitive interventions; 6 – Peer mediators or mentoring; 7 – ECD/ECE; 8 – Combined; 9 – Other; 10 – Not specified.

ⁱ Many study outcomes were not clear. Only the studies that explicitly discussed school-based interventions for the outcomes of interest were considered.

^j Conservatively coded as multilevel instead of whole-school because they did not indicate at which levels they intervened.

^k Effects for victimization were only extracted for studies which did not also have effects on bullying as it was assumed the victimization would relate to bullying instead of peer aggression.

Table 8. Effectiveness of peer aggression programmes to prevent perpetration.

Review	Prevention target				Intervention approach				Harmful effects	Outcome behavior^a	Significant moderators
	Universal: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Selected: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Indicated: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Not specified: PT^b/FU^c/NS^d: E^e (n)^f^/^g	Whole-school: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Multi-level: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Discrete: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h	Not specified: PT^b/FU^c/NS^d: E^e (n)^f^/^g – IT^h
Allen-Meares et al. (2013)	NS:1 (2)^f	PT: 1 (1)^f						NS: 1 (2)^f −5	No	1; 4	Not considered
								PT: 1 (1)^f – 9
Barnes et al. (2014)	NS:3 (15)^f	NS: 3 (6)^f					NS: 3 (20)^f – 5		Yes	1; 2; 4	Prevention target (universal vs selected)
Blank et al. (2010)^i	NS: 3 (6)^f				NS: 1 (2)^f – 3				No	1	Not considered
					NS: 3 (4)^f – 10
Bond et al. (2013)^i			PT: 2 (5)^f			PT: 2 (3)^f – 9	PT: 1 (2)^f – 9		No	1; 4	Not considered
							FU: 1 (1)^f – 9
			FU: 1 (3)^f			FU: 1 (2)^f – 9
Bonell, Wells, et al. (2013)	PT: 5 (2)^f				FU: 7 (1)^f – 9		PT: 5 (2)^f – 9		No	1; 2	Not considered
	FU: 6 (3)^f						FU: 5 (2)^f – 9
Durlak et al. (2011)	PT: 1 (112)^gFU: 1 (21)^g							PT: 1 (112)^g – 4	No	3 (including 1)	None
								FU: 1 (21)^g – 4
Dymnicki et al. (2011)	PT: 1 (26)^g							PT: 1 (26)^g – 1	Yes	1	Not considered
Fagan and Catalano (2013)^j	PT: 4 (3)^f	FU: 3 (4)^f	PT: 4 (1)^f	PT: 1 (1)^f		FU: 3 (2)^f – 7	PT: 4 (1)^f – 1		Yes	1	Not considered
						PT: 4 (3)^f – 8
				FU: 4 (1)^f		FU: 3 (3)^f – 8	PT: 1 (1)^f – 4
Farahmand et al. (2011)^i	PT: 4 (2)^f	PT: 6 (3)^f						PT: 5 (5)^f – 9	Yes	4	Not considered
	FU: 7 (1)^f	FU: 7 (1)^f						FU: 7 (2)^f – 9
Gansle (2005) ^k				PT: 1 (22)^g				PT: 1 (22)^g – 1; 4; 5	No	4 (including 1 and 3)	Random assignment and treatment duration^l
				FU: 1 (9)^g				FU: 1 (9)^g – 1; 4; 5
Gavine et al. (2016)^m	PT: 5 (2)^f					PT: 7 (1)^f – 1	PT: 1 (1)^f – 1		Yes	1	Not considered
	FU: 5 (4)^f					FU: 6 (2)^f – 1	NS: 4 (3)^f – 1
	NS: 5 (7)^f					NS: 5 (3)^f – 1	FU: 3 (1)^f – 3
							FU: 6 (1)^f – 9
							NS: 1 (1)^f – 9
Hahn et al. (2007)^n	NS: 1 (65)^g							NS: 1 (30)^g – 4	No	1; 2; 3; 4	Not considered
								NS: 1 (6)^g – 5
								NS: 1 (2)^g – 6
								NS: 1 (25)^g – 9
Hale et al. (2014)	PT: 1 (2)^f	PT: 1 (1)^f			PT: 1 (1)^f – 4	PT: 1 (1)^f – 5	PT: 1 (1)^f – 4		No	1	Not considered
					FU: 1 (1)^f – 9	FU: 1 (1)^f – 8
	FU: 1 (4)^f					FU: 1 (1)^f – 9	FU: 1 (1)^f – 9
Leff et al. (2010)	PT: 5 (4)^f		NS: 1 (3)^f			NS: 1 (2)^f – 5	PT: 4 (2)^f – 1		No	1; 2	Not considered
							NS: 1 (1)^f – 1
	NS: 1 (1)^f						PT: 7 (1)^f – 2
							PT: 1 (1)^f – 4
							NS: 1 (1)^f – 5
Limbos et al. (2007)^o	NS: 5 (17)^f	NS: 5 (5)^f						NS: 5 (22)^f – 1	No	1	Not considered
Moestue et al. (2013)	PT: 1 (3)^f		NS: 1 (1)^f			PT: 1 (1)^f – 1	NS: 1 (1)^f – 2		No	1; 3	Not considered
							PT: 1 (1)^f – 4
							PT: 1 (1)^f – 9
Mytton et al. (2006)^p		PT: 1 (34)^g						PT: 1 (18)^g – 1	No	1	School level
								FU: 1 (2)^g – 1
								PT: 1 (7)^g – 4/6
		FU: 1 (7)^g
								FU: 7 (2)^g – 4/6
								PT: 1 (7)^g – 8
								FU: 7 (3)^g – 8
Oliver et al. (2011)^q	PT: 6 (4)^f					PT: 7 (1)^f – 2	PT: 5 (3)^f – 2		No	1; 3; 4	Not analyzed
Park-Higgerson et al. (2008)^r	PT: 2 (4)^f	PT: 4 (16)^f	PT: 4 (1)^f			PT: 5 (4)^f – 5	PT: 1 (1)^f – 1		No	1; 2; 4	Not considered
						FU: 5 (2)^f – 5	PT: 1 (1)^f – 2
						PT: 7 (1)^f – 8	FU: 7 (1)^f – 3
						FU: 7 (1)^f – 8	PT: 4 (4)^f – 5
	FU: 5 (4)^f	FU: 5 (4)^f	FU: 4 (1)^f			PT: 1 (2)^f – 9	FU: 5 (3)^f – 5
						PT: 7 (1)^f – 10	PT: 2 (3)^f – 8
							FU: 5 (1)^f – 8
							PT: 1 (5)^f – 9
Reddy et al. (2009)		PT: 1 (8)^g	PT: 1 (12)^g					PT: 1 (22)^g– 10	No	4	No analysis completed
		FU: 1 (3)^g	FU: 1 (3)^g					FU: 1 (22)^g– 10
Reese et al. (2010)^s				PT: 1 (59)^g				PT: 1 (59)^g – 9	No	4	No analysis completed
Sancassiani et al. (2015)	PT: 3 (2)^f				PT: 7 (1)^f – 3				No	1; 3	Not considered
					PT: 1 (1)^f – 4
Schindler et al. (2015)				PT: 7 (31)^g				PT: 7 (31)^g – 7	No	4 (including 1)	Extent of focus on socio-emotional development
Sklad et al. (2012)^t	PT: 1 (35)^g							PT: 1 (35)^g – 4/5	No	1; 4	Duration of program, school level.
	FU: 1 (35)^g							FU: 1 (35)^g – 4/5
Stoltz et al. (2012)^u			PT: 1 (24)^g			PT: 1 (13)^g – 5	PT: 1 (11)^g – 5		Yes	4 (including 1, 2 and 3)	Age of participants
Tolan et al. (2013)		NS: 4 (2)^f	NS: 7 (1)^f			NS: 4 (2)^f – 6	NS: 7 (1)^f – 6		No	1	Not considered
Vidrine (n.d.)^v				FU: 1 (10)^g				FU: 1 (10)^g – 3	No	4 (including 1 and 2)	Age of participants
Vreeman and Carroll (2007)^w	PT: 3 (4)^f	PT: 1 (2)^f	NS: 7 (1)^f	PT: 1 (1)^f	PT: 2 (2)^f – 6	PT: 7 (1)^f – 5	PT: 1 (1)^f – 3		Yes	1; 2; 4	Not considered
							PT: 4 (2)^f – 4
			Combination U&I				PT: 1 (1)^f – 6
							NS: 7 (1)^f – 6
	NS: 7 (1)^f				NS: 7 (1)^f – 10	PT: 1 (1)^f – 9
			PT: 5 (1)^f
Wilson and Lipsey (2006b)^x	PT: 1 (73)^g							PT: 1 (73)^g ^– 5	No	1; 3; 4	Socioeconomic status, routine practice, frequency of sessions, implementation quality
Wilson and Lipsey (2006a)^y			PT: 1 (47)^g					PT: 1 (47)^g – 5	No	1; 3; 4	Attrition, special education
S. J. Wilson and Lipsey (2007)^z	PT: 1 (77)^g	PT: 1 (108)^g		PT: 7 (43)^g		PT: 7 (21)^g ^– 4/5/6/9	PT: 3 (228)^g ^– 4/5/6/9		No	1; 3	Student socio-economic status, attrition, student risk level, implementation quality

^a 1 – Aggression and/or violence (physical aggression/violence, physical fights or attacks, verbal aggression); 2 – Relational aggression; 3 – Conduct problems/disorder; 4 – Externalizing behaviors.

^b Post-test effects.

^c Follow-up effects.

^d Post-test or follow-up measurement was not specified in review.

^e Effectiveness rating based on rubric. We rated the effectiveness of interventions reviewed according to the following rubric: (1) Found overall to be effective (i.e. peer aggression reduced in 100% of the effects, or as reported by a meta-analysis); (2) Found to be mostly effective (i.e. peer aggression reduced in 75% or more of the reviewed studies’ effects as a result of the intervention); (3) Found to be slightly effective (i.e. peer aggression reduced in 56–74% of the reviewed studies’ effects as a result of the intervention); (4) Overall mixed effects (i.e. peer aggression reduced in 45–55% of the reviewed studies’ effects as a result of the intervention); (5) A minority of studies found an effect (i.e. peer aggression reduced in 26–44% of the reviewed studies’ effects as a result of the intervention); (6) Found to be mostly ineffective (i.e. peer aggression reduced in 25% or less of the reviewed studies’ effects as a result of the intervention); (7) Found overall to be ineffective (i.e. no effects/change on peer aggression overall, or as reported by a meta-analysis).

^f Number of primary studies used in calculating effect.

^g Based on effect derived from meta-analysis.

^h Intervention type: 1 – Violence prevention programmes (also includes conflict training, modified discipline, anger management); 2 – Classroom management programmes; 3 – Life skills programmes (also includes problem solving skills training, perspective taking skills training and coping skills training); 4 – Social-emotional programmes (also includes social skills programmes); 5 – Cognitive behavioral interventions (also includes cognitive interventions and interventions for behavior modification) or social cognitive interventions; 6 – Peer mediators or mentoring; 7 – ECD/ECE; 8 – Combined; 9 – Other; 10 – Not specified.

ⁱ Many study outcomes were not clear. Only the studies that explicitly discussed school-based interventions for the outcomes of interest were considered.

^j Effects in not specified column represent combined universal and selective programmes.

^k Figures based on the number of comparisons instead of the number of studies.

^l Not strictly moderators, but considered their correlation with effect size.

^m Conservatively coded as multilevel instead of whole-school because they did not indicate at which levels they intervened.

ⁿ In this review, violence refers to both victimization and perpetration. Placed effects in perpetration table only as these seemed to be more common, and their studies likely followed this trend.

^o Review provides unconservative estimates of a study’s effectiveness. Each study only needed one positive effect to be considered effective overall.

^p All descriptives and effects reflect studies using measures of the level or extent of actual aggressive behavior or physical acts of aggression, either observed or reported only.

^q Utilized the effects of the ICC.10 analysis only in determining effectiveness.

^r Selected interventions seemed to included indicated samples at times, we relied on their classification.

^s Figures based on the number of outcome measures instead of studies.

^t Post-test considered effects up to and including 6 months after completion of the intervention. Follow-up included outcomes measured at least 7 months after completion of an intervention.

^u The multilevel number might be inflated due to this review not separating whole-school interventions.

^v Effects based on most distal results.

^w Coded post-test and follow-up information according to their study design information.

^x Only 74% of the included studies had effects at immediate post-test, thus our results are coded as post-test overall.

^y Results thought to be most reflective of post-test findings, and selected and indicated intervention effects were combined. There were more indicated than selected interventions, so effects were placed under the former column.

^z Could not isolate all effects for universal programmes and selected/indicated programmes. As subject risk was selected (42%) and indicated (19%), effects for selected/indicated interventions were thought to be more reflective of selective interventions, thus they were placed under this category. Effects with p < .10 were not considered effective in our analysis.

Prevention of victimization

Only eight reviews considered program effectiveness for reducing peer victimisation. The specific type of victimization explored in evaluations was not often specified, but when it was, the focus was on physical and relational victimization. The vast majority of programmes were universal in terms of target, and the majority of these scored poorly in terms of effectiveness. The single selective intervention was found to be ineffective. Most were discrete interventions and of these, only cognitive behavioral programmes showed promise for preventing victimization. Violence prevention programmes showed some promise in preventing victimization only when implemented as a whole-school intervention. No harmful effects were noted in this area overall. These findings tentatively suggest that discrete, cognitive-behavioral programmes that specifically target the prevention of victimisation show promise, and that consideration should be given to ways they can be included in whole-school interventions.

Programmes to prevent perpetration

All 31 reviews considered the capacity of school-based interventions to reduce perpetration of peer aggression. Intervention effects on the perpetration of aggression or violence (verbal or physical) in particular were considered in nearly every review, followed by studies that assessed broader outcomes that may include aggression, such as externalizing behavior.

Universal interventions seem to have undergone the most testing, followed by selected interventions, interventions where this information was not specified, and then indicated interventions. The majority of these were scored as effective, with 58% of the unspecified interventions scoring a 1 and 89% of the indicated interventions scoring 1. There is some evidence that the effects of universal interventions endure beyond the immediate post-test. For selected and indicated interventions, these effects were largely only found at post-test. Interventions which did not specify their prevention target demonstrated more mixed effects for reducing peer aggression immediately after program completion; however longer-term follow-up effects were largely positive.

With regards to intervention approaches, discrete programmes had the most evidence for effectiveness, followed by multi-level and whole-school programmes – although it should be noted that approach was specified in less than half of the reviews. Socio-emotional programmes have been found to be one of the most promising approaches, while cognitive behavioral and peer mentoring/mediation interventions have also fairly consistently demonstrated positive results. There was a broad range in the duration of these programmes. Socio-emotional programmes generally seemed to offer around 16 sessions. Unfortunately, session number information was often not specified. Program sessions were also implemented at a varying rate; once or twice a week seemed fairly common. Various school (mostly teachers) and research personnel were often involved in their implementation as well. Other types of intervention were effective in some studies but ineffective or harmful in others. Very few studies considered the effectiveness of whole-school programmes, suggesting the need for further research on these types of interventions. Promisingly, across all reviews, harmful effects (i.e. increased reports of perpetration) were reported in very few studies.

Only a handful of the reviews considered moderators of program effects. Well implemented cognitive behavioral interventions and those with more sessions each week were found to be beneficial (Wilson & Lipsey, 2006b). Considering socio-emotional and cognitive behavioral programmes together there was mixed evidence for short program duration to be associated with positive effects (Gansle, 2005; Sklad, Diekstra, De Ritter, Ben, & Gravesteijn, 2012), however a trend towards younger students benefiting more from these types of interventions was found in two reviews (Sklad et al., 2012; Stoltz, van Londen, Dekovic, de Castro, & Prinzie, 2012).

Discussion

There is very little literature on prevention of teacher-on-student violence (including corporal punishment) and student-on-teacher violence, even though these forms of violence seem quite common (see, for instance; Burton & Leoschut, 2013; Chen & Wei, 2011; Lee, 2015). More promisingly, there is a great deal of literature addressing prevention of IPV and even more dealing with peer aggression at school, although there are substantial gaps even here.

One key gap in the field is that studies often only measure one outcome, even where a program is theoretically likely to reduce more than one form of violence. For instance, peer aggression and dating violence share common risk factors (Smallbone & McKillop, 2015), and reductions in dating violence are thus highly likely to follow from interventions to reduce peer violence. Similarly, victimisation is seldom measured as an outcome. Importantly, the field of violence prevention will only be advanced if specific effects on aggressive behavior are reported separately from other forms of externalizing behaviors.

More high quality studies are also needed: RCTs with longer follow-up periods, lower risk of bias, and which explore mediation and moderation effects, will allow us to understand which programmes have sustained effects, what theoretical perspectives drive effective programmes (and so to understand not only what programmes work, but also why they work), and which programmes are generalizable to which groups (Gottfredson et al., 2015; Whitaker et al., 2006, 2013).

Another bias in the literature is that research on the effectiveness of interventions was almost exclusively completed in wealthier regions, particularly in the USA. This is exceptionally problematic, as school violence is a global problem (see, for instance; Burton & Leoschut, 2013; Chen & Avi Astor, 2010; Due et al., 2008; Fernandez-Fuertes & Fuertes, 2010; Wubs et al., 2009). More studies in high-violence, low-resource contexts are urgently needed.

Some interventions were identified as harmful, in that they led to increasing reports of aggression. This may be because programmes increased awareness and thus increased reporting (Taylor et al., 2010a; Taylor, Stein, & Burden, 2010b), but it may also have been because of adverse reactions to the intervention (DeGue et al., 2014). It may also be an artefact of study design: studies with short follow-up periods will be unable to differentiate an increase in response to heightened awareness from those that actually cause increased aggression, as it takes time for reporting to stabilize in response to awareness and then to decline in response to an effective program.

Another important focus for new studies should be components of effective interventions (Whitaker et al., 2013). This could be done either through developing and testing new programmes that build on what has been learned about effective interventions (Whitaker et al., 2006), or through meta-analytic studies of successful programmes (see, for instance; Kaminski, Valle, Filene, and Boyle, 2008). Studies of this nature assist in identifying the ‘active ingredients’ in programmes (Embry & Biglan, 2008).

This review does have some limitations. Firstly, we included only systematic reviews, and the information we were able to extract from each review was dependent on what was reported. This strategy means that promising interventions that had not yet been included in a review would have been missed. Secondly, we were unable to determine the extent of primary study duplication across the reviews on peer aggression. Therefore, the true size of the evidence base on school-based violence prevention interventions remains somewhat unclear. Thirdly, we only included studies published in English. Thus, our results do not reflect the findings of any possible reviews on school violence interventions published in other languages.

Despite these limitations, it is clear that a number of violence prevention initiatives have been successfully delivered at school. Several promising interventions to prevent IPV could be identified. Cognitive behavioral, social-emotional and peer mentoring/mediation programmes were effective for preventing perpetration of peer violence, and cognitive behavioral and whole-school violence prevention programmes show promise for preventing peer victimisation. While the field needs considerable development in order to be regarded as having a strong evidence base, the existing literature does provide us with a good foundation for tackling this serious problem.

Funding

This work was supported by a grant from the Know Violence in Childhood Learning Initiative, via the Centre for Justice and Crime Prevention; and by two grants to the first author: a Department of Science and Technology – National Research Foundation (South Africa) Centre of Excellence in Human Development doctoral scholarship [grant number D20160038]; and a University of Cape Town Doctoral Research Scholarship. Opinions expressed, and conclusions arrived at, are those of the authors and are not to be attributed to the CoE in Human Development.

Disclosure statement

No potential conflict of interest was reported by the authors.

Acknowledgements

We thank Alex D’Angelo, librarian at the University of Cape Town, who provided enormous help with the search strategy for this review. We are also extremely grateful to our research assistants: Amy Scheepers, Christina Barnes, Dina Hammerschlag, Ashleigh Nestadt, Tamsyn Naylor, Alexa Berlein, Nina Abrahams, and Thomas Guattari-Stafford, who helped with completing data screening and extraction.

Appendix A.

List of searched databases, trial registries and online journals

Table

EBSCOHost databases

Academic Search Premier

Business Source Premier

Africa-Wide Information

AHFS Consumer Medication Information

ATLA Religion Database with ATLASerials

CINAHL

Communication & Mass Media Complete

ERIC

Health Source: Nursing/Academic Edition

Humanities International Complete

International Bibliography of Theatre & Dance with Full Text

Library, Information Science & Technology Abstracts

MasterFILE Premier

MEDLINE

Philosopher’s Index

PsycARTICLES

PsycCRITIQUES

PsycINFO

PsycTESTS

SocINDEX with Full Text

Teacher Reference Center

ProQuest databases

Environment Abstracts

ERIC

International Bibliography of the Social Sciences (IBSS)

International Index to Performing Arts Full Text

Library and information Science Abstracts (LISA)

PAIS International and PAIS archive

PILOTS: Published International Literature On Traumatic Stress

ProQuest Education Journals

Social Services Abstracts

Sociological Abstracts

OCLC FirstSearch

Medline

ERIC

PubMed

Medline

Wiley Online Library

Sage Journals Online – 2014 Premier Package

Web of Science

Africa Bibliography

British Education Index- the free collections

ERIC (directly at eric.ed.gov)

Cochrane Library

Campbell Collaboration Libraries

Open Grey

BDENF

Global Health

HISA

LILACS

MedCarib

WPRIM

Table

Trial registries

Clinical Trials Registry: www.clinicaltrials.gov

The Pan-African Clinical Trials Registry: http://www.pactr.org/

The WHO violence prevention trials registry: http://www.preventviolence.info/Trials

Table

Hand search journals
Aggression and Violent Behavior
International Journal of Violence and schools
Journal of School Violence
Journal of Injury and Violence Research
Youth Violence and Juvenile Justice
Violence and Victims
Journal of Aggression, Maltreatment and Trauma
The School Community Journal
Journal of School Health
Journal of Interpersonal Violence
Journal of School Psychology
Journal of Educational Psychology
School Psychology Quarterly
Journal of Applied School Psychology
Contemporary School Psychology
Psychology in the Schools
British Journal of Educational Psychology
School Psychology International
School Psychology Review
Educational Psychology

Appendix B.

References of included reviews

Allen-Meares, P., Montgomery, K. L., & Kim, J. S. (2013). School-based social-work interventions: A cross-national systematic review. Social Work, 58(3), 253–262. doi:10.1093/sw/swt022

Barnes, T. N., Smith, S. W., & Miller, M. D. (2014). School-based cognitive-behavioral interventions in the treatment of aggression in the United States: A meta-analysis. Aggression and Violent Behavior, 19(4), 311–321. doi:10.1016/j.avb.2014.04.013

Blank, L., Baxter, S., Goyder, E., Naylor, P., Guillaume, L., Wilkinson, A., … Chilcott, J. (2010). Promoting well‐being by changing behaviour: a systematic review and narrative synthesis of the effectiveness of whole secondary school behavioural interventions. Mental Health Review Journal, 15(2), 43–53. doi:10.5042/mhrj.2010.0371

Bond, C., Woods, K., Humphrey, N., Symes, W., & Green, L. (2013). Practitioner Review: The effectiveness of solution focused brief therapy with children and families: a systematic and critical evaluation of the literature from 1990–2010. J Child Psychol Psychiatry, 54(7), 707–723.

Bonell, C., Wells, H., Harden, A., Jamal, F., Fletcher, A., Thomas, J., … Moore, L. (2013). The effects on student health of interventions modifying the school environment: Systematic review. Journal of Epidemiology and Community Health, 67(8), 677–681. doi:10.1136/jech-2012–202247

De Koker, P., Mathews, C., Zuch, M., Bastien, S., & Mason-Jones, A. J. (2014). A systematic review of interventions for preventing adolescent intimate partner violence. Journal of Adolescent Health, 54(1), 3–13. doi:10.1016/j.jadohealth.2013.08.008

De La Rue, L., Polanin, J., Espelage, D., & Pigott, T. (2014). School-Based Interventions to Reduce Dating and Sexual Violence: A Systematic Review.

Campbell Systematic Reviews, 2006(6), 1-37. doi:10.4073/csr.2006.6

Campbell Systematic Reviews, 2011(4), 1–55. doi:10.4073/csr.2011.4

Campbell Systematic Reviews, 2014(7), 1–110. doi:10.4073/csr.2014.7

DeGue, S., Valle, L. A., Holt, M. K., Massetti, G. M., Matjasko, J. L., & Tharp, A. T. (2014). A systematic review of primary prevention strategies for sexual violence perpetration. Aggression and Violent Behavior, 19(4), 346–362. doi:10.1016/j.avb.2014.05.004

Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., & Schellinger, K. B. (2011). The Impact of Enhancing Students’ Social and Emotional Learning: A Meta-Analysis of School-Based Universal Interventions. Child Development, 82(1), 405–432. doi:10.1111/j.1467–8624.2010.01564.x

Dymnicki, A. B., Weissberg, R. P., & Henry, D. B. (2011). Understanding How Programs Work to Prevent Overt Aggressive Behaviors: A Meta-Analysis of Mediators of Elementary School-Based Programs. Journal of School Violence, 10(4), 315–337.

Fagan, A. A., & Catalano, R. F. (2013). What works in youth violence prevention: A review of the literature. Research on Social Work Practice, 23(2), 141–156. doi:10.1177/1049731512465899

Farahmand, F. K., Grant, K. E., Polo, A. J., Duffy, S. N., & DuBois, D. L. (2011). School-Based Mental Health and Behavioral Programs for Low-Income, Urban Youth: A Systematic and Meta-Analytic Review. Clinical Psychology-Science and Practice, 18(4), 372–390. doi:10.1111/j.1468–2850.2011.01265.x

Gansle, K. A. (2005). The effectiveness of school-based anger interventions and programs: A meta-analysis. Journal of School Psychology, 43(4), 321–341. doi:10.1016/j.jsp.2005.07.002

Gavine, A. J., Donnelly, P. D., & Williams, D. J. (2016). Effectiveness of universal school-based programs for prevention of violence in adolescents. Psychology of Violence, 6(3), 390–399. doi:10.1037/vio0000052

Hahn, R., Fuqua-Whitley, D., Wethington, H., Lowy, J., Crosby, A., Fullilove, M., … Dahlberg, L. (2007). Effectiveness of universal school-based programs to prevent violent and aggressive behavior: a systematic review. Am J Prev Med, 33(2 Suppl), S114–129. doi:10.1016/j.amepre.2007.04.012

Hale, D. R., Fitzgerald-Yau, N., & Mark Viner, R. (2014). A Systematic Review of Effective Interventions for Reducing Multiple Health Risk Behaviors in Adolescence. American Journal of Public Health, 104(5), e19-e41. doi:10.2105/AJPH.2014.301874

Leff, S. S., Waasdorp, T. E., & Crick, N. R. (2010). A Review of Existing Relational Aggression Programs: Strengths, Limitations, and Future Directions. School Psychology Review, 39(4), 508–535.

Limbos, M. A., Chan, L. S., Warf, C., Schneir, A., Iverson, E., Shekelle, P., & Kipke, M. D. (2007). Effectiveness of interventions to prevent youth violence a systematic review. Am J Prev Med, 33(1), 65–74. doi:10.1016/j.amepre.2007.02.045

Moestue, H., Moestue, L., & Muggah, R. (2013). Youth violence prevention in Latin America and the Caribbean: a scoping review of the evidence. NOREF.

Mytton, J., DiGuiseppi, C., Gough, D., Taylor, R., & Logan, S. (2006). School-based secondary prevention programmes for preventing violence. Cochrane Database Syst Rev(3), CD004606. doi:10.1002/14651858.CD004606.pub2

Oliver, R., Reschly, D., & Wehby, J. (2011). The Effects of Teachers’ Classroom Management Practices on Disruptive, or Aggressive Student Behavior: A Systematic Review.

Park-Higgerson, H.-K., Perumean-Chaney, S. E., Bartolucci, A. A., Grimley, D. M., & Singh, K. P. (2008). The Evaluation of School-Based Violence Prevention Programs: A Meta-Analysis. Journal of School Health, 78(9), 465–479. doi:10.1111/j.1746-1561.2008.00332.x

Reddy, L. A., Newman, E., De Thomas, C. A., & Chun, V. (2009). Effectiveness of school-based prevention and intervention programs for children and adolescents with emotional disturbance: a meta-analysis. J Sch Psychol, 47(2), 77–99. doi:10.1016/j.jsp.2008.11.001

Reese, R. J., Prout, H. T., Zirkelback, E. H., & Anderson, C. R. (2010). Effectiveness of school-based psychotherapy: A meta-analysis of dissertation research. Psychology in the Schools, 47(10), 1035–1045. doi:10.1002/pits.20522

Sancassiani, F., Pintus, E., Holte, A., Paulus, P., Moro, M. F., Cossu, G., … Lindert, J. (2015). Enhancing the Emotional and Social Skills of the Youth to Promote their Wellbeing and Positive Development: A Systematic Review of Universal School-based Randomized Controlled Trials. Clinical practice and epidemiology in mental health : CP & EMH, 11(Suppl 1 M2), 21–40. doi:10.2174/1745017901511010021

Schindler, H. S., Kholoptseva, J., Oh, S. S., Yoshikawa, H., Duncan, G. J., Magnuson, K. A., & Shonkoff, J. P. (2015). Maximizing the potential of early childhood education to prevent externalizing behavior problems: A meta-analysis. J Sch Psychol, 53(3), 243–263. doi:10.1016/j.jsp.2015.04.001

Sklad, M., Diekstra, R., De Ritter, M., Ben, J., & Gravesteijn, C. (2012). Effectiveness of school-based universal, social, emotional, and behavioural programs: Do they enhance students’ development in the area of skill, behaviour, and adjustment? . Psychology in the Schools, 49(9), 892–909. doi:10.1002/pits.21641

Stoltz, S., van Londen, M., Dekovic, M., de Castro, B. O., & Prinzie, P. (2012). Effectiveness of individually delivered indicated school-based interventions on externalizing behavior. International Journal of Behavioral Development, 36(5), 381–388. doi:10.1177/0165025412450525

Tolan, P., Henry, D., Schoeny, M., Bass, A., Lovegrove, P., & Nichols, E. (2013). Mentoring Interventions to Affect Juvenile Delinquency and Associated Problems: A Systematic Review. Campbell Systematic Reviews, 2013(10), 1–48. doi:10.4073/csr.2013.10

Vidrine, S. (n.d). A Meta-Analysis of Interventions Targeting Executive Function to Improve Externalizing Behavior (08876177). Retrieved from http://search.ebscohost.com/login.aspx?direct = true&db = aph&AN = 97825234&site = ehost-live

Vreeman, R. C., & Carroll, A. E. (2007). A systematic review of school-based interventions to prevent bullying. Archives of Pediatrics & Adolescent Medicine, 161(1), 78–88. doi:10.1001/archpedi.161.1.78

Whitaker, D. J., Morrison, S., Lindquist, C., Hawkins, S. R., O’Neil, J. A., Nesius, A. M., … Reese, L. R. (2006). A critical review of interventions for the primary prevention of perpetration of partner violence. Aggression and Violent Behavior, 11(2), 151–166. doi:10.1016/j.avb.2005.07.007

Whitaker, D. J., Murphy, C. M., Eckhardt, C. I., Hodges, A. E., & Cowart, M. (2013). Effectiveness of primary prevention efforts for intimate partner violence. Partner Abuse, 4(2), 175–195. doi:10.1891/1946-6560.4.2.175

Wilson, S. J., & Lipsey, M. W. (2006a). The Effects of School-Based Social Information Processing Interventions on Aggressive Behavior: Part II: Selected/Indicated Pull-Out Programs: A Systematic Review.

Wilson, S. J., & Lipsey, M. W. (2006b). The Effects of School-Based Social Information Processing Interventions on Aggressive Behavior, Part I: Universal Programs. Campbell Systematic Reviews, 2006(5), 1–42. doi:10.4073/csr.2006.5

Wilson, S. J., & Lipsey, M. W. (2007). School-Based Interventions for Aggressive and Disruptive Behavior: Update of a Meta-Analysis. American journal of preventive medicine, 33(2 Suppl), S130-S143. doi:10.1016/j.amepre.2007.04.01

Appendix C.

Excluded reviews' reasons for exclusion and references

Table

Review	Reason for exclusion
Barlow, Smailagic, Ferriter, Bennett, and Jones (2010)	Too few school-based studies on outcomes of interest included
Baskin et al. (2010)	Unclear if outcomes of interest were included in review, and if they were to what extent
Beelmann and Raabe (2009)	Review of reviews
Bonell et al. (2013)	Protocol of a review, not an actual review
Bowman-Perrott, Burke, Nan, and Zaini (2014)	Too few school-based studies on outcomes of interest included
Candelaria, Fedewa, and Ahn (2012)	Could not clearly determine all of the studies which were school-based and which considered outcomes of interest
Cobb, Sample, Morgen, and Johns (2006)	Could not clearly determine all of the studies which were school-based and which considered outcomes of interest
Edwards and Hinsz (2014)	Too few school-based studies on outcomes of interest included
Ehiri, Hitchcock, Ejere, and Mytton (2007)	Protocol of a review, not an actual review
Fellmeth, Heffernan, Nurse, Habibula, and Sethi (2013)	Could not clearly determine which studies that were included in analysis were school-based
Ferguson, Miguel, Kilburn, and Sanchez (2007)	Effects for bullying and other types of violence were not reported separately. Could not determine extent to which violence outside of bullying was represented in statistics
Grove, Evans, Pastor, and Mack (2008)	Could not clearly determine how many school-based interventions were included, and thus how represented they were in the statistics
R. Hahn et al. (2007)	A published version of this report was also found in our searches and thus used instead.
R. A. Hahn et al. (2005)	Too few school-based studies on outcomes of interest included
Hankin, Hertz, and Simon (2011)	Too few school-based studies on outcomes of interest included
January, Casey, and Paulson (2011)	No clear effects on outcome behaviors of interest i.e., there was no way to determine to what extent aggression was an outcome which was considered in included studies and thus how represented it was in the analysis
Klasen and Crombag (2013)	Too few school-based studies on outcomes of interest included
Kristjansson et al. (2006)	Too few school-based studies on outcomes of interest included
D. M. Maggin, Chafouleas, Goddard, and Johnson (2011)	Did not consider outcomes of interest
Daniel M. Maggin, Johnson, Chafouleas, Ruberto, and Berggren (2012)	Outcome behaviors of interest comprised less than 5% of outcome behaviors considered in review. No way of determining the effects specific to outcome behaviors of interest either
Matjasko et al. (2012)	Review of reviews
McCart, Priester, Davies, and Azen (2006)	School-based intervention effects could not be separated from the effects for interventions implemented elsewhere. No way of knowing how many school-based interventions were included either
Meirelles dos Santos and Giglio (2012)	Too few school-based studies on outcomes of interest included
Montgomery and Maunders (2015)	Too few school-based studies on outcomes of interest included
Ozabaci (2011)	Too few school-based studies on outcomes of interest included
Parker and Turner (2013)	Too few school-based studies on outcomes of interest included
Piquero et al. (2008)	Too few school-based studies on outcomes of interest included
Piquero, Jennings, Farrington, and Jennings (2010)	Unclear to what extent outcomes of interest were included and separating school-based effects was impossible to do
Polanin and Espelage (2015)	Primary study
Reichow, Barton, Boyd, and Hume (2014)	Did not consider outcomes of interest
Sentenac et al. (2012)	Too few school-based studies on outcomes of interest included
Silverman et al. (2008)	Too few school-based studies on outcomes of interest included
Singh et al. (2011)	Too few school-based studies on outcomes of interest included
Solomon, Klein, Hintze, Cressey, and Peller (2012)	Unclear to what extent outcomes of interest are included and represented in statistics
Sugimoto-Matsuda and Braun (2014)	Did not consider outcomes of interest.
Ting (2009)	Did not consider outcomes of interest
Vannest, Davis, Davis, Mason, and Burke (2010)	No distinct separation of the effects for the outcome behaviors of interest and other behaviors
Walsh, Zwi, Woolfenden, and Shlonsky (2015)	Too few school-based studies on outcomes of interest included
Weisburd, Telep, Hinkle, and Eck (2008)	Too few school-based studies on outcomes of interest included
Wilson and Institute for Public Policy Studies (2005)	Showed significant similarity to Wilson and Lipsey (2007) article. Later article was chosen to be included in review as it included a greater number of studies and was published more recently

Barlow, J., Smailagic, N., Ferriter, M., Bennett, C., & Jones, H. (2010). Group-based parent-training programmes for improving emotional and behavioural adjustment in children from birth to three years old. Cochrane Database of Systematic Reviews(3). doi:10.1002/14651858.CD003680.pub2

Baskin, T. W., Slaten, C. D., Crosby, N. R., Pufahl, T., Schneller, C. L., & Ladell, M. (2010). Efficacy of Counseling and Psychotherapy in Schools: A Meta-Analytic Review of Treatment Outcome Studies 1Ψ7. The Counseling Psychologist, 38(7), 878–903. doi:10.1177/0011000010369497

Beelmann, A., & Raabe, T. (2009). The effects of preventing antisocial behavior and crime in childhood and adolescence: Results and implications of research reviews and meta-analyses. European Journal of Developmental Science, 3(3), 260–281.

Bonell, C., Jamal, F., Harden, A., Wells, H., Parry, W., Fletcher, A., … Moore, L. (2013). Systematic review of the effects of schools and school environment interventions on health: evidence mapping and synthesis. doi:10.3310/phr01010

Bowman-Perrott, L., Burke, M. D., Nan, Z., & Zaini, S. (2014). Direct and Collateral Effects of Peer Tutoring on Social and Behavioral Outcomes: A Meta-Analysis of Single-Case Research. School Psychology Review, 43(3), 260–285.

Candelaria, A. M., Fedewa, A. L., & Ahn, S. (2012). The effects of anger management on children’s social and emotional outcomes: A meta-analysis. School Psychology International. doi:10.1177/0143034312454360

Cobb, B., Sample, P. L., Morgen, A., & Johns, N. R. (2006). Cognitive-Behavioral Interventions, Dropout, and Youth With Disabilities: A Systematic Review. Remedial and Special Education, 27(5), 259–275.

Edwards, S. R., & Hinsz, V. B. (2014). A Meta-Analysis of Empirically Tested School-Based Dating Violence Prevention Programs. SAGE Open, 4(2). doi:10.1177/2158244014535787

Ehiri, J. E., Hitchcock, L. I., Ejere, H. O. D., & Mytton, J. A. (2007). Primary prevention interventions for reducing school violence. Cochrane Database of Systematic Reviews(1). doi:10.1002/14651858.CD006347

Fellmeth, G. L., Heffernan, C., Nurse, J., Habibula, S., & Sethi, D. (2013). Educational and Skills-Based Interventions for Preventing Relationship and Dating Violence in Adolescents and Young Adults: A Systematic Review. Campbell Systematic Reviews, 14. doi:10.4073/csr.2013.14

Ferguson, C. J., Miguel, C. S., Kilburn, J. C., & Sanchez, P. (2007). The Effectiveness of School-Based Anti-Bullying Programs: A Meta-Analytic Review. Criminal Justice Review, 32(4), 401–414. doi:10.1177/0734016807311712

Grove, A. B., Evans, S. W., Pastor, D. A., & Mack, S. D. (2008). A meta-analytic examination of follow-up studies of programs designed to prevent the primary symptoms of oppositional defiant and conduct disorders. Aggression and Violent Behavior, 13(3), 169–184. doi:10.1016/j.avb.2008.03.001

Hahn, R., Fuqua-Whitley, D., Wethington, H., Lowy, J., Liberman, A., Crosby, A., … Dahlberg, L. (2007). The Effectiveness of Universal School-Based Programs for the Prevention of Violent and Aggressive Behavior: A Report on Recommendations of the Task Force on Community Preventive Services. Morbidity and Mortality Weekly Report (MMWR), 56(RR07), 1–12.

Hahn, R. A., Bilukha, O., Crosby, A., Fullilove, M. T., Liberman, A., Moscicki, E., … Task Force on Community Preventive, S. (2005). Firearms laws and the reduction of violence: a systematic review. Am J Prev Med, 28(2 Suppl 1), 40–71. doi:10.1016/j.amepre.2004.10.005

Hankin, A., Hertz, M., & Simon, T. (2011). Impacts of Metal Detector Use in Schools: Insights From 15 Years of Research*. Journal of School Health, 81(2), 100–106. doi:10.1111/j.1746-1561.2010.00566.x

January, A. M., Casey, R. J., & Paulson, D. (2011). A Meta-Analysis of Classroom-Wide Interventions to Build Social Skills: Do They Work? School Psychology Review, 40(2), 242–256.

Klasen, H., & Crombag, A.-C. (2013). What works where? A systematic review of child and adolescent mental health interventions for low and middle income countries. Social Psychiatry and Psychiatric Epidemiology, 48(4), 595–611. doi:10.1007/s00127-012-0566-x

Kristjansson, E., Robinson, V., Petticrew, M., MacDonald, B., Krasevec, J., Laura Janzen, … Tugwell, P. (2006). School Feeding for Improving the Physical and Psychosocial Health of Disadvantaged Students. Campbell Systematic Reviews, 14. doi:10.4073/csr.2006.14

Maggin, D. M., Chafouleas, S. M., Goddard, K. M., & Johnson, A. H. (2011). A systematic evaluation of token economies as a classroom management tool for students with challenging behavior. J Sch Psychol, 49(5), 529–554. doi:10.1016/j.jsp.2011.05.001

Maggin, D. M., Johnson, A. H., Chafouleas, S. M., Ruberto, L. M., & Berggren, M. (2012). A systematic evidence review of school-based group contingency interventions for students with challenging behavior. Journal of School Psychology, 50(5), 625–654. doi:10.1016/j.jsp.2012.06.001

Matjasko, J. L., Vivolo-Kantor, A. M., Massetti, G. M., Holland, K. M., Holt, M. K., & Dela Cruz, J. (2012). A systematic meta-review of evaluations of youth violence prevention programs: Common and divergent findings from 25 years of meta-analyses and systematic reviews. Aggression and Violent Behavior, 17(6), 540–552. doi:10.1016/j.avb.2012.06.006

McCart, M. R., Priester, P. E., Davies, W. H., & Azen, R. (2006). Differential Effectiveness of Behavioral Parent-Training and Cognitive-Behavioral Therapy for Antisocial Youth: A Meta-Analysis. Journal of Abnormal Child Psychology, 34(4), 525–541. doi:10.1007/s10802-006-9031-1

Meirelles dos Santos, M. B., & Giglio, J. S. (2012). P-1158 – Group play therapy for children promising findings of a systematic revision of literature. European Psychiatry, 27, 1–1. doi:10.1016/S0924-9338(12)75325-7

Montgomery, P., & Maunders, K. (2015). The effectiveness of creative bibliotherapy for internalizing, externalizing, and prosocial behaviors in children: A systematic review. Children and Youth Services Review, 55, 37–47. doi:10.1016/j.childyouth.2015.05.010

Ozabaci, N. (2011). Cognitive behavioural therapy for violent behaviour in children and adolescents: A meta-analysis. Children and Youth Services Review, 33(10), 1989–1993. doi:10.1016/j.childyouth.2011.05.027

Parker, B., & Turner, W. (2013). Psychoanalytic/Psychodynamic Psychotherapy for Children and Adolescents Who Have Been Sexually Abused: A Systematic Review.

Piquero, A. R., Farrington, D., Jennings, W. G., Tremblay, R., Piquero, A., & Welsh, B. (2008). Effects of Early Family/Parent Training Programs on Antisocial Behavior and Delinquency: A Systematic Review.

Piquero, A. R., Jennings, W., Farrington, D., & Jennings, W. G. (2010). Self-Control Interventions for Children Under Age 10 for Improving Self-Control and Delinquency and Problem Behaviors: A Systematic Review.

Polanin, J., & Espelage, D. (2015). Using a Meta-analytic Technique to Assess the Relationship between Treatment Intensity and Program Effects in a Cluster-Randomized Trial. Journal of Behavioral Education, 24(1), 133–151. doi:10.1007/s10864-014-9205-9

Reichow, B., Barton, E. E., Boyd, B. A., & Hume, K. (2014). Early Intensive Behavioral Intervention (EIBI) for Young Children with Autism Spectrum Disorders (ASD): A Systematic Review. Campbell Systematic Reviews, 9. doi:10.4073/csr.2014.9

Sentenac, M., Arnaud, C., Gavin, A., Molcho, M., Gabhainn, S. N., & Godeau, E. (2012). Peer victimization among school-aged children with chronic conditions. Epidemiologic reviews, 34, 120–128. doi:10.1093/epirev/mxr024

Singh, R. D., Jimerson, S. R., Renshaw, T., Saeki, E., Hart, S. R., Earhart, J., & Stewart, K. (2011). A Summary and Synthesis of Contemporary Empirical Evidence Regarding the Effects of the Drug Abuse Resistance Education Program (D.A.R.E.). Contemporary School Psychology, 15, 93–102.

Silverman, W. K., Ortiz, C. D., Viswesvaran, C., Burns, B. J., Kolko, D. J., Putnam, F. W., & Amaya-Jackson, L. (2008). Evidence-based psychosocial treatments for children and adolescents exposed to traumatic events. J Clin Child Adolesc Psychol, 37(1), 156–183. doi:10.1080/15374410701818293

Solomon, B. G., Klein, S. A., Hintze, J. M., Cressey, J. M., & Peller, S. L. (2012). A meta-analysis of school-wide positive behavior support: An exploratory study using single-case synthesis. Psychology in the Schools, 49(2), 105–121. doi:10.1002/pits.20625

Sugimoto-Matsuda, J. J., & Braun, K. L. (2014). The Role of Collaboration in Facilitating Policy Change in Youth Violence Prevention: a Review of the Literature. Prevention Science, 15(2), 194–204. doi:10.1007/s11121-013-0369-7

Ting, S.-M. R. (2009). Meta-Analysis on Dating Violence Prevention Among Middle and High Schools. Journal of School Violence, 8(4), 328–337. doi:10.1080/15388220903130197

Vannest, K. J., Davis, J. L., Davis, C. R., Mason, B. A., & Burke, M. D. (2010). Effective Intervention for Behavior With a Daily Behavior Report Card: A Meta-Analysis. School Psychology Review, 39(4), 654–672.

Walsh, K., Zwi, K., Woolfenden, S., & Shlonsky, A. (2015). School-Based Education Programmes for the Prevention of Child Sexual Abuse: A Systematic Review.

Weisburd, D., Telep, C. W., Hinkle, J. C., & Eck, J. E. (2008). The Effects of Problem-Oriented Policing on Crime and Disorder. Campbell Systematic Reviews, 14. doi:10.4073/csr.2008.14

Wilson, S. J., & Lipsey, M. W. (2005). The Effectiveness of School-Based Violence Prevention Programs for Reducing Disruptive and Aggressive Behavior: A Meta-analysis by Centre for Evaluation Research and Methodology, Institute for Public Policy and Vanderbilt.

Appendix D.

Data extraction document

Part B: AMSTAR

Table

			Yes	No	Can’t answer	Comments
1.	Question:	Was an ‘a priori’ design provided?
	Explanation:	The research question and inclusion criteria should be established before the conduct of the review
	Note:	Need to refer to a protocol, ethics approval, or pre-determined/a priori published research objectives to score a ‘yes’
2.	Question:	Was there duplicate study selection and data extraction?
	Explanation:	There should be at least two independent data extractors and a consensus procedure for disagreements should be in place
	Note:	Two people do study selection, two people do data extraction, consensus process or one person checks the other’s work
3.	Question:	Was a comprehensive literature search performed?
	Explanation:	At least two electronic sources should be searched. The report must include years and databases used (e.g. Central, EMBASE, and MEDLINE). Key words and/or MESH terms must be stated and where feasible the search strategy should be provided. All searches should be supplemented by consulting current contents, reviews, textbooks, specialized registers, or experts in the particular field of study, and by reviewing the references in the studies found
	Note:	If at least two sources + one supplementary strategy used, select ‘yes’ (Cochrane register/central counts as two sources; a grey literature search counts as supplementary)
4.	Question:	Was the status of publication (i.e. grey literature) used as an inclusion criterion?
	Explanation:	The authors should state that they searched for reports regardless of their publication type. The authors should state whether or not they excluded any reports (from the systematic review), based on their publication status, language etc.
	Note:	If review indicates that there was a search for ‘grey literature’ or ‘unpublished literature’, indicate ‘yes’. SIGLE database, dissertations, conference proceedings, and trial registries are all considered grey for this purpose. If searching a source that contains both grey and non-grey, must specify that they were searching for grey/unpublished lit
5.	Question:	Was a list of studies (included and excluded) provided?
	Explanation:	A list of included and excluded studies should be provided.
	Note:	Acceptable if the excluded studies are referenced. If there is an electronic link to the list but the link is dead, select ‘no’
6.	Question:	Were the characteristics of the included studies provided?
	Explanation:	In an aggregated form such as a table, data from the original studies should be provided on the participants, interventions and outcomes. The ranges of characteristics in all the studies analyzed e.g., age, race, sex, relevant socioeconomic data, disease status, duration, severity, or other diseases should be reported
	Note:	Acceptable if not in table format as long as they are described as above
7.	Question:	Was the scientific quality of the included studies assessed and documented?
	Explanation:	‘A priori’ methods of assessment should be provided (e.g. for effectiveness studies if the author(s) chose to include only randomized, double-blind, placebo controlled studies, or allocation concealment as inclusion criteria); for other types of studies alternative items will be relevant
	Note:	Can include use of a quality scoring tool or checklist, e.g., Jadad scale, risk of bias, sensitivity analysis, etc., or a description of quality items, with some kind of result for EACH study (‘low’ or ‘high’ is fine, as long as it is clear which studies scored ‘low’ and which scored ‘high’; a summary score/range for all studies is not acceptable)
8.	Question:	Was the scientific quality of the included studies used appropriately in formulating conclusions?
	Explanation:	The results of the methodological rigor and scientific quality should be considered in the analysis and the conclusions of the review, and explicitly stated in formulating recommendations
	Note:	Might say something such as ‘the results should be interpreted with caution due to poor quality of included studies’. Cannot score ‘yes’ for this question if scored ‘no’ for question 7
9.	Question:	Were the methods used to combine the findings of studies appropriate?
	Explanation:	For the pooled results, a test should be done to ensure the studies were combinable, to assess their homogeneity (i.e. Chi-squared test for homogeneity, I^2 ). If heterogeneity exists a random effects model should be used and/or the clinical appropriateness of combining should be taken into consideration (i.e. is it sensible to combine?)
10.	Note:	Indicate ‘yes’ if they mention or describe heterogeneity, i.e., if they explain that they cannot pool because of heterogeneity/variability between interventions
	Question:	Was the likelihood of publication bias assessed?
11.	Explanation:	An assessment of publication bias should include a combination of graphical aids (e.g. funnel plot, other available tests) and/or statistical tests (e.g. Egger regression test, Hedges-Olken)
	Note:	If no test values or funnel plot included, score ‘no’. Score ‘yes’ if mentions that publication bias could not be assessed because there were fewer than 10 included studies.
	Question:	Was the conflict of interest included?
	Explanation:	Potential sources of support should be clearly acknowledged in both the systematic review and the included studies
	Note:	To get a ‘yes’, must indicate source of funding or support for the systematic review AND for each of the included studies
		Total AMSTAR Score:

Part C: Extraction

Please extract the descriptive information from each study first. Information relating to each field needs to be marked with an ‘X’ if relevant to the review. If a field is relevant, please also specify the number of primary studies in the review this information pertains to. If the information in a particular field is not specified please specify this using the NS (not specified) option.

Descriptive information

Table

1.	No. of studies in review	Not specified
2.	No. of studies on school-based interventions with violent behavior as an outcome	Not specified
3.	No. of studies on school-based interventions with ONLY attitudes towards violence as an outcome	Not specified

Table

Prevention target		Characteristic relevant	No. of primary studies
1	Universal only
2	Selected only
3	Indicated only
4	Combination
5	Not specified
Approach
1	Whole-school only
2	Discrete only
3	Combination
4	Not specified
Content
1	Comprehensive only
2	Specific only
3	Combination
4	Not specified
School level
1	Pre-primary only
2	Primary only
3	High only
4	Combination PP + P
5	Combination P + H
6	Combination all
7	Not specified
Participant gender
1	Female-only
2	Male-only
3	Mixed
4	Not specified
Study design
1	Randomised controlled trial
2	Quasi-experimental
3	Not specified
WHO regions		Characteristic relevant	No. of primary studies for each country in region
1	North America
2	South America
3	Western Pacific
4	Europe
5	Eastern Mediterranean
6	Africa
7	South East Asia
8	Not specified

Effects information

1. For meta-analyses with school-based interventions which have effects on violence overall please mark the appropriate column in the table below with an ‘X’.

Table

Reduction in violence (effective)	No effect on violence (ineffective)	Increase in violence (harmful)

2. For reviews where primary studies’ effects need to be individually extracted, please add the name of each primary study which needs to have their effects extracted first. Then, for each of these relevant primary studies mark the appropriate column in the table below with an ‘X’.

Table

No.	Primary study name	Reduction in violence (effective)	No effect on violence (ineffective)	Increase in violence (harmful)
1
2
3
4
5
6
7
	Sub-totals:
Grand total number of effects:

3. Were harmful effects reported or found?

Yes/No

4. Did the reviews include individual primary studies with effect sizes on relevant outcomes?

Yes/No

5. Specific type of outcome behavior considered in review (e.g. physical aggression): _________________________________________________________________________________

6. Specific type of intervention considered in review (e.g. social-emotional program): _________________________________________________________________________________