Mental health promotion and problem prevention in schools: what … · 2015-10-19 · EMBASE, ERIC, CINALH, Sociological Abstracts, ASSIA, Psycinfo, the Cochrane Database of Systematic

Mental health promotion and problem preventionin schools: what does the evidence say?

KATHERINE WEARE* and MELANIE NINDEducation, University of Southampton, Southampton, Hampshire SO17 1BJ, UK*Corresponding author. Email: [email protected]

SUMMARY

The European Union Dataprev project reviewed workon mental health in four areas, parenting, schools, theworkplace and older people. The schools workpackagecarried out a systematic review of reviews of work onmental health in schools from which it identified evi-dence-based interventions and programmes andextracted the general principles from evidence-basedwork. A systematic search of the literature uncovered 52systematic reviews and meta-analyses of mental health inschools. The interventions identified by the reviews hada wide range of beneficial effects on children, familiesand communities and on a range of mental health,social, emotional and educational outcomes. The effectsizes associated with most interventions were generallysmall to moderate in statistical terms, but large in termsof real-world impacts. The effects associated with inter-ventions were variable and their effectiveness could notalways be relied on. The characteristics of more effective

interventions included: teaching skills, focusing on posi-tive mental health; balancing universal and targetedapproaches; starting early with the youngest childrenand continuing with older ones; operating for a lengthyperiod of time and embedding work within a multi-modal/whole-school approach which included such fea-tures as changes to the curriculum including teachingskills and linking with academic learning, improvingschool ethos, teacher education, liaison with parents,parenting education, community involvement and coor-dinated work with outside agencies. Interventions wereonly effective if they were completely and accuratelyimplemented: this applied particularly to whole-schoolinterventions which could be ineffective if notimplemented with clarity, intensity and fidelity. Theimplications for policy and practice around mentalhealth in schools are discussed, including the suggestionof some rebalancing of priorities and emphases.

Key words: Mental health; Social and emotional learning; Schools; Children; Systematic review

BACKGROUND

Mental health and schools

Childhood and adolescence provide key oppor-tunities to develop the foundations for mentalhealth and prevent mental health problems,and the school is a unique resource to helpachieve this. Schools can help tackle theproblem of the substantial number of childrenand young people who experience mentalhealth problems. Around 25% of children andyoung people in the developed world have an

identifiable mental health problem (Hardenet al., 2001), of whom 10% fulfil criteria for amental health disorder. Schools can alsopromote positive mental health and createresilience, providing the child or young personwith resources to thrive and, in adverse con-ditions, to cope by buffering negative stressors.For children who come from less thanoptimum home backgrounds and neighbour-hoods the intervention of the school can bethe turning point for many children with fewother supports (Gross, 2008).

Health Promotion International, Vol. 26 No. S1

doi:10.1093/heapro/dar075

# The Author (2011). Published by Oxford University Press. All rights reserved.

For Permissions, please email: [email protected]

i29

by guest on Decem

ber 5, 2012http://heapro.oxfordjournals.org/

Dow

nloaded from

http://heapro.oxfordjournals.org/

The importance of the school for mentalhealth, and the opportunities it provides forinterventions have been evident for some time,and the last two decades have seen considerablegrowth in mental health research and interven-tions. There are literally thousands of schoolmental health interventions in operation acrossthe world, some of which have been evaluated.These go under many names: mental health,‘social and emotional learning’ (SEL), ‘emotion-al literacy’, ‘emotional intelligence’, ‘resilience’,‘lifeskills’ and ‘character education’ (Weare,2010). The world leader in terms of interventionsis the USA, generating the most interventionsand investing the most in evaluation. Thousandsof what are effectively mental health interven-tions are operating with various levels of demon-strable success. Of these, �20 majorinterventions are consistently identified as suc-cessful by rigorous systematic reviews (Zinset al., 2004; CASEL, 2010). Australia is also thescene of thriving work with some interventionsstarting to produce robust and positive evalu-ations (Adi et al., 2007a; Shucksmith et al., 2007).

The objectives of the DataPrev schools workpackage

Against this background, one of the workpackages of the DataPrev project reviewedmental health interventions in schools. Theschools’ work package aimed to clarify the evi-dence for and create a database of keyevidence-based principles, approaches andinterventions that are relevant to Europe andproduce policy and practice guidelines to assistpolicy-makers as they select approaches andinterventions for implementation. A full reportis available on the Dataprev website. http://dataprevproject.net/

Objectives of this paper

This paper will draw upon the review of theschools work package. It will describe the meth-odology used in the systematic review andoutline its main findings in terms of the qualityand content of the reviews, the impact of inter-ventions across a range of mental health issuesand outline the themes that emerged from thereview, including the apparent characteristics ofmore effective interventions. It will discuss theimplications of these finding for work in schoolmental health promotion, including in Europe.

METHODOLOGY

Identifying reviews

As there were already many good qualityreviews of primary studies in the field, both theschools and the parenting work packages didnot look at primary studies but instead soughtexisting good quality systematic reviews, reviewsof reviews, data synthesis, data extraction,meta-analyses and evidence-based databases.The scope was: from 1990, school-aged childrenand young people (4–19 years in mainstream,special and independent institutions), andincluded universal, targeted, indicated, school-based, and classroom-based interventions,including those in which schools worked withfamilies and the community, to improve mentalhealth, to prevent mental illness and problemsand/or tackle mental health problems.

The databases searched included MEDLINE,EMBASE, ERIC, CINALH, SociologicalAbstracts, ASSIA, Psycinfo, the CochraneDatabase of Systematic Reviews, DARECENTRAL, SIGLE, and the Social SciencesCitation Index. Additionally, reviews were foundthrough research agencies in the field that userigorous evaluation methods, personal contactswith established reviewers, pursuing referencesfrom previous reviews and overviews, and hand-searching two journals: Advances in SchoolMental Health Promotion and InternationalJournal of School Mental Health.

In keeping with most modern definitionsmental health was broadly conceptualized andover 80 search terms were used to reflect thewide and inclusive nature of the field, parallelingthe concepts of mental health and search termsused in the parenting review. Mental health wasseen as including positive wellbeing, and sogeneric terms such as wellbeing and quality oflife and descriptions of positive mental statessuch as happiness and self-esteem were used. Itwas thought to encorporate mental health skillsand capacities, so terms such as communicationand resilience were included. It was also seen asincluding both internalizing and externalizingmental health problems, so terms such asdepression and anger were used. It was appreci-ated that taking such a broad and inclusive viewwould result in a set of reviews which would besomewhat heterogeneous in terms of focus,subject, target group and so on, but, in line withother good quality large scale reviews entitle

i30 K. Weare and M. Nind

by guest on Decem


Dow

nloaded from

http://dataprevproject.net/






‘mental health and wellbeing’ in recent years[e.g. (Adi, 2007a,b; Shucksmith et al., 2007)] itwas felt important to reflect the breath of thefield and bring together findings from a range ofstudies to try to achieve a comprehensive picturewhich reflected current broad concept of mentalhealth.

Content and thematic analysis

Data were extracted using two standardizedforms. One noted aspects of the content: thefocus of review, aims of the intervention, whodelivered, frequency and duration, population,setting and timing. The second noted results:number of included studies, relevant outcomesincluding effects sizes where given, findings andauthors’ conclusions. The reviews were subjectedto content and thematic analysis, with recurrentthemes and trends identified and particularattention paid to any quantitative estimates ofeffectiveness. Descriptive data on the reviews,and the authors’ results and conclusions aresummarized in Table 1.

Critical appraisal and the weighting of evidencefrom the reviews

A third standardized form was used for criticalappraisal, using criteria, informed by a seminalpaper by (Oxman et al., 1994), by the practiceof other well-conducted reviews of reviews inthis area (Browne et al., 2004; Adi et al.,2007a,b) and by the parallel review of reviewsof parenting by the Dataprev project (StewartBrown and Shrader McMillan, this volume).The criteria used are outlined in Table 2, theyincluded the relevance of the focus of thereview, whether the interventions included tookplace only in schools and related communityand family settings (the focus of this review) orincluded clinical settings (not within the focusof this review). Further criteria were whetherthe review addressed a focussed question;included only studies with an element of control(RCTs and CCTs) had a stated effective, appro-priate and comprehensive search and reviewstrategy, appraised the quality of the studiesincluded, provided a meta-analysis and/or datasynthesis, and included quantitative presen-tation of results quantitatively with effects sizes,percentages and/or confidence intervals.

Table 2 rates the quality of the 52 reviews.Reviews were graded to the extent to which they

met the various criteria. The reviews on whichthis review placed most weight and from whichthe key results are derived were those judged tobe of high quality (which only included thosewhere the interventions reviewed did not includeclinical contexts, as it was felt that this inclusionwas a dilution of the relevance of the findings).Those of medium quality were then consulted tosupport or shed further light on the key resultsalready identified. Reviews of low quality wereused as additional support only where there wasalready strong evidence from high-and medium-quality reviews. (The starring system used inTable 2 has been used throughout the text, withreferences to the reviews starred appropriatelyto help make the weight of evidence clear to thereader.)

RESULTS

The reviews

Over 500 studies were identified, of which 52reviews met the inclusion criteria.

Universal/targeted

Most (46) of the reviews were universal in scope,i.e. they targeted all children in the group,including those without problems, and 14 ofthese also explored the impact of interventionsand approaches on targeted or indicated popu-lations within their larger sample (discussed inmore detail below). Six reviews were entirelyfocused on targeted and/or indicated popu-lations, focusing on children with or showingsigns of various mental health problems (2) vio-lence and aggression (2) and emotional andbehavioural problems (2). See Table 1 for moredetails.

Critical review and quality of the reviews

As Table 2 shows, 27 of the reviews were ofhigh quality (6 or 7 criteria met), 18 were ofmedium quality (5 criteria met) and 7 of lowquality (4 or less criteria met). The major areaof methodological weakness was the inclusionof studies without an element of control, mostcommonly interrupted time line (18 reviews).The second most common weakness was thefailure to enumerate results in any way (15reviews). Third weakness was the inclusion of

Mental health promotion and problem prevention in schools i31

by guest on Decem


Dow

nloaded from


Table 1: Critical review and quality assessment

Author (year) Review focused onlyon interventionscentred aroundschools, i.e. notclinical?

Clearlyfocusedquestion?

Only controlled trials(RCTS, CCTs) included?

Stated andappropriate andcomprehensivesearch strategy?

Quality of studies assessedand used to guide results?

Substantialmeta-analysis/data synthesis?

Results presented to allowquantitative assessment ofimpact?

Quality1

(Adi et al., 2007a) Yes Yes Yes Yes Yes Yes Yes ***(Adi et al., 2007b) Yes Yes Yes Yes Yes Yes Yes ***(Bayer et al., 2009) Yes No Yes Yes Yes Yes No, a narrative review **(Beelman et al., 1994) No Yes Yes Yes Yes Yes Yes **(Beelman and Losel, 2006) No Yes Yes Yes Yes Yes Yes **(Berkowitz and Bier, 2007) Yes Yes No, also included pre and

post designYes Yes Yes Yes ***

(Blank et al., 2009) Yes Yes No, also includedinterrupted time series

Yes Yes Yes Yes ***

(Browne et al., 2004) Yes Yes No, included reviews usingwide range of studies butmostly with controls

Yes Yes Yes Yes ***

(Catalano et al., 2002) Yes Yes Yes Yes Yes Yes Yes ***(Clayton et al., 2001) Yes No No, included wide range of

studiesYes No, programmes with low

quality evaluations andinconclusive resultsincluded as ‘promising’

Yes No *

(Diekstra, 2008a) Yes Yes No, review of reviews usingwide range of studies

Yes Yes Yes No, described features of thereview

*

(Diekstra, 2008b) Yes Yes Yes Yes Yes Yes Yes ***(Durlak and Wells, 1997) No Yes Yes Yes Yes Yes Yes **(Durlak et al., 2007) Yes Yes Yes Yes Yes Yes Yes ***(Durlak and Weissberg,

2007)Yes Yes Yes Yes Yes Yes Yes ***

(Durlak et al., 2011) Yes Yes Yes Yes Yes Yes Yes ***(Ekeland et al., 2004) No Yes Yes Yes Yes Yes Yes **(Farrington and Ttofi, 2009) Yes Yes Yes Yes Yes Yes Yes ***(Gansle, 2005) Yes Yes Yes Yes Yes Yes Yes ***(Garrard and Lipsey, 2007) Yes Yes Yes Yes Yes Yes Yes ***(Green et al., 2005) Yes No No, review of reviews but no

requirement that theprimary research usedcontrols or comparisongroups

Yes Yes Yes No, described features of thereviews

*

(Greenberg et al., 2001) No Yes Yes Yes Yes Yes No, descriptively by featuresof the programmes

**

(Hahn, 2007) Yes No No, included wide range ofdesigns

Yes Yes Yes Yes **

(Haney and Durlak, 1998) No Yes Yes Yes Yes Yes Yes **(Harden et al., 2001) No Yes No, included wide range of

designsYes Yes Yes Yes, for the systematic

review and in depthreview

**

(Hoagwood and Erwin, 1997) Yes No Yes Yes Yes Yes No, results described assignificant, programmes aseffective, or not

**

i32K

.W

eare

an

dM

.N

ind

by guest on December 5, 2012 http://heapro.oxfordjournals.org/ Downloaded from


(Horowitz and Garber, 2006) No Yes Yes Yes Yes Yes No, results described aseffective, mixed or noteffective

**

(Kraag et al., 2006) Yes Yes Yes Yes Yes Yes Yes ***(Maxwell et al., 2008) No No No, included wide range of

designsYes No, wide range of studies

included some of mediummethodological quality

Yes No, thematically anddescriptively.

*

(McCarthy and Carr, 2002) Yes Yes No, included wide range ofdesigns

Yes Yes Yes No, descriptive/narrativereview

**

(Merry et al., 2004) No Yes Yes Yes Yes Yes Yes **(Mytton et al., 2002) Yes Yes Yes Yes Yes Yes Yes ***(Neil and Christensen, 2007) Yes Yes No, included before and

afterYes, of studies

fromAustraliaonly

Yes No, descriptive/narrativereview

**

(O’Mara et al., 2006) Yes Yes Yes Yes Yes Yes Yes ***(Park-Higgerson et al., 2008) Yes Yes Yes Yes Yes Yes Yes ***(Payton et al., 2008)

(indicated)Yes Yes Yes Yes Yes Yes ***

(Reddy et al., 2009) Yes Yes No, included variousresearch designs

Yes Yes Yes Yes ***

(Rones and Hoagwood,2000)

Yes Yes Yes Yes Yes Yes No descriptive/narrativereview

**

(Schachter et al., 2008) Yes Yes No, included pre-post test orpost test

Yes No, most studies classified aspoor

Yes No *

(Scheckner et al., 2002) Yes Yes Yes Yes Yes Yes Yes ***(Shucksmith et al., 2007) Yes Yes Yes Yes Yes Yes ***(Sklad et al., 2010) Yes Yes Yes Yes Yes Yes Yes ***(Stage and Quiroz, 1997) Yes Yes No used the interrupted

time seriesYes Yes Yes Yes **

(Tennant et al., 2007) Yes Yes No, review of reviews,looking at systematicreviews

Yes Yes Yes No (for one study only) **

(Tilford et al., 1997) Yes No No, used wide range ofdesigns

Yes No No No, narrative review *

(Vreeman and Carroll, 2007) Yes Yes Yes Yes Yes Yes Yes ***(Waddell et al., 2007) No Yes Yes Yes Yes Yes No (for one study only) **(Wells et al., 2004) Yes Yes Yes Yes Yes Yes No, narrative synthesis ***(Wilson et al., 2003) Yes Yes No, included pre-post test

designYes Yes Yes Yes ***

(Wilson and Lipsey, 2006a) Yes Yes Yes Yes Yes Yes Yes ***(Wilson and Lipsey, 2006b) Yes Yes Yes Yes Yes Yes Yes ***(Wilson and Lipsey, 2007) Yes Yes No, included pre-post test Yes Yes Yes Yes ***

1Key for assessment for quality.***, high quality 6 or 7 of the criteria met, including no interventions in clinical settings included (coloumn 2).**, medium quality, 5 criteria met.*, low quality ,4 criteria met.

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i33



Table 2: Focus and results of reviews

Author (year) Number and focus ofincluded studies

Mental health outcomes measured Authors’ results Authors’ conclusions

(Adi et al., 2007a) 31 studies, 15 RCTs and 16CCTs describing 30interventions in primaryschools that take auniversal approach topromoting mentalwellbeing but notprimarily focused onviolence or bullying

Broad range, including:emotional wellbeing (including happiness and confidence, and the

opposite of depression/anxiety);psychological wellbeing (including resilience, mastery, confidence,

autonomy, attentiveness/involvement and the capacity tomanage conflict and problem solve)

social wellbeing/good relationships with others

Overall effect sizes were calculated for only 4 of the 31 studies, onegood quality RCT: 0.15 and 0.30; two moderate quality RCTs:0.37 and 0.25 and one moderate quality CCT: 0.27 and 0.18, Allsuggesting small to medium effects of interventions on mentalhealth. Good evidence to support multi-componentprogrammes, covering classroom curricula and the schoolenvironment, which include significant teacher training anddevelopment and support and training for parenting. Typicallylong-term involving children for between 1 and 3 years.

Some evidence that short-term stress and coping programmesdelivered by psychologists are effective in the short term.Effectiveness may be enhanced by addition of a programme forparents. More evidence is needed on sustainability andeffectiveness of psychologists versus teachers in providing suchinterventions

Reasonable quality evidence that short-term conflict resolutionprogrammes delivered by teachers and involving peer mediationare effective in the short term.

Reasonable quality evidence that long-term programmes coveringsocial problem solving, social awareness and emotional literacy,in which teachers reinforce the classroom curriculum in allinteractions with children are effective in the long term evenwhen delivered alone

Some evidence to support further trials of programmes in whichretired volunteers are recruited to help in schools Insufficientevidence to make recommendations relating to the optimumbalance of universal and targeted approaches, but there wassome evidence that the combination may be effective

There are no trials identified in this systematic review to showdifferential effects according to age, gender, ethnic or socialgroups

Clear positive impact ofmulti-component/wholeschool approaches

Need more research on thecontent and process ofdelivery of interventions(including the content andapproach to teachertraining and parentingsupport, barriers andfacilitators toimplementation) and themost effectivecombination of targetedand universal approaches

Need more research onpromising programmes todevelop coping skills andreduce stress and anxietyand other short-termclass-based programmes,e.g. conflict resolution, toassess long-termeffectiveness, and crosscultural adaptability

Good quality CCTs ofprogrammes adopting ahealth promoting schoolapproach to mental healthpromotion should beundertaken using a rangeof robust outcomemeasures, positive as wellas negative, andmeasuring long-termimpact

Secondary research is neededto update reviews ofmeasures of child mentalhealth and primaryresearch to developmeasures which fill gapsin availability

i34K

.W

eare

an

dM

.N

ind



(Adi et al., 2007b) 17 studies reported in 23papers—11 RCTs and 6CCTs in primary schoolsthat take a universalapproach to promotingmental wellbeing primarilyfocused on violence orbullying

Focused only on measures of violence and proxy measures ofaggression, e.g. antisocial behaviour or social skills, ininterventions

Most common outcomes measured—teacher, peer or self-reportedmeasures of behaviour problems or social competence

Some trials used observations of children’s behaviour, by teachersand/or independent observers

Some trials used child reports of self-victimization or victimizationof peers, or children’s knowledge of bullying

A few studies reported event-based data such as exclusions orexpulsions from school due to violence, or visits to the schoolnurse for injury

Two studies with long-term follow-up reported on arrests and courtappearances for delinquency

6/17 trials showed a clearly positive impact, and 8/17 trials showed apossibly positive impact. The most common violence-relatedoutcomes measured were teacher-, peer- or self-reportedmeasures of behaviour—either behaviour problems or socialcompetence

Only 2 of the 17 contained effects sizes. One showed a standardizedmean difference of 0.41 with effects maintained in seven studiesreporting 12 months follow-up, the other effect sizes were foundof the order of 0.1 for universal interventions and 0.3 fortargeted or indicated populations

There is evidence from three out of four ‘moderate’ quality RCTs,and two out of two good quality CCTs of the effectiveness ofmulti-component programmes, which typically combine socialskills development curriculum, teacher training in managementof behaviour and parenting education

Moderate to good evidence that a multi-component programmewhich aims to change school ethos (PeaceBuilders) waseffective, including at 2 year follow-up in improving outcomesrelated to violence and mental health, measured by teacherreports of social competence and aggression

The evidence relating to curriculum only programmes (e.g. secondstep) suggests short, but not longer term effectiveness

There is some evidence that the Good Behaviour Game waseffective in the short term, not evident at 2 and 6 yearsfollow-up for all children but some evidence it reduced violencein the most aggressive boys. This programme may be useful incombination with others

There was some evidence of short-term effectiveness of the OlweusAnti Bullying programme after a year, but not evident at 2 yearfollow-up

Widespread evidence that programmes have more effect on boysthan girls, white than black children, high- than low-risk children

Clear positive impact ofmulti-component/wholeschool approaches Promisingevidence for approacheswhich attempt to change theschool culture and ethos,changing values, attitudes andbehaviours relating to the wayboth staff and students treateach other. Such wholeschool programmes couldhave enduring impact on theschool culture, so the‘intervention’ would notnormally have an endpoint.Long-term evaluation wouldneed to track these childrenthrough into secondaryschools. Whole schoolinterventions that do notchange school ethos andvalues are unlikely to showsustainable changes Higherimpact on high-risk childrenmay be the ‘ceilingeffect’ Main impact is onboys. Need to developprogramme which take theviolent behaviourdemonstrated by girls intoaccount

(Bayer et al.,2009)

50 RCTS from which 6programmes wereidentified that showedeffectiveness and could betranslated into anAustralian context

Emotional and behavioural problems Most programmes targeted children’s behavioural problems, and afew targeted emotional problems. At school age, the GoodBehaviour Game class programme showed evidence ofeffectiveness. Interventions exist primarily for behaviour and, toa lesser extent, emotional problems, and could be disseminatedfrom research to mainstream in Australia, ensuring fidelity tooriginal programmes

Future research shoulddevelop programmestargeting emotionalproblems, and replicateeffective programmes forbehaviour problems inquality populationtranslation trials.Randomized trial methodsin staged roll-outs candetermine population costbenefits for children’smental health withoutdelaying dissemination

(Beelman et al.,1994)

49 studies of interventionsthat teach socialcompetence to children

Measures of social competence, e.g. social-cognitive skills and socialadjustment

Social competence training showed moderate effect sizes. However,effect sizes were lower than in previous studies. Two mainproblems were identified: First, significant effect sizes were foundonly when direct goal criteria (e.g. social-cognitive skills) wereevaluated, whereas there were few effects on broader constructs(e.g. social adjustment). Second, long-term effects were weak

Further primary studies areneeded on thegeneralization andmaintenance of change

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i35



Table 2: Continued



(Beelman andLosel, 2006)

84 research reports with 127treatment-controlcomparisons on socialskills training to preventanti-social behaviour andpromote socialcompetence

Measures of A small but significant overall positive effect of d ¼ 0.39 atpost-intervention and d ¼ 0.28 at 3 months follow-up

Effect sizes where somewhat greater for outcome measures of socialcompetence than for measures of anti-social behaviour,particularly when delinquency was assessed

Slight tendency for more intensive treatments to be more effective,e.g. at follow-up most intensive treatments were the only ones toimpact on anti-social behaviour (d ¼ 0.30)

Cognitive-behavioural programmes were the only ones thatimpacted significantly on anti-social behaviour (d ¼ 0.50) andgenerally had the best results in terms of generalization overtime and on outcome criteria, compared with cognitive orbehavioural only

Authors, trainers and supervised students had more effect thanteachers (d ¼ 0.47 versus d ¼ 0.33)

Indicated approaches had higher effect sizes than universalapproaches

A small but significant overallpositive effect of frominterventions

Because most studies dealtwith small sample sizes,non-official outcome dataand measurements after,1 year the results shouldbe treated with caution.Further high-qualitystudies with long-termoutcomes are needed,particularly outside theUSA

(1) Anti-social behaviour, e.g. aggression, delinquency, disruptionand/or

(2) Social competence, e.g. social interaction skills, pro-socialbehaviour, self-control or social problem solving

Broad range of data used, e.g. reports, observations, official records,but had to be reported in sufficient detail to permit reliableeffect size computation

(Berkowitz andBier, 2007)

33 effective programmes wereidentified reported by 64empirical studies, plus 5meta-analyses andliterature reviews

Examined to identify themost common effects andthe most common sharedpractices of charactereducation programmes

Sociomoral cognitionSexual behaviourProsocial behaviours and attitudesProblem-solving skillsKnowledge about riskDrug useRelationshipsViolence/aggressionSchool behaviourKnowledge/attitudes about riskEmotional competencyAcademic achievementAttachment to schoolPersonal moralityCharacter knowledgeCommunicative competenceAttitudes towards teachers

The most commonly reported effects of character education weresocio-moral cognition, pro-social behaviours and attitudes,problem-solving skills, reduced drug use, reduced violence/aggression, school behaviour, knowledge and attitudes aboutrisk, emotional competency, academic achievement, attachmentto school and decreased general misbehaviour

The percentage of tests of that variable that were significantlypositive were, in order, sexual behaviour (91%), characterknowledge (87%) socio-moral cognition (74%), problem-solvingskills (64%), emotional competency (64%) relationships (62%),attachment to school (61%), academic achievement (59%)andcommunicative competency (50%)

For those 10 programmes that assessed fidelity, there was a cleartrend for complete and accurate implementation to result inmore outcome effectiveness than incomplete or inaccurateimplementation

Character education can workwhen implemented withfidelity and broadly, and has avery robust impact. Effectivecharacter education tends toinclude: professionaldevelopment; studentinteractive pedagogicalstrategies; an explicit focus oncharacter/ethics; directtraining of social andemotional competencies;modelling of character;aligned classroom/behaviourmanagement strategies andcommunity service and/orservice learning

i36K

.W

eare

an

dM

.N

ind



(Blank et al.,2009)

37 studies which aim topromote emotional andsocial wellbeing bymodifying behaviouramong children insecondary education(aged 11–18), taking auniversal approach

30 on negative and 7 onpositive behaviour

Self-reported attitude/behaviour: (28 of which 3 þve, 25 2ve)Teacher behaviour/attitude rating (2 2ve)Parent behaviour/attitude rating (1 þve)Academic achievement (2 ve)Routine data (6 2ve)Effectiveness of programme (1 þve, 2 2ve)Observed school wide changes (2 þve)

5 studies calculated effect sizes. An RCT study of an aggressionviolence intervention showed that a programme with whichincluded parental and community involvement was moreeffective than a programme of social development curriculum,with effect sizes 0.41 and 0.31, respectively

An RCT study of a ‘trans-theoretical based bullying curriculum’delivered by the internet showed that the intervention groupwere four times more likely not to participate in bullying (effectsize 0.42). An RCT study of a violence prevention curriculumshowed very small difference in violence scale ratings (effect size0.1)

A CBA study of a programme to reduce aggression and violenceshowed effect sizes of 0.5–0.73 in measures of endorsement ofsocial exclusion and tolerance of physical and verbal aggression

A CBA study of a behavioural management programme to reduceaggression and violence which measured self-reported aggressionand found an effect size of 0.02

The effect sizes demonstrated are therefore highly variable which isunsurprising given the heterogeneity of included interventionsand outcome measures

On balance, in 8 out of 11 well-conducted studies, there wasevidence of the effectiveness of good quality universalinterventions to support curriculum approaches to whole schoolinterventions, which aim to promote positive behaviours,however three well-conducted studies did not support this

The evidence on the effectiveness of the role of teachers, andexternal agencies in delivering interventions, and on involvingparents, was equivocal

On balance there was reasonable evidence to support involvingyoung people as peer educators/mediators in interventions

The lack of well-conductedstudies for this age groupin the school setting makeit hard to draw firmconclusions

The results of this reviewbroadly support thetheoretical literature onwellbeing in secondaryschools including thedifferentiation betweeninterventions which aimto promote positivebehaviour andinterventions which aimto prevent negativebehaviour

The literature to supportwhole school/multi-componentinterventions in general isnot well developed,especially in terms ofgood quality effectivenessstudies. The vast majorityof interventions identifiedin the review are based inthe classroom and take acurriculum approach.More research is neededon whole school/multi-componentapproaches at secondarylevel

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i37



Table 2: Continued



(Browne et al.,2004)

23 reviews of effective andefficient mental health(focusing on mentalhealth problems)non-clinical programmesfor school aged children

Reviews discussed efforts to reduce deficiencies related todepression, anxiety, externalizing/internalizing or otherpsychological/social problems, reductions in risky behaviours,increase competence and resilience through various protectivestrategies

1 study calculated effects sizes. At post-test from 0.41 to 1.70 (smallto large). At follow-up from 0.60 to 1.69 (medium to large)

Otherwise commented on the characteristics of more effectiveprogrammes: universal rather than targeted; multi-modal, i.e.multiple, integrated elements involving family, school andcommunity; for younger children (but programmes for olderchildren also effective); specific aim rather than broad andunfocussed; theoretically based; interactive rather thaninformation only/didactic, involved families; positive rather thanfear based; long term with follow-up rather than short term/intensive; adults as supports and mentors; peer mentoring

Effect sizes decreased over time for knowledge and skills acquisitionand behaviour reduction suggesting the need for periodicfollow-up and reinforcement

Best practices include: early,long-term interventionincluding reinforcement;follow-up and anecological focus withfamily and communitysector involvement;consistent adult staffingand interactive,non-didactic programmingadapted to gender, ageand cultural needs

Need to encourageinteragency co-operation,ensure services reachappropriate segments ofthe population; replicateof best practices andpublicise informationabout benefits and costsavings

(Catalano et al.,2002)

25 effective and robustlyevaluated programmes topromote positive youthdevelopment

Not explicitly stated, but programmes selected for match with goalsof positive youth development, i.e.

19 of the effective programmes showed positive changes in youthbehaviour, including significant improvements in interpersonalskills, quality of peer and adult relationships, self-control,problem solving, cognitive competencies, self-efficacy,commitment to schooling and academic achievement

24 of the effective programmes showed significant improvements inproblem behaviours, including drug and alcohol use, schoolmisbehaviour, aggressive behaviour, violence, truancy, high-risksexual behaviour and smoking

Although one-third of the effective programmes operated in only asingle setting, for the other two-thirds, combining the resourcesof the family, the community, and the community’s schools werethe other ingredients of success

Effective programmes shared common themes and principles. Allsought to strengthen social, emotional, cognitive and/orbehavioural competencies, self-efficacy, and family andcommunity standards for healthy social and personal behaviour.Seventy per cent also targeted healthy bonds between youth andadults, increased opportunities and recognition for youthparticipation in positive social activities. Ninety-six per cent usedtraining manuals or other forms of structured curricula. Eightyper cent lasted 9 months or more

Although a broad range ofstrategies produced theseresults, the themes commonto success involved methodsto: strengthen social,emotional, behavioural,cognitive and moralcompetencies; buildself-efficacy; shape messagesfrom family and communityabout standards for positiveyouth behaviour; increasehealthy bonding with adults,peers and younger children;expand opportunities andrecognition for youth whoengage in positive behaviourand activities; providestructure and consistency inprogramme delivery andintervene with youth for atleast 9 months or more

Bonding Social Emotional Cognitive Behavioural and moralcompetence Self-determination Spirituality Self-efficacy Clearand positive identity Belief in the future Recognition for positivebehaviour Opportunities for prosocial involvement Prosocialnorms (healthy standards for behaviour)

i38K

.W

eare

an

dM

.N

ind



(Clayton et al.,2001)

Very wide range Qualitative analysis. Concluded there are many effectiveprogrammes aimed at anti-violence, conflict resolution andpeace which is encouraging

However, all used different methods and tools of evaluation

Need more RCTs andlong-term evaluationsusing multiple datasources

Violence prevention shouldbe founded on soundtheory, e.g. childdevelopment, shouldteach skills as well associal norms, find ways towork effectively withchildren without acuteproblems, be tailored tospecific populations,include adequate teachertraining and strengthenself-worth

Pro-social behaviour, e.g. aggression, violenceImproved skills e.g. social decision making, resiliencePro-social attitudes, e.g. friendliness, empathyPositive self-concept

(Diekstra, 2008a) 19 reviews of interventionsthat take a universalapproach social andemotional learning

Wide range of outcomes, reflecting the broad nature of the fieldChanges in social and emotional skills, attitudes to self and others,

externalizing and behavioural problems and disorders, antisocial/criminal behaviour, drug use/abuse, internalizing or emotionalproblems and disorders—stress, anxiety, depression, suicidaltendencies, attitudes towards school including truancy andabsence, school test scores and grades

Very strong and significant impact of many programmes on socialand emotional skills and attitudes to self and others

Strong and widespread impact on externalizing problems, less so oninternalizing problems, but still a significant impact (e.g. 10%reduction in depression)

Strong impact on attitudes to school and test scores and gradesFew studies had follow-up evaluations—those that do showed good

stability over timeMuch variability in effects, probably due to differences in

implementationShort-term programmes much less effective than long term.Programmes showed most dramatic impact on high-risk children, but

impacted on all children, on all ages and both gendersTeachers as effective as other professionals in delivering

programmes, although need training. Only when school staffconduct the intervention does student academic performanceimprove significantly—possibly because school staff are involvedin both aspects of school, and SEL/SFL impacts on wider schoolculture where school staff involved in delivery

SEL/SFL programmes workand meet a wide varietyof goals

Strong support for universalapproach across all ageranges

The most effectiveprogrammes aretheoretically consistent,highly interactive, use avariety of didactic forms,are implemented in smallgroups, cover bothspecific and general lifeskills and are cast withinsupporting communitiesand environmentalstrategies

To be effective programmeshave to run for at least 3/6months with at leastweekly sessions, ideallywith boosters later

More research needed onwhat the people whodeliver programmes needto be successful, e.g.characteristics, training,support.

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i39



Table 2: Continued



(Diekstra, 2008b) 76 studies of universalschool-based programmes

17 of these conducted outsideof the USA

Social skillsAnti-social behaviourSubstance abusePositive self-imageAcademic achievementMental disorders/healthPro-social behaviour

Universal school-based SEL/SFL programmes generally havepositive effects on a number of desirable outcomes. In the shortterm, the largest effects are on social-emotional skills, attitudestowards self, pro-social behaviour, academic achievement andreduction in anti-social behaviour. These effects decrease in thelong term but remain significant. Some effects increase, e.g. thereduction in mental disorders

Overall effect size of USA and non-US studies similar for the onlyoutcome on which comparison possible, social–emotional skills

SEL/SLF works and isbeneficial to children aroundthe globe

(Durlak andWells, 1997)

177 primary preventionprogrammes designed toprevent behavioural andsocial problems in childrenand adolescents

60 outcome measures used, summarized as:Problems/symptomsExternalizingInternalizingAcademic achievementSociometric statusCognitive processesPhysiological measures

Most categories of programmes produced outcomes similar to orhigher in magnitude than those obtained by many otherestablished preventive and treatment interventions in the socialsciences and medicine

Programmes modifying the school environment, individually focusedmental health promotion efforts, and attempts to help childrennegotiate stressful transitions yield significant mean effectsranging from 0.24 to 0.93. In practical terms, the averageparticipant in a primary prevention programme surpasses theperformance of between 59 and 82% of those in a control group,and outcomes reflect an 8–46% difference in success ratesfavouring prevention groups

Most categories of programmes had the dual benefit of significantlyreducing problems and significantly increasing competencies

Priorities for future researchinclude clearer specificationof intervention proceduresand programme goals,assessment of programmeimplementation, morefollow-up studies anddetermining howcharacteristics of theintervention and participantsrelate to different outcomes

(Durlak et al.,2007)

526 universalcompetence-promotionoutcome studies ofpositive youthdevelopment programmesexploring effects onschools, families andcommunities

Effects of interventions on schools, families and communities, or acombination

64% of the positive youth development interventions attemptedsome type of microsystemic or mesosystemic change involvingschools, families or community-based organizations in an attemptto foster developmental competencies in children andadolescents. Only 24% of the reports provided quantitative dataon the change that occurred in targeted systems. However, 6 ofthe 7 post-mean effect sizes were statistically significant andranged in magnitude from 0.34 (for family environment) to 0.78(for classroom level change), ranging from modest to large inmagnitude. The only non-significant (and negative) post-meaneffect of 20.26 (youths’ bonding to community adults) was basedon only two interventions

Attempts to change socialsystems—schools, familiesand communities affectingchildren and adolescentscan be successful

Future work should measuremore thoroughly theextent to which thesystemic changes that aretargeted throughintervention are achieved,and investigate how suchchanges contribute to thedevelopment andsustainability of desirableoutcomes

i40K

.W

eare

an

dM

.N

ind



(Durlak andWeissberg,2007)

73 after-school programmesthat attempted to promotepersonal and social skills

Feelings and attitudes—child self-perceptions, e.g. self-esteem,self-concept, self-efficacy) and bonding to school (e.g. positivefeelings about school and teachers)

Behavioural adjustment—positive social behaviours (e.g. expressionof feelings, positive interactions, assertiveness), problembehaviours (e.g. aggression and rebelliousness) and drug use,legal and illegal

School performance—on achievement tests, grades and schoolattendance

Youth who participate in after-school programmes improvesignificantly in feelings and attitudes, indicators of behaviouraladjustment and school performance. Overall average effect sizewas 0.22.

After-school programmes succeeded in improving youths’ feelings ofself-confidence and self-esteem (0.34), school bonding (positivefeelings and attitudes towards school, 0.14), positive socialbehaviours (0.19), reduction in problem behaviours such asaggression, non-compliance and conduct problems (0.18)achievement test scores (0.18), school grades (0.11) and schoolattendance (0.10).

Programmes that used evidence-based skill training approaches, andwere sequenced, active, focused and explicit were consistentlysuccessful in producing multiple benefits for youth (mean effectsizes ranged from 0.24 to 0.35) while those that did not use suchprocedures were not successful in any outcome area

After school programmes canproduce multiple benefitsthat pertain to youths’personal, social andacademic life

To be effective they need touse evidence—approaches—besequenced, active, focusedand explicit

(Durlak et al.,2011)

207 studies of universal socialand emotional learningprogrammes

Social-emotional skillsAttitudes towards self and othersPositive social behavioursConduct problemsEmotional distressAcademic performance

11% improvement in achievement tests; 25% improvement in socialand emotional skills; 10% decrease in classroom misbehaviour,anxiety and depression. These effects held during follow-upperiods of at least 6 months

School staff can effectively deliver social and emotional learningThe grand study-level mean for all 207 interventions was 0.28

(CI ¼ 0.25–0.29) Effect sizes: SEL skills 0.60, Academicperformance 0.28, emotional distress 0.25, positive socialbehaviours 0.24, attitudes 0.23, conduct problems 0.20

Effective for all grades and agesStudent academic performance significantly improved only when

school personnel conducted the intervention. Programmes thatencountered implementation problems were less effective thaninterventions without any apparent implementation problems

Programmes that were sequenced, active, focused explicit (SAFE)and well implemented were consistently successful, those thatdid not were not

No clear evidence that multi-component better than single—possibly because they were less often SAFE and wellimplemented

Current findings offer strongsupport for the value ofclassroom andschool-based SELinterventions when SAFEand well implemented

More research needed on keyfeatures ofimplementation

(Ekeland et al.,2004)

23 trials with children andyoung people of exerciseprogrammes thatmeasured the impact onself esteem

Self-esteem, variously measured Generally, the trials were small, and only one was assessed to have alow risk of bias.

13 trials compared exercise alone with no intervention. 8 wereincluded in the meta-analysis, and overall the results wereheteregeneous. One study with a low risk of bias showed astandardised mean difference (SMD) of 1.33 (95% CI 0.43 to2.23), while the SMD’s for the three studies with a moderaterisk of bias and the four studies with a high risk of bias was 0.21(95% CI 20.17 to 0.59) and 0.57 (95% CI 0.11 to 1.04),respectively. 12 trials compared exercise as part of acomprehensive programme with no intervention. Only 4provided data sufficient to calculate overall effects, and theresults indicate a moderate short-term difference in self-esteemin favour of the intervention [SMD 0.51 (95% CI 0.15 to 0.88)].

The results indicate thatexercise has positiveshort-term effects onself-esteem in childrenand young people. Sincethere are no knownnegative effects ofexercise and manypositive effects onphysical health, exercisemay be an importantmeasure in improvingchildren’s self-esteem.

These conclusions are basedon several smalllow-quality trials.

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i41



Table 2: Continued



(Farrington andTtofi, 2009)

89 reports describing53 different programme

evaluations on preventingbullying and victimization

Bullying or victimization had to be included as outcome measures.Bullying and victimization could be measured using self-reportquestionnaires, peer ratings, teacher ratings or observational data

School-based anti-bullying programmes are effective in reducingbullying and victimization (being bullied). On average, bullyingdecreased by 20–23% and victimization decreased by 17–20%.The effects were generally highest in the age-cohort designs andlowest in the randomized experiments. Firmer disciplinarymethods, parent training/meetings, video, co-operative groupwork and longer and more intense programmes were significantlyassociated with a decrease in victimization. Work with peers wasassociated with an increase in victimization

Results obtained so far inevaluations ofanti-bullying programmesare encouraging

New anti-bullyingprogrammes should bedesigned and tested basedon the key programmeelements and evaluationcomponents that found tobe most effective

(Gansle, 2005) 26 studies of school-basedprogrammes that focusedon anger or includedanger as a dependentvariable

Externalizing behaviour and angerInternalizing and anger, e.g. depression, shyness, anxietySocial skills, e.g. peer relations, social competence, self-controlBeliefs and attitudes, e.g. self-efficacy, self-esteem, locus of control,

intent to use non-violent behavioursAcademic, e.g. measured achievement, grades, academic

engagement and attendance

Across outcomes, the weighted mean effect size of the interventionspost-treatment was 0.31, which is modest but significant andsimilar to other behavioural approaches in related fields. Thelargest effects were found for anger and externalizingbehaviours, internalizing and social skills, with mean effect sizesof 0.54, 0.43 and 0.34, respectively. Longer interventions,focused on behavioural activities more effective

Socially focused interventions (e.g. generating responses, makingeye contact) worked better than self-focused interventions (e.g.recognizing and labelling emotions)

No differences for group comparisons by school setting, specialeducation status, entrance criteria or treatment agents

Anger management compareswell with other social andemotional educationinterventions

Interventions that are moremethodologicallyrigorous, are longer, aremore socially focused andinclude more behaviouralcomponents are morelikely to benefit students

(Garrard andLipsey, 2007)

36 studies on children andyoung people of conflictresolution education(CRE)

General construct of CRE used and studies selected that fell withinit

Targets anti-social behaviour promotes cooperation, empathy andrespect

Sometimes has secondary goals of emotional and social growth andwellbeing, critical thinking and improving school climate

Small amount of research and of uneven coverageMean effect size of 0.25 for the 36 studies was statistically significant

and represents improvements in problem behaviours that is ofpractical significance

Positive effect observed for different types of intervention, e.g.explicit direct skills instruction, or embedded in curriculum, orpeer mediation

Relatively small effect on younger students, 9 and under, andgreater effect on older

Majority of beneficial effects shown for shorter programmes, 2 h aweek for 15 h on average

CRE should be takenseriously as a tool fortreating school-basedanti-social behaviour

Particularly indicated forolder school students 9þand especially inadolescence

Shorter programmes can beeffective

Need to focus on whataspects of implementationmake programmes moreeffective

(Green et al.,2005)

8 reviews of interventions toimprove the social andemotional wellbeing ofprimary school-agedchildren

Problem-solving skillsAlternative thinking strategies and the promotion of self-esteem

reduction in aggressive behaviourBullying and violence preventionTeaching children to cope with stressful experiences and with

educational transitions

Intervention characteristics associated with more effective outcomes:promoting positive mental health rather than the prevention ofmental illness; continuous and long term; whole schoolapproach, focusing on school climate and environment ratherthan on individual change; opportunities for practice in range ofcontexts, addressed self-concept, self-esteem and coping skills;combining universal and targeted programmes

Schools have a role in mentalhealth promotion.Conclusions limited by shortduration of studies, lack ofdetail of interventions,identified outcomes andsocio-demographic data andthe relationship betweenprocesses and outcomes

i42K

.W

eare

an

dM

.N

ind



(Greenberg et al.,2001)

34 studies of preventionprogrammes preventingmental health disorders inschool-aged children

Universal: violence prevention, social/cognitive skillbuildingprogrammes, changing the school ecology, multi-component

Externalizing behaviour: anger, aggression, conduct disorderInternalizing: depression, anxiety, suicide, stress

Important and meaningful progress has been made in preventionresearch with children, families and schools during the last twodecades. There have been advances in the theory, design andevaluation of programmes, and there are a growing number ofprogrammes with documented efficacy of beneficial impact onthe reduction on psychiatric symptomology

Multi-year programmes more likely to foster enduring benefitsMore effective programmes start in the preschool and early

elementary yearsPreventive interventions are best directed at risk and protective

factors rather than at problem behaviours. Feasible andcost-effective to target multiple negative outcomes

Interventions should be aimed at multiple domains, changinginstitutions and environments as well as individuals

Prevention programmes that focus independently on the child’sbehaviour are not as effective as those that also ‘educate’ thechild and focus on teacher and family, home and school, and theneeds of schools and neighbourhood

There is no single programmecomponent that canprevent multiple high-riskbehaviours. A package ofcoordinated, collaborativestrategies andprogrammes is required ineach community

School ecology should be acentral focus ofintervention

To link to other communitycare systems and createsustainability, preventionprogrammes will need tobe integrated with systemsof treatment

(Hahn, 2007) 53 studies of the effects ofuniversal school-basedprogrammes to preventviolent and aggressivebehaviour

Reported or observed aggression or violenceConduct disorderMeasures of externalizing behaviourActing out (aggressive, impulsive, or delinquency, school records of

suspensions or disciplinary referrals)

For all grades combined, the median effect was a 15.0% relativereduction in violent behaviour among students who received theprogramme. Effects found at all school levels, Effectsdiminished slightly over time at the end of the intervention

All school programme intervention strategies (e.g. informational,cognitive/affective and social skills building) and programmefoci (e.g. disruptive or antisocial behaviour, bullying, datingviolence) similarly were associated with reduced violentbehaviour

No clear association for frequency or duration of programme

The number of studies in thissystematic review overalland the number of studiesat each grade level, ofadequate quality,consistency of effect, andeffect size, provide strongevidence that universalschool-based programmesare associated withdecreases inviolence-related outcomes.Beneficial results werefound at all school levelsexamined, frompre-kindergarten throughhigh school

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i43



Table 2: Continued



(Haney andDurlak, 1998)

102 studies covering 120programmes, whichindicated significantimprovement in children’sand adolescents’self-esteem andself-concept, andbehavioural, personalityand academic functioning

Self-esteem and self-concept Internalizing problems, externalizingproblems, mixed problems

Using measures of overt behaviour (using behavioural observationsor rating scales), personality functioning (e.g. self-reports ofanxiety or depression)

The weighted mean ES for all 120 interventions was 0.27, suggestinga modest overall impact

Interventions specifically focused on changing self-esteem andself-concept were significantly more effective (mean effectsize ¼ 0.57) than programmes focused on another target, such asbehaviour or social skills (0.10)

Treatment programmes were also more effective (0.47) than primaryprevention programmes (0.09) in changing self-esteem

Four variables emerged as significant predictors of self-esteemoutcomes: two methodological features (type of design andcontrol group), the use of a theoretical or empirical rationaleand the type of programme (treatment or prevention)

Programmes can influenceself-esteem andself-concept

Need to focus on self-esteemand self-conceptspecifically, and not hopeother focusedinterventions will impactindirectly

Future research needs toexamine the causalconnection betweenchanges occurring inself-esteem and otherareas of adjustment,assess intervention successfor different ethnic groupsand for children ofdifferent ages and sex,and determine thelong-term impact ofinterventions

(Harden et al.,2001)

345 studies concerned withmental health and youngpeople screened andmapped:

7 rigorous ones subjected tosystematic review

30 fairly rigorous onessubjected to in-depthreview

Very varied. Mental health, with specific focus on prevention ofsuicide and self-harm, and associated depression, and the promotionof self-esteem and coping strategies

Mapping exercise: most (72%) studies in schools, vast majority inUSA. About half on prevention: most common focus preventionof suicide, self-harm or behaviour problems. About half onpromoting positive mental health, e.g. self-esteem, self-conceptor coping skills. Half universal, rest focused on ‘at risk’.Barriers—major focus on psychological/individual rather thansocial/environmental factor. Quality of the studies was veryvariable: half judged ‘potentially sound’

7 systematic reviews and 30 in-depth studies had mixed positive/negative on the evidence of mental health promotion.Interventions to promote positive self-esteem limited effect, butmore effective if self-esteem is the main focus. Evidence onprevention of suicide and self-harm limited, some evidence thatdiscussing suicide may be harmful

12 studies of young people’s views: YP do not relate to the term‘mental health’, have sophisticated understandings of copingstrategies, a wide range of social and environmental concernsand find traditional health promotion irrelevant

Proceed with caution as theevidence for mentalhealth promotion is mixed

Involve and listen to youngpeople

If trying to developself-esteem, then focus onit specifically

Avoid universal suicideprevention education

(Hoagwood andErwin, 1997)

16 studies of effectiveness ofschool-based mentalhealth services forchildren

DepressionLocus of controlPeer acceptanceAggressionBehavioural problems

Three types of interventions found to have empirical support fortheir effectiveness. Cognitive-behavioural therapy especially fordepression has strong evidence. Social skills training has reasonableevidence. Teacher consultation (i.e. educating teachers andexamining the effects on pre-referral practices and problembehaviours) has promising evidence based on one intervention

Need to: investigateeffectiveness of withwider range of psychiatricdisorders; broaden therange of outcomes;examine the combinedeffectiveness of theseinterventions; link withhome-based interventions

i44K

.W

eare

an

dM

.N

ind



(Horowitz andGarber, 2006)

30 studies of programmesaimed at preventingdepression in children andadolescents

Measures of depression and anxiety Wide range in degree of success of programmesWeighted overall mean effect size post-intervention was 0.16, and at

follow-up was 0.11, i.e. small but significantMean effect size for selective prevention programmes was 0.30,

greater than effect size of universal prevention programmes(mean effect size 0.12). Probably because baseline depression inuniversal approach not high

Effects of indicated and selective programmes were not significantlydifferent. No clear effect of gender or age

Studies were in practice treatment (i.e. improvement in symptoms ofthe intervention group) rather than prevention (increase insymptoms in the control group but not in intervention.)—possibly due to most not having long follow-ups

Need long-term evaluationsPremature to abandon

universal programmes,but should focus onselected and targeted

(Kraag et al.,2006)

19 studies of schoolprogrammes targetingstress management orcoping skills in schoolchildren

Various measures of stress and mental health outcomesCategorized into four groups: symptoms of stress; social behaviour,

coping/social skills; self-efficacy/self-esteem

Overall effect size for the programmes was 21.51 [95% confidenceinterval (CI) 22.29, 20.73], indicating a positive effect. However,heterogeneity was significant (pb.001). Sensitivity analyses showedthat study quality and type of intervention were sources ofheterogeneity influencing the overall result (p valuesb.001). Theheterogeneity in quality may be associated with methodologicaldiversity and differences in outcome assessments, rather than varietyin treatment effect. Effect was calculated per intervention type, andpositive effects were found for stress symptoms with a pooled effectsize of 20.865 (95% CI: 21.229, 20.502) and for coping with apooled effect size of 23.493 (95% CI: 26.711, 20.275)

It is tentatively concludedthat school programmestargeting stressmanagement or copingskills are effective inreducing stress symptomsand enhancing copingskills

Future research should useclear quality criteria andstrive for less diversity inmethodology andoutcome assessment

(Maxwell et al.,2008)

20 studies of mental healthand emotional wellbeingin children and youngpeople

Not statedPrimary studies covered wide range of outcomes involving mental

health and emotional wellbeing

In schools, sustained broad-based mental health promotionprogrammes combined with more targeted behavioural andcognitive-behavioural therapy (CBT) for those children withidentifiable emotional wellbeing and mental health needs, offerevidence of a demonstrably effective approach. There is areasonably strong evidence base to support targeted work with bothparents and children

While systematic reviews areoften seen as offering theonly reliable basis onwhich programmaticdecisions should be made,the case has been madehere that broadening theevidence base may bebeneficial in providingevidence of practicallybased promising studiesintegrating evaluationfindings from recent localprogrammes, rather thanrelying too extensively onresearch conducted inother contexts

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i45



Table 2: Continued



(McCarthy andCarr, 2002)

4 studies of bullying inschools

Bully/victim problems, e.g. as, exposure to various physical, verbal,indirect, racial or sexual forms of bullying/harassment, variousforms of bullying other students, pro-bully and pro-victimattitudes, and the extent to those in the social environment areinformed about and react to the bullying

4 studies, between 1989 and 1997, 2 from Norway, 1 UK, 1 Canada.All were whole school

2 programmes were effective, 2 not. Programmes that wereimplemented completely, consistently in accordance with theguidelines and with external training, consultancy and supportwere effective. Those that were not, were not

Whole school bullyingprevention programmes caneffectively reduce bothreports of bulling and reportsof being bullied both in theshort and longer term. Theireffectiveness is determined bythe degree to whichprogramme integrity ismaintained and support,training and consultancyprovided

(Merry et al.,2004)

21 studies eligible forinclusion, 13 of which ofsufficient quality formeta-analysis, ofprogrammes that aim toprevent depression in theyoung

Primary outcomesPrevention of depression indicated by reduction in depressive

symptoms on pre-post-assessment (early intervention) orreduction in onset of depressive symptoms or disorder measuredby depression scores on a rating scales

Psychological interventions (skills based) were effective comparedwith non-intervention immediately after the programmes weredelivered with a significant reduction in scores on depressionrating scales for targeted [standardized mean difference (SMD)of 20.26 and a 95% confidence interval (CI) of 20.40 to20.13]. Some studies showed a decrease in depressive illnessover a year

Small but statistically significant changes in depression scoresfollowing psychologically based interventions such asstress-management or problem-solving skills (d ¼ 20.26, 95%CI: 20.40 to 20.13). The results also showed slightly betterlevels of effectiveness for targeted (RD ¼ 20.26, CI: 20.40 to0.13) than for universal programmes (d ¼ 20.21)

While small effect sizes were reported, these were neverthelessassociated with a significant reduction in depressive episodes

There was only one knowledge-based intervention and no evidencefor its effectiveness

Reports of effectiveness for boys and girls were contradictoryThe quality of many studies was poor, and only two studies made

allocation concealment explicit

The results from skills-basedinterventions arepromising.Knowledge-basedapproaches do not appearto work

It is likely that girls and boyswill respond differently tointerventions and a moredefinitive delineation ofgender specific responsesto interventions would behelpful

Secondary outcomes

(1) General adjustment(2) Academic/work function(3) Social adjustment(4) Cognitive style(5) Suicidal ideation/attempts(6)

i46K

.W

eare

an

dM

.N

ind



(Mytton et al.,2002)

44 studies of school-basedviolence preventionprogrammes for childrenidentified as at risk foraggressive behaviour

Violent injuries, observed or reported aggressive or violentbehaviours, and school or agency responses to aggressivebehaviours

For the 28 trials that assessed aggressive behaviours, the pooleddifference between study groups was 20.36 (95% confidenceinterval, 20.54 to 20.19) in favour of a reduction in aggressionwith intervention. For the nine trials that reported data on schoolor agency responses to aggression, the pooled difference was20.59 (95% confidence interval, 21.18 to 0.01). Subgroupanalyses suggested greater effectiveness in older students andwhen administered to mixed-sex groups rather than to boys alone

Programmes modestlyreduced both aggressivebehaviours and school oragency actions in response.Effects were similarregardless of whether theprogrammes focused ontraining in skills ofnon-response (e.g. conflictresolution or angercontrol) or on training insocial skills or socialcontext changes.School-based violenceprevention programmesmay produce reductions inaggressive and violentbehaviours in children whoalready exhibit suchbehaviour. These results,however, need to beconfirmed in large,high-quality trials

(Neil andChristensen,2007)

24 trials of 9 Australianschool-based preventionand early interventionprogrammes for anxietyand depression

Programmes that addressed symptoms of anxiety or depression in aschool context, or increased student resilience through thedevelopment of positive coping skills

Most programmes were based on cognitive behaviour therapy,interpersonal therapy or psycho-education. Six were universalinterventions, two were indicated programmes and one was atreatment programme. Most were associated with short-termimprovements or symptom reduction at follow-up

A number of schoolsprogrammes producepositive outcomes.However, even wellestablished programmesrequire further evaluation

(O’Mara et al.,2006)

145 studies of theeffectiveness ofinterventions to enhancevarious aspects ofself-concept

Programmes were included if they contained a measure ofself-concept or another related self-concept construct (e.g.self-esteem, self-efficacy), which could be either a global measure(e.g. self-esteem) or a specific domain (e.g. academicself-concept)

Overall, interventions on various aspects of self concept weresignificantly effective (d ¼ 0.51, 460 effect sizes). These effectsdo not systematically diminish over time

The largest mean effect size of all the moderator categories was forinterventions aimed at enhancing a specific self-concept facetand that also measured that specific self-conceptdomain(d ¼ 1.16)

Intervention effects were substantially larger for facets ofself-concept that were logically related to the intervention thanfor unrelated facets of self-concept

Self-concept is notuni-dimensional, it ismulti-dimensional

Interventions are moreeffective than previouslythought if we use amulti-dimensional modelof self concept

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i47



Table 2: Continued



(Park-Higgerson,2008)

26 studies of violenceprevention in schools

Externalizing, aggressive or violent behaviour (i.e. scores ofaggression, use of violence/violent or externalizing behaviour),empathy, impulse control, anger management

Overall, the intervention groups did not have significant effects inreducing aggression and violence when compared with thecontrol groups (ES ¼ 20.09, 95% CI ¼ 20.23 to 0.05, withheterogeneity p , 0.00001)

Comparing programmes with different features, there was nosignificant difference between interventions, althoughprogrammes that used non-theory-based interventions, focusedon at-risk and older children, and employed interventionspecialists had slightly stronger effects in reducing aggressionand violence. Interventions using a single approach had a mildpositive effect on decreasing aggressive and violent behaviour(effect size ¼ 20.15, 95% CI ¼ 20.29 to 20.02, p ¼ 0.03)

This meta-analysis did notfind any differentialeffects for 4 of the 5programmecharacteristics. This wascontrary to expectation,exemplifying thecomplexity of identifyingeffective programmestrategies

Small effect sizes, missingpretests, differences inoutcome focus, smallsample sizes andheterogeneity among theincluded studies may havecontributed to the lack ofsignificant findings forseveral of the programmecharacteristics

(Payton et al.,2008)

80 studies of indicated socialand emotional learning(SEL) programmes, i.e.that identify and workwith students displayingearly signs of behaviouralor emotional problems

SEL skills, attitudes towards self and others, positive socialbehaviour, conduct problems, emotional distress, academicperformance

Significant mean effect sizes ranging from 0.38 for improvedattitudes towards self, school and others to 0.77 for improvedsocial and emotional skills were achieved in all six outcomecategories studied. Participants in these indicated SELprogrammes received significantly greater benefits acrossoutcome categories than did participants in the control groups.Although the magnitude of these effects was generally lower atfollow-up, they were still significant in five out of the sixcategories (all except academic performance)

SEL interventionprogrammes for studentsexhibiting adjustment orlearning problems workedfor a wide range ofpresenting problems wereeffective when deliveredby either school ornon-school personnel, andhad significant outcomeswhether they includedonly one or multipleprogramme components.Such programmes shouldbe recommended aspotentially successfuloptions for promotingyouth wellbeing andadjustment both duringand after school hours

i48K

.W

eare

an

dM

.N

ind



(Reddy et al.,2009)

29 studies of school-basedprevention andintervention programmesfor children andadolescents at-risk of andwith emotionaldisturbance (ED)

Externalizing behaviour problems in the home or schoolInternalizing behaviour problems in the home or schoolSocial skills in the home or schoolAdaptive functioning at schoolActive engagement with task

The prevention and intervention programmes produced meanweighted ESs of 1.00 at post-test (unweighted ES of 1.49) and 1.35at follow-up (unweighted ES of 2.25). The intervention programmeswere most effective in improving externalizing behaviour problemsin the home (weighted ES of 2.46) and at school (weighted ES of1.27), social skills at school (weighted ES of 2.39), general academicskills (weighted ES of 1.78) and internalizing behaviour problems inthe home (weighted ES of 1.59)

Results offer initial supportfor the idea thatprevention andintervention programmesimplemented in schoolsare generally effective inalleviating the early onsetof emotional andbehavioural symptoms andreducing persistentsymptoms found amongchildren and adolescentswith ED

(Rones andHoagwood,2000)

47 studies of school-basedmental health programmes

Emotional and behavioural problemsDepressionConduct problemsStress managementSubstance use

There are a robust group of school-based mental health programmeswith evidence of an impact across a variety of emotional andbehavioural problems in children

Key features of successful programme implementation include (i)consistent implementation; (ii) inclusion of parents, teachers orpeers; (iii) use of multiple modalities—(e.g. the combination ofinformational presentations with cognitive and behavioural skilltraining); (iv) integration of programme content into generalclassroom curriculum and (v) developmentally appropriateprogramme components

Strategies that facilitated implementation included thecommunication of programme goals, rationale and componentsto school staff; the provision of feedback on programme effects;the development of plans to overcome barriers toimplementation and the specification of individualresponsibilities and multi-component programmes that targetedthe ecology of the child

Successful interventions included teacher training in classroommanagement techniques, parent training in child managementand child cognitive–social skills training

There are a strong set ofuniversal programmes,but we need moretargeted approaches, andmore programmes forolder students

We know something aboutthe factors that make forsuccessfulimplementation, but needto know more

Need more work onschool-based outcomes,e.g. attendance, andschool-related behaviour(as measured bydisciplinary referrals,suspensions andretention)

(Schachter et al.,2008)

40 evaluation studies of theeffects of school-basedinterventions on mentalhealth stigma

Mental health stigma, mainly around depression and schizophrenia Five limitations within the evidence base constituted barriers todrawing conclusive inferences about the effectiveness and harmsof school-based interventions: poor reporting quality, a dearth ofrandomized controlled trial evidence, poor methods quality forall research designs, considerable clinical heterogeneity andinconsistent or null results

There exists enoughsuggestive evidence toinform a future researchdirection, which takesbehavioural change as itsprimary outcome. Goldstandard research designsand methods are required

(Scheckner et al.,2002)

16 studies of interventions toaddress conflict, anger andaggression

Pro-social behaviour and skills, conflict resolution, angermanagement and resolution, reducing aggression

4 studies had strong effects sizes: Peacebuilders (1.49), SMART(students managing anger resolution together) (0.96), abibliotherapy programme in Israel based on reading and media(0.84) and first step to success (85). 4 others had moderate effectsizes, a further bibliotherapy programme (0.69), a CBTprogramme with aggressive boys (0.53), 2 programmes focusingon violence-free relationships (0.45) and a social cognitive groupintervention (0.39)

Programme impactsignificantly affected bythe use ofcognitive-behaviouralstrategies, multi-settingatmosphere (2 of the 4strong effect programmes),primary (elementaryschool) prevention, aqualified programmeleader and longer length ofprogramme

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i49



Table 2: Continued



(Shucksmithet al., 2007)

32 studies of targeted/indicated activities aimedat promoting the mentalwellbeing of children inprimary education

Emotional wellbeing (including happiness and confidence, and theopposite of depression) Psychological wellbeing (includingautonomy, problem solving, resilience, attentiveness/involvement) Social wellbeing (good relationships with others,and the opposite of conduct disorder, delinquency, interpersonalviolence and bullying)

CBT-based programmes for anxiety transferred successfully betweencountries

Brief targeted interventions for anxiety successful in groupsParent trainingþchild group CBT adds benefitsChildren of divorce and anxious school refusers benefitted from

CBT-based skills training. Depressive symptoms can beprevented by CBT plus social problem-solving

Peer mentoring of aggressive with non-aggressive children helpsdevelop prosocial skills and social standing

Gains from multi-component programmes are modest, given theircost. Social problem solving and the development of positivepeer relations have strongest programme effects

Two studies showed improved academic achievement as significantoutcomes of multi-component interventions

Complex longitudinal multi-component studies support the case forearly intervention with aggressive disruptive children and forproviding booster interventions

Recruitment and retention of parents is a major challengeParents may prefer targeted children to be treated at school rather

than home

Shifts in quality and focusacross the time period.1990s saw proliferation ofsmall-scale studies.Longer and larger studiesand evaluations are morerecent and long-termevidence thus still lacking

The majority of the includedstudies were US based.This may limit theapplicability of to othersettings

Early studies usedexperimental designs andclinical staff to deliversmall-scale interventionsto small samples ofchildren. Theirapplicability to real-lifeclassroom settings istherefore suspect

Later studies (almostexclusively in the USA)show massive sums ofmoney in large multicomponent longitudinaltrials. The results are veryuseful and are showingthe way towards thedesign of more effectiveinterventions, yet theremust be serious doubts asto the availability of suchresources within normaleducation budgets

i50K

.W

eare

an

dM

.N

ind



(Sklad et al.,2010)

75 studies of universalinterventions with schoolchildren that includedelements ofsocio-emotional learning

Social-emotional skills and attitudes (direct outcomes):Social-emotional skills (e.g. social competence, conflict resolution

skills):

(1) Positive self-image (e.g. self-efficacy, self-esteem)(2) Behavioural adjustment (second order effects):(3) Anti-social behaviour (e.g. aggressive behaviour, disruptive

behaviour)(4) Pro-social behaviour (e.g. altruistic behaviour, helping others)(5) Substance abuse (e.g. tobacco, alcohol and marijuana use)(6) Mental health disorders (e.g. internalizing symptoms, anxiety,

depression or suicidality)(7) Academic achievement on core subjects, such as reading and

math

Programmes had significant beneficial short-term effects academicachievement: 0.50, antisocial behaviour 20.48, mental disorder20.16, positive self-image 0.69, prosocial behaviour 0.59, socialskills 0.74, substance abuse -0.11

Weak, but statistically significant immediate effects on mentaldisorders and substance abuse

Long-term effects were significant for most outcomes, with theexception of positive self-perception. Academic achievement:0.25 antisocial behaviour 20.17, mental disorder 20.37, positiveself-image 0.08, prosocial behaviour 0.13, social skills 0.50substance abuse 20.20

The long-term largest beneficial effect was found for mentaldisorders, for which the effect was moderate in size and largerthan the immediate effect size—a ‘sleeper effect’:

All other long-term effect sizes, with the exception of the effect sizefor positive self-image, were also statistically significant, yet theirsizes were small. Positive self-image was the only outcomeparameter that showed no statistically significant effect ofprogrammes at the follow-up

Effect sizes at the follow-up were statistically significantlyheterogeneous for all outcome categories except academicachievement, pro-social behaviour and social skills. Theheterogeneity of effect sizes for the remaining four categories(antisocial behaviour, mental disorders, positive self-image andsubstance abuse) was high: 76–93%

Universal school-based SELprogrammes generallyhave positive effects onreduction or prevention ofmental problems anddisorders, and a numberof other desirableoutcomes. Theseoutcomes includeenhancement of socialand emotional skills;positive attitudes towardsself and others, promotionof academic achievementand prevention ofantisocial behaviour

The heterogeneity of theeffect sizes suggests thatthere are importantfactors or moderators thataffect the effectiveness ofprogrammes on analysedoutcome categories

Earlier research shows thatprogrammes showstronger effects on directoutcomes than onincidental or indirectoutcomes Therefore, toestablish an unbiasedgeneral estimate of theeffectiveness of SELprogrammes on anyparticular outcome, morestudies should be carriedout on multipurposeprogrammes. In addition,authors reporting theresults of targeted SELprogrammes should beencouraged to measureand report a widespectrum of outcomesrather than focusing on afew target outcomes

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i51



Table 2: Continued



(Stage andQuiroz, 1997)

99 studies that usedinterventions to decreasedisruptive classroombehaviour in publiceducation settings

Conduct disorder Externalizing behaviours A total of 223 effect sizes yielded a mean effect size of –0.78Studies using teacher rating scales were less likely to evidence

reductions in disruptive classroom behaviours compared tostudies using behavioural observation methodologies

Students treated in self-contained classrooms were more likely toevidence a reduction in disruptive classroom behaviour.

With the exclusion of studies using teacher rating scales, comparisonof treatment interventions showed no statistically reliabledifferences due to the large variability in the relativeeffectiveness for students treated

Results indicate thatinterventions to reducedisruptive classroombehaviour yield comparableresults to other meta-analyticstudies investigating theeffectiveness ofpsychotherapy for childrenand adolescents

(Tennant et al.,2007)

20 interventions to promotemental health and preventmental illness in childrenwhich included schools(plus 7 on parenting)

Parenting skillsAnxiety and depression preventionSelf-esteemViolence and aggression prevention

Included studies targeted a range of risk and protective factors, anda range of populations (including both parents and children). While,many lacked methodological rigour, overall, the evidence is stronglysuggestive of the effectiveness of a range of interventions inpromoting positive mental wellbeing, and reducing key risk factorsfor mental illness in children

A variety of programmeshave been shown to beeffective in promotingchildren’s mental health,albeit with modest effectsizes. Based on thisevidence, arguments areadvanced for thepreferential provision ofearly preventiveprogrammes

(Tilford et al.,1997)

Mental health of children andyoung people. Numbersnot stated

Not stated There is evidence for the effectiveness of classroom programmes onself-esteem, self-concept and coping skills of children andadolescents and for children coping with divorce

Children who have particular difficulties may benefit from targetedinterventions

Separate self-concept activities may have a value with minoritygroups rather than more general life skills approaches

All children need access to ahealth educationcurriculum.

Try structured programmeson self-concept andcoping skills for childrenand young people

Identify the needs of childrenexperiencing stressful lifeevents: their needs shouldbe met through theco-ordinated activities ofeducation, health andsocial care professionals

More evaluation anddissemination of findingsoutside the USA

More research onmulti-componentapproaches

More long-term projects andmore follow-up needed

i52K

.W

eare

an

dM

.N

ind



(Vreeman andCarroll, 2007)

26 studies of school-basedinterventions to decreasebullying

Bullying (bullying, victimization, aggressive behaviour and schoolresponses to violence)

Outcomes indirectly related to bullying (school achievement,perceived school safety, self-esteem and knowledge or attitudestoward bullying)

The types of interventions could be categorized as curriculum (10studies), multidisciplinary or ‘whole-school’ interventions (10studies), social skills groups (4 studies), mentoring (1 study) andsocial worker support (1 study)

Only 4 of the 10 curriculum studies showed decreased bullying, but3 of those 4 also showed no improvement in some populations

Of the 10 studies evaluating the whole-school approach, 7 revealeddecreased bullying, with younger children having fewer positiveeffects

Three of the social skills training studies showed no clear bullyingreduction

The mentoring study found decreased bullying for mentoredchildren

The study of increased school social workers found decreasedbullying, truancy, theft and drug use

Many school-basedinterventions directlyreduce bullying, withbetter results forinterventions that involvemultiple disciplines/wholeschool interventions.Curricular changes lessoften affect bullyingbehaviours

Outcomes indirectly relatedto bullying are notconsistently improved bythese interventions

(Waddell et al.,2007)

15 RCTs on programmes toprevent conduct disorder,anxiety and depression

Preventing conduct disorder (9 studies of 8 programmes)Preventing depression (4 studies)Anxiety (1 study)All three (1 study)

Ten RCTs demonstrated significant reductions in child symptomand/or diagnostic measures at follow-up. The most noteworthyprogrammes, for conduct disorder targeted at-risk children inthe early years using parent training or child social skills trainingfor anxiety, employed universal CBT training in school-agechildren; and for depression, targeted at-risk school-agechildren, also using CBT

Effect sizes for noteworthy programmes were modest butconsequential

Few Canadian studies and few that evaluated costs

Prevention programmes arepromising but replicationRCTs are needed todetermine effectivenessand cost-effectiveness inCanadian settings

Four programme types shouldbe priorities forreplication: targetedparent training and childsocial skills training forpreventing conductdisorder in children’searly years; and universaland targeted CBT forpreventing anxiety anddepression in children’sschool-age years

Conducting RCTs throughresearch-policypartnerships would enableimplementation inrealistic settings whileensuring rigorousevaluation

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i53



Table 2: Continued



(Wells et al.,2003)

17 studies of 16 interventionsto promote mental healthin schools

9 measured negative aspects of mental health—aggression, conductproblems or antisocial behaviour (5), depression or suicidaltendencies (4)

6 measured personal and interpersonal behaviours that underpinmental health—problem-solving (4) conflict resolution (1)emotional awareness (1)

4 measured aspects of positive mental health and these all focusedon self-concept or self-esteem

Most of the included studies were relatively small for investigationsof school health promotion interventions, involving ,500children in between one and six schools. Even among the mostrobust studies methodological flaws were common. Only threestudies took account of cluster design methodology in theanalysis

The most positive evidence of effectiveness was obtained forprogrammes that adopted a whole-school approach, wereimplemented continuously for more than a year, and were aimedat the promotion of mental health as opposed to the preventionof mental illness. Those that aimed to improve children’sbehaviour and were limited to the classroom were less likely tobe effective. The results of some studies were, however, atvariance with these generalizations, suggesting that otherunidentified factors may also be important in determiningsuccess

Universal school mentalhealth promotionprogrammes can beeffective and long-terminterventions that aim topromote the positivemental health of all pupilsand involve changes tothe school climate likelyto be more successfulthan brief class-basedmental illness preventionprogrammes

Optimum approach might bea combination ofuniversal and targetedapproaches

Methodological flaws in someof the studies indicate theneed for further researchand there is also a needfor robust studies of theseprogrammes outside theUSA

(Wilson et al.,2003)

172 studies of experimentaland quasi-experimentalstudies of school-basedprogrammes withoutcomes representingaggressive and/ordisruptive behaviour

The review included studies of any school-based programme forwhich aggressive behaviour was measured as an outcome variable

Effect sizes were 0.1 for universal interventions and 0.3 for targetedor indicated populations

Most studies were conducted on demonstration programmes; thefew studies of routine practice programmes showed muchsmaller effects

Programme effects did not vary greatly with the age, gender orethnic mix of the research samples

Interventions were generally more effective when they wereimplemented well and relatively intense, used one-on-oneformats, and were administered by teachers

Behavioural approaches and counselling showed the largest effects,followed by academic programmes and separate schools/classroom, social competence training with and withoutcognitive-behavioural components followed close behind andmultimodal and peer mediation programmes

Higher risk youth showed greater reductions in aggressivebehaviour, and poorly implemented programmes producedsmaller effects, and different types of programmes weregenerally similar in their effectiveness, other things equal

Though not representative ofroutine practice, thedemonstrationprogrammes yieldedencouraging evidenceabout what practiceprogrammes mightachieve under favourablecircumstances

A range of strategies work, solong as they are wellimplemented

Programmes are mosteffective in contextswhere the base rates ofaggressive behaviour arehigh enough formeaningful reduction tobe possible

i54K

.W

eare

an

dM

.N

ind



(Wilson andLipsey, 2006a)

73 research studies describedin 89 reports of universalschool based socialinformation processinginterventions onaggressive behaviour

Violence, aggression, fighting, person crimes, disruptive behaviourproblems, acting out, conduct disorder, externalizing problems

Students who participated in social information processing showedless aggressive and disruptive behaviour after treatment thenstudents who did not receive the programme. The overallweighted mean effect size was 0.21 which was statisticallysignificant, though not large. Nearly 80% of the effect sizevalues were positive

Children from low socioeconomic status families or from schoolswith a large proportion of low income students achieved greaterbenefit than students from higher socioeconomic statuscommunities. Studies with higher quality methods (e.g. thosewith random assignment or low attrition) did not produce better(or worse) outcomes than studies using less rigorous methods.

More intensely delivered programmes more effectiveResearch and demonstration programmes and those that had no

obvious implementation difficulties produced the largest effects.Programs delivered under routine circumstances were the leasteffective, independent of implementation quality, maybebecause they tended to be less intense

The overall mean effect sizeof 0.21 indicates thatuniversal social informationprocessing programs areeffective for reducingaggressive and disruptivebehaviour

(Wilson andLipsey, 2006b)

68 reports of 47 studies ofuniversal school-basedsocial informationprocessing interventionson aggressive behaviour

Aggressive behaviour, i.e. violence, aggression, fighting, personcrimes, disruptive behaviour, acting out, conduct disorder,externalizing problems

Clear positive programme effect. At-risk and behaviour problemstudents who participated in social information processingprograms showed less aggressive and disruptive behaviour aftertreatment than students who did not receive a programme. Theoverall weighted mean effect size was 0.26, which wasstatistically significant. Over 60% of the effect size values werepositive

Studies with greater amounts of attrition tended to show smallerprogramme impact than those with little attrition. There were nosignificant differences between experimental andquasi-experimental studies

Generally greater reductions in aggressive behaviour were found forhigher risk students. However, programmes for special educationstudents were significantly less effective than those for regulareducation students—they may have had problems that were tooserious to respond to relatively short interventions

The overall mean effect sizeof 0.26 indicates that targetedand indicated socialinformation processingprogrammes are effective forreducing aggressive anddisruptive behaviour in at riskand problem students

Continued

Men

tal

hea

lthp

rom

otio

nan

dp

rob

lemp

reventio

nin

scho

ols

i55



Table 2: Continued



(Wilson andLipsey, 2007)

249 experimental andquasi-experimental studiesof school-basedprogrammes withoutcomes representingaggressive and/ordisruptive behaviours

The review included studies of any school-based programme forwhich aggressive or disruptive behaviour was measured as anoutcome variable

The programmes were effective. Positive overall intervention effectswere found on aggressive and disruptive behaviour and otherrelevant outcomes (0.20–0.35)

The most common and most effective approaches were universalprogrammes (0.21) and targeted programmes (0.29) for selected/indicated children.

Multi-component/comprehensive programmes did not showsignificant effects (0.05, not statistically significant) which issurprising and counter-intuitive. It may be that their broad scopeis associated with some dilution of the intensity and focus of theprogrammes so that students have less engagement with themthan with the programmes in the universal and selected/indicated categories, and that as proportionately fewer of theprogrammes in this category involved the cognitively orientedtreatment modalities that were the most widely represented onesamong the universal and selected/indicated programmes

Effects for special schools or classrooms were modest butstatistically significant (0.11)—not clear whether relatively lowimpact is because all is being done already that can be, or thatproblems are so severe that these programmes cannot reachthem

Routine programmes delivered by teachers did not have significantlyworse effect sizes to those delivered by professionals

Different treatment modalities (e.g. behavioural, cognitive, socialskills) produced largely similar effects. Effects were larger forprogrammes that reported few implementation problems (0.32),with more frequent sessions (0.40 over a longer period of time(0.34), and those involving students at higher risk for aggressivebehaviour 0.21). For the universal programmes, the greatestbenefits appeared for younger students (027) and students fromeconomically disadvantaged backgrounds (0.21). For theselected/indicated programmes, it was students alreadyexhibiting problematic behaviour who showed the largest effects(0.21)

The mean effect sizes forthese types ofprogrammes represent adecrease in aggressive/disruptive behaviour thatis likely to be of practicalsignificance to schools

Schools seeking preventionprogrammes may choosefrom a range of effectiveprogrammes with someconfidence that whateverthey pick will be effective.Without the researcherinvolvement thatcharacterizes the greatmajority of programmesin this meta-analysis,schools might bewell-advised to givepriority to those that willbe easiest to implementwell in their settings

i56K

.W

eare

an

dM

.N

ind



non-school work in clinical contexts (8 reviews),which automatically meant a review was gradedas being of no more than moderate quality. All52 reviews used a stated and appropriate andcomprehensive search strategy; 51 provided ameta-analysis or narrative data synthesis; 47assessed the quality of studies and used theirassessment to guide results and 46 asked aclearly focused question.

Geographical location of the reviews

Just over half (27) of the reviews were carriedout by researchers based in the USA, the restwere from the UK (13), the Netherlands (3)Germany (2), Canada (2), Australia (2), NewZealand (1), Norway (1) and the Netherlandsand Belgium combined (1).

Transferability

Several reviews considered whether interven-tions that are generally USA based might betransferable to a different context or country (Afew trials had taken place in European contextsof evidence-based interventions that originatedelsewhere, and some trials in other countries(Diekstra, 2008b***). Such conclusions as couldbe drawn from the small evidence base werepositive. Bayer et al. identified some evidence-based programmes (Bayer et al., 2009)** thatwere potentially transferable to Australia.Diekstra found the overall effect size of USAand non-USA studies similar for the onlyoutcome on which comparison was possible,social and emotional skills (Diekstra,2008b)***, while Shucksmith et al. found thatCBT-based interventions targeted at reducinganxiety disorders had been transferred success-fully between several countries (Shucksmithet al., 2007)***.

Impact and effectiveness of interventions

Fifty of the 52 reviews came to a positive assess-ment of the evidence they reviewed, concludingthat one or more of the interventions had atleast small effects and/or were in some way‘effective’. The remaining two reviews wereinconclusive rather than negative, and citedmethodological weaknesses as the reason whythey could not come to firm conclusions.(Park-Higgerson et al., 2008*** on violence

prevention interventions and Schachter et al.,2008* on mental health stigma interventions.)

Only four minor examples of apparentadverse effects were reported across hundredsof interventions reviewed. All were small effectsand two were concerned with apparentincreases in bullying after interventions, whichlargely appeared to be connected with the useof peer groups for children who bullied (Adiet al., 2007b***; Shucksmith et al., 2007***;Blank et al., 2009***).

Overall effects

Interventions reviewed had wide-ranging ben-eficial effects on individual children and youngpeople, on classrooms, families and commu-nities and on an array of mental health, social,emotional and educational outcomes.

‘Internalizing’ mental health problems

Nineteen reviews included ‘internalizing’ mentalhealth problems and disorders such asdepression and anxiety among the outcomesthey examined. All concluded that the overallimpact of work was positive. Where reviewsused numbers to summarize effects, the impactappeared on the whole to be small to modest.Focusing only on the 9 reviews that were con-cerned only with work in schools (10 othersincluded interventions in clinical contexts) 3showed ES of 0.10–0.50 (Payton et al., 2008***;Reddy et al., 2009***; Sklad et al., 2010***), andone (Browne et al., 2004***) suggested modestto large impacts with ES of 0.41–1.70. Evidencefrom moderate quality reviews, most of whichincluded work in clinical as well as school con-texts, was consistent with these positive results,with the four reviews that enumerated resultsshowing effect sizes that varied from small tolarge, between 0.16 and 0.93 (Durlak and Wells,1997**; Haney and Durlak, 1998**; Merry et al.,2004**; Horowitz and Garber, 2006**)The other11 reviews that did not provide a numericalset of results also claimed that the impact ofinterventions on internalizing mental healthproblems was positive.

The impact on higher risk children was gener-ally consistently shown to be higher than thaton children with milder problems, and quitestrong, with average ES of �1.00, rising to 2.46for some specific selective interventions andmeasures (Horowitz and Garber, 2006**;


by guest on Decem


Dow

nloaded from


Browne et al., 2004***; Reddy et al., 2009***;Payton et al., 2008***).

Positive mental health and wellbeing and socialand emotional learning (SEL)

The impact on positive mental, emotional andsocial health and wellbeing in general showedpositive and small to moderate effects of inter-ventions, with ES of 0.15–0.37 (Adi et al.,2007a)***. Durlak and Weissberg (Durlak andWeissberg, 2007)*** and Durlak et al. (Durlaket al., 2011)*** both found that well-implemented SEL interventions had mean ES of0.24–0.35, and Durlak et al. calculated a grandstudy-level mean ES of 0.28 for 207 SEL inter-ventions (Durlak et al., 2011)***. Three otherreviews showed impacts on social and emotionalskills and competences to be positive, and mod-erate to strong effects (ES 0.5–1.49) (Catalanoet al., 2002***; Scheckner et al., 2002***;Berkowitz and Bier, 2007***). Impacts on self-esteem and self-confidence were consistentlyshown to be moderate across a range of high-quality reviews, with ES of 0. 34–0.69 across fivereviews (Haney and Durlak, 1998**; Ekelandet al., 2004***; O’Mara et al., 2006***; Durlakand Weissberg, 2007***; Sklad et al., 2010***).

‘Externalizing’ problems: violence, bullying,conflict and anger

Fifteen reviews addressed were of interventionswhich aimed to reduce and prevent violence,bullying, conflict and anger, using various over-lapping terminologies. Ten focused predomi-nantly on aggression, violence and conflictresolution, four focused predominantly on bul-lying, one focused on anger and one focused ondisruptive classroom behaviour.

Again, the impact on universal populationswas positive and small (ES 0.1 on average) butgenerally markedly stronger for high-risk chil-dren (ES 0.21–0.35 on average) (Catalano et al.,2002***; Mytton et al., 2002***; Scheckneret al., 2002***; Wilson et al., 2003***; Wilsonand Lipsey, 2006a***; Adi et al., 2007b***;Garrard and Lipsey, 2007***; Hanh et al.,2007**; Blank et al., 2009***; Farrington andTtofi, 2009***). Impact was generally strongerfor older students than younger (Farrington andTtofi, 2009***). Cognitive–behavioural inter-ventions also consistently showed a larger effectthan average with an ES of 0.5 (Beelman and

Losel, 2006**; Shucksmith, et al., 2007***).Targeting children who have violent or bullyingbehaviour, and especially carrying out peer-based work with them, in which difficult chil-dren work together, generally had an adverseeffect, with more bullying and victimizationresulting (Shucksmith et al., 2007***; Farringtonand Ttofi, 2009***).

Attitudes to school and academic achievements

Four studies assessed the impact of variousinterventions on aspects of children’s behaviourand attitudes towards school and reported. ESrelating to commitment to schooling that weresmall to moderate (ES 0.14–0.6) (Catalanoet al., 2002***; Berkowitz and Bier, 2007***;Durlak and Weissberg, 2007***; Sklad et al.,2010***) and of a similar magnitude forachievement in test scores and school grades(ES 0.11–0.5) (Durlak and Weissberg, 2007***;Sklad et al., 2010***; Durlak et al., 2011***).

Classrooms and families

One study (Durlak and Weissberg, 2007)***addressed the impact of various SEL interven-tions on surrounding environments, and foundpositive results (ES 0.34 for family environ-ments; 0.78 for classroom environments).

Emergent themes across reviews, including thecharacteristics of more effective interventions

Although the reviews were very heterogeneous,covering many different interventions, issues,topics and populations, undertaken across a20-year period and of varied quality, there wasin practice a considerable overlap betweenthem, with some key interventions cropping uptime again. However, the apparent impact ofmost interventions was variable, leading to theconclusion that the effectiveness of any inter-vention cannot be relied upon. Most interven-tions only worked sometimes, some did notwork at all, and some were considerably moreeffective than average in some circumstances.There is clearly more to being effective thansimply carrying out an intervention, even if welldesigned. The more recent reviews, particularly,recognized this problem and most of themincluded an analysis of the characteristics ofinterventions that appeared to be linked witheffectiveness. It is therefore possible to make


by guest on Decem


Dow

nloaded from


some tentative data synthesis of the key themesthat emerged, many of which are concernedwith the features of effective interventions.

The balance between targeted and universalapproaches

Most of the reviews (48) were of universalapproaches and the 46 that were able to come toconclusive results concluded that universalapproaches they reviewed had a positive impact.Fourteen reviews looked at both targeted and uni-versal approaches, and nine of them were able tocome to comparative conclusions. As we havealready suggested, interventions across the wholerange of outcomes assessed were consistentlyshown to have a more dramatic effect on higherrisk children (Haney and Durlak, 1998**; Wilsonet al., 2003***; Beelman and Losel, 2006**;Wilson and Lipsey, 2006b***; Adi et al., 2007b***;Waddell et al., 2007**; Diekstra, 2008b***;Horowitz and Garber, 2006**; Park-Higgersonet al., 2008***). Two reviews hypothesized thatthis was due to the ‘ceiling effect’ with populationswithout overt problems not having the same scopefor improvement. (Horowitz and Garber, 2006**;Adi et al., 2007b***).

Two reviews concluded that some redressingof the balance was needed, with a greateremphasis needing to be placed on targeted(Rones and Hoagwood, 2000**; Beelman andLosel, 2006**). However, three reviews con-cluded that universal approaches provided amore effective context for working with stu-dents with problems than targeted or indicatedalone (Wells et al., 2003***; Browne et al.,2004***; Diekstra, 2008a*).

Adi et al. addressed this issue specifically andconcluded that both universal and targetedapproaches have their place, and appear to bestronger in combination found insufficient evi-dence to make recommendations relating to theoptimum balance of universal and targetedapproaches (Adi et al., 2007a)***.

Several reviews addressed the question ofdifferential impact but many of the findingswere inconclusive, or not replicated in otherstudies, and no substantial or clear resultsemerged: Adi et al., in a large and recent reviewof the whole field, found no trials to showdifferential effects according to age, gender,ethnic or social groups (Adi et al., 2007a)***.

Develop skills and competences

Improvements in skills were an outcomeexplored in 10 reviews across a wide range ofissues and there was consensus among them thatteaching skills and developing competence is acentral part of any comprehensive and effectiveintervention (Catalano et al., 2002***; Berkowitzand Bier, 2007***; Durlak and Weissberg,2007***; Shucksmith et al., 2007***) in a recentreview of targeted approaches to mental healthand wellbeing in primary schools, which took abroad view of the field concluded that the morecomplex and effective interventions, despitetheir different branding, offered a very similarmix of CBT and social skills training for children,training of parents and teachers in appropriatereinforcement and better methods of discipline,and that this mix was very similar whatever theproblem or diagnosis, for internalizing problems,such as depression and anxiety, as well as forexternalizing behaviours, such as conduct dis-orders. Looking at several of wide range ofissues explored by the various good qualityreviews, the acquisition of social and emotionalskills and competences was associated with awide range of specific outcomes including posi-tive youth development (Catalano et al.,2002***; Durlak and Weissberg, 2007***) char-acter education (Berkowitz and Bier, 2007)*** areduction in depression and anxiety (Shucksmithet al., 2007***; Waddell et al., 2007**;Blank et al., 2009***), conduct disorders(Shucksmith et al., 2007***; Waddell et al.,2007**) violence (Mytton et al., 2002)***, bully-ing (Farrington and Ttofi, 2009***) conflict(Garrard and Lipsey, 2007***; Waddell et al.,2007**) and anger (Gansle, 2005***).

Three reviews concluded that the teaching ofskills had more, and longer term, impact whenmental health issues were integrated into thegeneral classroom curriculum than when theskills were focused on in isolation and thatinterventions covering social problem solving,social awareness and emotional literacy, inwhich teachers reinforce the classroom curricu-lum in all interactions with children, were par-ticularly effective (Rones and Hoagwood,2000**; Adi et al., 2007a***; Berkowitz andBier, 2007***).

Six reviews nominated CBT as a particularlyeffective approach, suggesting that it impactedon anti-social behaviour (Beelman and Losel,2006**), violence and aggression (Wilson et al.,


by guest on Decem


Dow

nloaded from


2003***), conduct disorder (Waddell et al.,2007**), pro-social behaviour and skills, conflictresolution, anger management and resolution,and reduced aggression (Scheckner et al., 2002,2007***) and anxiety and depression (Neil andChristensen, 2007**; Shucksmith et al., 2007***;Waddell et al., 2007**).

Teaching methodologies

Only one review (Hahn, 2007**) concludedthat the teaching methodologies employedby interventions made no difference: thebalance of the evidence suggested that thechoice of teaching strategies and methodswas highly influential over interventions’effectiveness.

Five reviews suggested the need for a positiveand holistic approach. Greenberg et al.(Greenberg et al., 2001)** and Wells et al.(Wells et al., 2003)*** concluded that behav-ioural strategies on their own were unlikely tobe effective, and Merry et al. (Merry et al.,2004)** came to a similar conclusion about theineffectiveness of information only strategies.These three reviews concluded that interven-tions need also to ‘educate’ the child throughimpacting on attitudes, values, feelings andbehaviour. Three reviews concluded that inter-ventions were more effective if they werepositive rather than fear or problem based(Wells et al., 2003***; Browne et al., 2004***;Green et al., 2005*) while Browne et al.(Browne et al., 2004)*** felt that interventionswere more effective if they addressed the needsof the whole child, rather than just seeing themas a ‘problem’.

Six reviews looked at the issues of teachingmethods. Five concluded that more effectiveinterventions used active rather than didacticteaching methods, employing interactivemethods such as games, simulations and smallgroup work (Browne et al., 2004***, Berkowitzand Bier, 2007***; Durlak and Weissberg,2007***; Diekstra, 2008a*; Durlak et al.,2011***). Two reviews suggested that usingmultiple modalities (a range of integrated andcoordinated methods, groups, levels of interven-tion, one-to-one and whole-class work) wasmore effective than using just one or twoapproaches (Rones and Hoagwood, 2000**;Browne et al., 2004***).

Whole-school/multi-component approaches

Five reviews, across a wide topic range, con-cluded that it is necessary for effectiveness tomove beyond an individual, classroom and cur-riculum focus alone, and embed such workwithin a whole-school, complex, multi-component approach involving a wide range ofpeople, agencies, methods and levels of inter-vention, and mobilizing the whole school as anorganization (Catalano et al., 2002***; Wellset al., 2003***; Adi et al., 2007a***,b***).Vreeman and Carroll found only 4 of the 10curriculum and social skills based studies theyreviewed showed decreased bullying, and 3 ofthose 4 also showed no improvement in somepopulations, while of 10 studies evaluating thewhole-school approach, 7 revealed decreasedbullying across the board (Vreeman andCarroll, 2007)***. Catalano et al. found that, indeveloping pro-social behaviour and social com-petence, the more components an interventioncovered the better. However, two recentreviews have come to different conclusions(Catalano et al., 2002)***. Wilson and Lipseyreviewing interventions to prevent violence(Wilson and Lipsey, 2007)*** and Durlak et al.reviewing interventions to develop SEL, bothfound that multi-component interventions didnot show significant effects compared withinterventions which only involved one aspect ofschool life, a finding which the reviewers foundcounter-intuitive and contrary to their expec-tations from previous evidence (Durlak et al.,2011)***. Both hypothesized that it may be thatthe broad scope of some of the more recentmulti-component interventions is associatedwith some dilution of the intensity and focusand with weaker implementation, so that stu-dents have less engagement with these interven-tions. The importance of the quality of theimplementation is an issue we will return tolater.

School ethos and culture

Six reviews discussed the importance of school‘environments’ and efforts to change them topromote mental health (Durlak and Wells,1997**; Greenberg et al., 2001**; Catalanoet al., 2002**; Wilson et al., 2003***). Somereviews explored this issue in more detail, for-mulating the concept of school ‘ethos’ andculture to describe the underlying values and


by guest on Decem


Dow

nloaded from


attitudes that the school represents, particularlyin relation to the way staff and students treatone another, the development of bondsbetween youth and adults, and increased oppor-tunities and recognition for youth participationin positive social activities. Two reviews ofprimary schools attempt to promote mentalhealth and wellbeing, both in general (Adiet al., 2007a***) and in relation to the preven-tion of violence (Adi et al., 2007b***) found theresults of interventions to influence and changeschool ethos and culture were positive and verypromising for future research. One review(Greenberg et al., 2001**), discussing evidenceon preventing mental disorder, concluded thatschool ecology should in future be a centralfocus of intervention.

Agents of transmission

The interventions analysed in the variousreviews were transmitted by many differentagents. Eleven reviews explored the issue ofeffective leadership and which agents weremore effective.

Scheckner et al., looking at interventions thatpromote pro-social behaviour and skills, foundthat intervention impact was significantlyaffected by having a qualified interventionleader (Scheckner et al., 2002)***, while severalreviews (e.g. Adi et al., 2007a***,b***;Berkowitz and Bier, 2007***; Diekstra, 2008a*)commented on the need for extensive and inten-sive training for those involved in leadership.

Many of the early interventions used clinicalstaff to deliver small-scale demonstration pro-grammes using experimental designs and invol-ving small samples of children, usually focusedon specific issues and with short-term evalu-ation. This approach has remained common fortargeted interventions. Shucksmith et al.(Shucksmith et al., 2007)*** concluded that itwas particularly appropriate when the interven-tions are starting out. Adi et al. suggested usingspecialist staff was effective in short-term stressand coping interventions (Adi et al., 2007a)***,while Blank et al. suggested it was useful forinterventions to address for anxiety anddepression (Blank et al., 2009)***. However,Shucksmith et al. (Shucksmith et al., 2007)***concluded that this use of specialist staff wasunsustainable in the longer term and forlarger-scale and universal interventions. So,more recently there has been a shift away from

specialist staff, both across and, over time,within interventions to implement interventionsin real-life circumstances, using those routinelyinvolved in the life of the school, such as tea-chers, parents and sometimes peers.

There was contradictory evidence on theabsolute effectiveness of teachers comparedwith specialist staff. Three reviews suggestedthat teachers are not as effective as specialiststaff (Wilson et al., 2003***; Beelman andLosel, 2006**; Wilson and Lipsey, 2006a***)and hypothesized that this was because inter-ventions delivered under routine circumstanceswere less intense. However, three other reviewsconcluded that teachers can be as effective asspecialists (Adi et al., 2007a***; Wilson andLipsey, 2007***; Diekstra 2008a*). Wilson andLipsey (Wilson and Lipsey, 2007)***, Durlaket al. (Durlak et al., 2011)*** and Diekstra(Diekstra, 2008a)*concluded that it is particu-larly important that teachers are involved ifinterventions are to get to the heart of theschool process. Durlak et al. (Durlak et al.,2011)*** and Diekstra (Diekstra 2008a) con-cluded that only when school staff conduct theintervention does student academic perform-ance improve significantly and mental healthstart to impact on school culture—possiblybecause school staff are involved in bothaspects of school life and can bring it alltogether.

Some reviews included interventions whichinvolved peer work. The evidence on its effec-tiveness was mixed. Six reviews found thatpeers can be an effective and significant part ofsome types of mental health interventions.Rones and Hoagwood (Rones and Hoagwood,2000)**, Adi et al. (Adi et al., 2007a)*** andGarrard and Lipsey (Garrard and Lipsey,2007)*** all reported reasonable evidence thatpeer mediation in conflict resolution is effectivein the short term, and Blank et al. found it to beeffective in the longer term (Blank et al.,2009)***. Browne (Browne et al., 2004)*** andShucksmith et al. (Shucksmith et al., 2007)***found some evidence that peer norming(putting children with problems with thosewithout) has at least short-term modest impactson the mental health of children with problems.However, Farrington and Ttofi (Farrington andTtofi, 2009)*** and Shucksmith et al.,(Shucksmith et al., 2007)*** found that peerwork which is only carried out with childrenwho bully increased their subsequent bullying


by guest on Decem


Dow

nloaded from


and victimization of other children, with bully-ing children reinforcing one another’s attitudesand behaviours.

Family and community involvement

Durlak et al. found 64% of the positive youthdevelopment interventions attempted some typeof microsystemic or mesosystemic change invol-ving schools, families or community-basedorganizations (Durlak and Weissberg, 2007)***.Four reviews all looking at broad issues such aspositive youth development and mental health(Greenberg et al., 2001**; Catalano et al.,2002***; Browne et al., 2004***; Diekstra,2008a*) concluded that such engagement withand support from families and communities ishelpful, with Greenberg (Greenberg et al.,2001)** suggesting that it is more effective thanprevention programmes which focus only andindependently on the child’s behaviour. BothBrowne (Browne et al., 2004)*** andGreenberg (Greenberg et al., 2001)** commen-ted on the importance of embedding interven-tions within multi-disciplinary teams andcommunities to provide support.

The involvement of parents was nominated by10 reviews as a key component of effectivemulti-component interventions. Parental invol-vement was reported as increasing, effectivenessfor pro-social youth development (Catalanoet al., 2002***; Durlak et al., 2007***), universalinterventions to promote mental health (Wellset al., 2003***; Adi et al., 2007a***), stress andcoping interventions (Adi et al., 2007a***),interventions to reduce violence and bullying(Adi et al., 2007b***; Blank et al., 2009***;Farrington and Ttofi, 2009***) targetedapproaches to prevent mental disorders(Greenberg et al., 2001**; Shucksmith et al.,2007***) and conduct disorder (Waddell et al.,2007**). Shucksmith et al. suggested that this isbecause, when involved, parents can supportand reinforce at home the messages children arelearning at school (Shucksmith et al., 2007)***.

Durlak and Weissberg suggested that theeffect is two-way: looking at positive youthdevelopment interventions that attempted tochange schools, families and community-basedorganizations they found some statistically sig-nificant changes in families and communities asa result of school-based interventions, rangingfrom modest to large effects (Durlak andWeissberg, 2007)***.

Age and stage

Eleven reviews explored the issue of age inrelation to the impact of interventions. Fiveconcluded that it is generally important to startinterventions early, with younger children(Durlak and Wells, 1997**; Greenberg et al.,2001**; Browne et al., 2004**; Shucksmith et al.,2007***; Waddell et al., 2007**), although twosuggested that the age of introduction may notbe crucial (Adi et al., 2007a***; Durlak et al.,2011***) with reviews of interventions toprevent bullying, conflict and violence, targetingolder students particularly suggesting thatworking with older students is more effective(Mytton et al., 2002***; Garrard and Lipsey,2007***; Farrington and Ttofi, 2009***).*****

However, Blank et al. (Blank et al., 2009)***commented that there is little work for onwhich to base comparisons. In their rare reviewof work for students over 11 years, they foundfew studies with older children, and that thosefew were one-off interventions in local contexts,focused on improving behaviour, reducing vio-lence and bullying and developing pro-socialbehaviour and skills, without involvement ofparents or the wider school.

The balance of evidence pointed to startingearly, with well designed and implementedinterventions and then continuing with olderstudents. Three reviews concluded that therewas evidence both for intensive interventions inthe early years and for supportive ‘booster’ ses-sions later. (Browne et al., 2004***; Shucksmithet al., 2007***; Diekstra, 2008a*).

Length and intensity of interventionsand of their evaluations

Fifteen reviews produced results that related tohow long interventions should last and howintensive they should be to be effective. Someof the evidence on the associations betweenimpact and duration was inconclusive (Hahn,2007**; Blank et al., 2009***), but overall someclear patterns emerged.

None of the reviews concluded that singlebrief interventions have any worthwhile role.Three reviews produced some evidence insupport of short-term interventions (8–10weeks) for specific and mild problems for con-flict resolution (Garrard and Lipsey, 2007***;Adi et al., 2007a***) and minor anxiety and


by guest on Decem


Dow

nloaded from


emotional disorders (Shucksmith et al.,2007***).

However, the majority of reviews found inter-ventions of at least 9 months to a year to bemore effective, especially in broad areas and/orin response to more severe problems. Longerand more intense interventions appeared to bemore effective than brief ones for positivemental health (Wells et al., 2003***; Greenet al., 2005*), positive youth behaviour(Catalano et al., 2002)*** preventing violenceand bullying (Scheckner et al., 2002***; Adiet al., 2007b***; Farrington and Ttofi, 2009***),anger (Scheckner et al., 2002***; Gansle,2005***) and preventing mental disorders(Greenberg et al., 2001**).

Most of the evaluations of the includedstudies took place immediately after the inter-vention and as we have seen there was reason-able evidence of at least small to moderateshort-term impact. But with recent largerstudies, some long-term effects were emerging.Generally, effects gradually decreased in thelong-term but remained significant (Horowitzand Garber, 2006**; Diekstra, 2008b***).

High-quality implementation

Eleven reviews commented on the issue of theimpact of intervention quality on effectiveness,with all concluding that it was an importantdeterminant. Wilson et al. (Wilson et al.,2003)***, Wilson and Lipsey (Wilson andLipsey, 2006a)*** and Durlak et al. (Durlaket al., 2011)*** found that interventions thathad no obvious implementation difficulties pro-duced the larger effects than those with difficul-ties. Wilson and Lipsey (Wilson and Lipsey,2006a***) concluded that schools seeking pre-vention interventions might be well-advised togive priority to those that will be easiest toimplement well in their settings. SimilarlyBerkowitz and Bier (Berkowitz and Bier,2007)***, reviewing range of interventions oncharacter education, concluded that there was aclear trend for complete and accurateimplementation to result in more outcomeeffectiveness than incomplete or inaccurateimplementation. Two concluded that implemen-tation was of overriding importance. Durlakand Weissberg (Durlak and Weissberg,2007)*** and Durlak et al. (Durlak et al., 2011)concluded that interventions were not effectiveat all if they were based only on loose

guidelines and broad principles, but also neededhigh levels of intensity, consistency, clarity andprogramme fidelity.

Some of the key interrelated features of high-quality implementation identified by high-quality implementation were:

† a sound theoretical base (Browne et al.,2004)*** explicitness—specific, well-definedgoals and rationale, communicated effectivelyto staff and leaders through thorough trainingand linked explicitly with the interventioncomponents (McCarthy and Carr, 2002**;Rones and Hoagwood, 2000**; Browne et al.,2004***; Sklad et al., 2010***);

† a direct, intense and explicit focus on thedesired outcome rather than using a differentfocus and hoping for indirect effects (Hardenet al., 2001**; O’Mara et al., 2006***; Durlakand Weissberg, 2007***; Durlak et al.,2011***);

† explicit guidelines, possibly manualized(McCarthy and Carr, 2002**; Durlak andWeissberg, 2007***; Durlak et al., 2011***)thorough training and quality control consist-ent staffing and the specification of individualresponsibilities (Browne et al., 2004***);

† complete and accurate implementation.(Berkowitz and Bier, 2007; McCarthy andCarr, 2002**; Catalano et al., 2002; Durlakand Weissberg, 2007***; Durlak et al.,2011)***.

DISCUSSION

The impact of work on mental health in schools

This review endorses the importance of work topromote mental health and prevent mentalhealth problems in schools. It confirms the find-ings of earlier reviews and recent overviews(e.g. Jenkins and Barry, 2007) that over the last25 years a strong group of school mental healthprogrammes have emerged with clear andrepeated evidence of positive impact. Therewere very few examples of adverse effects,which is reassuring in the face of some concernsthat have been expressed about the ‘dangers’ ofwork in this area (e.g. Ecclestone and Hayes,2009). The cumulative evidence is of impactthat is small to moderate in statistical terms andstronger in the short than long term. In terms ofspecific impacts, there was a small to moderateimpact of universal interventions on positive


by guest on Decem


Dow

nloaded from


mental health, mental health problems and dis-orders, violence and bullying and pro-social be-haviour. In all these areas, the effects ofinterventions was dramatically higher, and quitestrong, when targeted at higher risk children,and markedly stronger than average for somespecific interventions and on some measures.The impact of interventions on social andemotional skills and competences was moderateto strong. Impacts on commitment to schoolingand academic achievements were small to mod-erate, and on family and classroom environ-ments, moderate.

The effects may be ‘small to moderate’ in stat-istical terms but they represent effects that inthe real-world are important and relativelylarge. As two reviewers, Durlak and Wells(Durlak and Wells, 1997)*** and Stage andQuiroz (Stage and Quiroz, 1997)** commented,these are outcomes similar to, or higher in, mag-nitude than those obtained by many other estab-lished preventive and treatment interventions inthe social sciences and medicine. Durlak andWells estimated that in practical terms, theaverage participant in the primary preventioninterventions they reviewed surpassed the per-formance of between 59 and 82% of those in acontrol group, and outcomes reflect an 8–46%difference in success rates favouring preventiongroups (Durlak and Wells, 1997)***. Durlaket al. calculated that the effect sizes from the207 SEL interventions they reviewed averagedout to an 11% improvement in achievementtests, a 25% improvement in social andemotional skills and a 10% decrease in class-room misbehaviour, anxiety and depression,effects which held up for at least 6 months afterthe intervention (Durlak et al., 2011)***. It istherefore important that work on mental healthpromotion and problem prevention in schoolsbe endorsed, continued and expanded.

The characteristics of effective interventions

Reviews showed considerable variation ofeffects. It is clear that many types of interven-tions can be effective, sometimes strikingly so,but that their effectiveness cannot be relied on.Since information about the variability of effectsbecame clear, the discussion has moved awayfrom simple impact to focus more on the some-times subtle characteristics of effective interven-tions. The findings about effective characteristicsthat emerge from more recent reviews over the

last 10 years contain some important keymessages, which may involve redressing someimbalances in policy and practice.

Linking with academic learning

Several studies showed generally positiveimpacts of various interventions on aspects ofchildren’s learning, behaviour and attitudestowards school, such as achievement in testscores and school grades, commitment to schooland school attendance. They also showedgreater and longer term impact of interventionswhen mental health issues were integrated intothe general classroom curriculum than focusedon in isolation. It is increasingly accepted bythose keen to promote mental health in schoolsthat linking this activity with the goals ofschooling is vital, both to improve the impact ofinterventions and to ensure that hard pressedschools are to justify a concern with mentalhealth to sceptical staff, parents, governors andfunding bodies (Zins et al., 2004). Thoseinvolved school mental health at every levelneed to ensure they work closely education anddemonstrate how work to improve the mentalhealth of students can benefit what schools seeas their core business: academic learning andachievement, school attendance and behaviourfor learning.

Balancing universal and targeted interventions

Within mental health promotion and problemprevention generally the emphasis has longbeen on a positive approach to mental health inschools, and universal approaches that targeteveryone. This review generally endorses thisfocus on the universal, but with some importantcaveats. It was clear that universal approacheson their own were not as effective as those thatadded in a robust targeted component, and thatinterventions had a more dramatic effect onhigher risk children. It would appear fromcurrent evidence that the best informedapproach is to include both universal and tar-geted approaches, which appear to be strongerin combination, although the exact balance hasyet to be determined. It may well be thatmental health promotion in schools needs toredress the balance somewhat in favour of morework on targeted approaches, while continuingto embed and integrate them within a robustuniversal approach.


by guest on Decem


Dow

nloaded from


Starting early and continuing with olderstudents

Evidence from this review suggests that it isbest to start early and intervene with the young-est children, particularly in broader areas thatdevelop generic social and emotional skills.However, interventions also need to be longterm, including over several years, and newwork on specific problems, such as violence andbullying, with older students is effective too,especially if it builds on a generic skills base.Once an effective intervention has run, regularbooster sessions with older students appear tobe helpful to overcome the recurrent problemof diminution of effects of intervention in thelonger term, which in many cases is as little as 6months.

Using leaders appropriately

The reviews found that many different types ofleaders had a role to play within interventions.All need extensive and intensive training.Specialist, often clinically trained staff proved tobe effective at the start the process of interven-tion development and dissemination. However,for interventions to be sustainable, andembedded in the life of the school, and in orderto start to affect academic achievement, routinestaff, and in particular teachers, need to takeover. Peers can be effective, for example inmediation, although care has to be taken toavoid putting together students with difficulties,particularly of the aggressive type, as they canreinforce on another. Families and communitiescan add strength and depth to work in schools ifappropriately involved, and can help to supportand reinforce the messages children are learningat school.

Putting skills at the centre

The centrality of skills, and, in particular, workusing CBT approaches to any effective mentalhealth intervention, was strongly supported byseveral reviews in this study. The use of holistic,educative and empowering theories and interac-tive pedagogical methods was endorsed bymany of the reviews which found that behav-ioural and information-based approaches anddidactic methodologies were not nearly as effec-tive. In terms of the theories that underpinskills acquisition, European theory tends to be

holistic, emphasizing not just behaviour changeand knowledge acquisition, but also changes inattitudes, beliefs and values, while Europeanhealth education has long pioneered activeclassroom methodologies, involving experientiallearning, classroom interaction, games, simu-lations and groupwork of various kinds (Weare,2000).

Using a whole-school approach

Most of the reviews reported here that dis-cussed the issue suggested that skills work aloneis not enough, and that for optimal impact,skills work needs to be embedded within awhole-school, multi-modal approach which typi-cally includes changes to school ethos, teachereducation, liaison with parents, parenting edu-cation, community involvement and coordinatedwork with outside agencies. The whole-schoolapproach, well implemented, has long beenseen as more effective in terms of outcome thana skills focused, curriculum based, approachalone.

Taking a whole-school approach is in linewith a great deal of international policy andpractice. Within Europe, and in other parts ofthe world such as Australia, programme devel-opment by agencies such as the WHO, EC, andvarious national governments has been highlyinfluenced by the settings approach of theWHO, with its focus on creating healthyenvironments. This has given rise within Europeto large-scale, agency-led, whole-school pro-grammes such as Health Promoting Schools(Schools for Health in Europe, 2010), HealthySchools (Healthy Schools, 2011), Social andEmotional Aspects of Learning (DES, 2010)and the Good and Healthy School (Paulus,2009). Australia has the state-led Mindmatters(Mindmatters, 2009) programme and thegovernment-led Kidsmatters (Kidsmatter, 2009)framework.

The need for rigorous implementationof whole-school approaches

All of these whole-school programmes arepopular, widespread and well thought of bypractitioners and policy-makers. However, it isnotable that the evidence generated by themhas been weak in terms of hard outcomes andhas not resulted in evaluations that are robustenough to feature in systematic reviews. It is of


by guest on Decem


Dow

nloaded from


concern that none of these high-profile pro-grammes was therefore able to meet the evalu-ation requirements of this review and beincluded here as examples of evidence-basedinterventions.

It is, however, clearly possible for whole-school approaches to result in hard outcomesand appear in a systematic review. So what isgoing wrong with many agency-led Europeanand Australian whole-school approaches interms of failure to produce outcomes whichlead to their inclusion in a systematic review? Aset of findings from this study may cast light onthis problem.

Some recent reviews identified here suggestthat some whole-school approaches, includingin the USA, are failing to show impact (Wilsonand Lipsey, 2007; Durlak et al., 2011). Authorsattribute this to a lack of consistent, rigorousand faithful implementation which is causingthese approaches to become too diluted andlack impact. The necessary characteristics theyidentify for effective implementation includehaving specific, well-defined goals and rationale,a direct and explicit focus on desired outcomes,explicit guidelines, possibly manualized,thorough training and quality control and com-plete and accurate implementation.

The European and Australian style and thetype of whole-school approaches it generatestend to promote ‘bottom up’ principles such asempowerment, autonomy, democracy and localadaptability and ownership (WHO, 1997). Allthe agency-led whole-school programmesnamed above have produced a wealth of well-planned materials, guidelines and advice, butare also deliberately non-prescriptive and prin-ciples based. This flexible and non-prescriptivestyle is echoed in wider approaches to acrossEurope and Australia, mental health, whichemphasize the need for end-user involvementand the lay voice. This approach contrasts withthe US style of more top-down, manualizedapproaches, with scripts, prescriptive trainingand a strict requirement for programme fidelity.

There are strong reasons to retain the demo-cratic European and Australian approach forlarge-scale programmes for mental health. It isgenerally seen as providing essential supportivestructures, positive climates, empowered com-munities and end-user involvement, which leadsto well-rooted and long-lasting changes of atti-tudes and policies that are necessary to supportsustainable changes in mental health. However,

it is clear that, on its own this style of approachalso makes it challenging to achieve hard out-comes and measurable changes. There may needto be a balancing of this style with some morefocused, and more prescriptive elements, as hasbeen achieved already in some of the moredemonstrably effective whole-school pro-grammes. Those involved in approaches tomental health in schools right across the globe,including in Europe should consider having theconviction to build on what is now known, con-solidate and formalize their guidance and pro-cedures and provide a greater level of clarity anddirection for future developments to ensure con-sistent implementation of clear, evidence based,interventions. Unless mental health interven-tions are delivered with clarity and fidelity,approaches which would work well ifimplemented consistently, are likely to continueto be too diluted and vague to show real impact.

Meanwhile, on a practical level, advice onimplementation offered by such as Wilson thatschools only undertake interventions that fittheir context and which they can easilyimplement with fidelity and rigour, seems wise.Larger, multi-faceted interventions would alsodo well to ensure that meet the implementationcriteria for clarity and fidelity outlined above,and give concrete advice on such issues aswhere to start, how to set measurable goals andevaluate them, and the need to prioritize, tophase in changes slowly and ensure that theyare properly embedded before going on to thenext.

Strengths and limitations of the evidence

This review of 52 reviews of mental health inschools is the largest so far undertaken. Itincludes 32 studies not included in a set of foursubstantial and good quality reviews of the fieldconducted for NICE in 2007 and 2009 (NationalInstitute of Clinical Excellence, UK) (Adi et al.,2007a,b; Shucksmith et al., 2007; Blank et al.,2009). Seventeen of the additional reviews werepublished more recently and 15 were newlyincluded, probably due to the wider searchterms used in this review. This suggests that theconclusions of this review are based on the mostrecent and the broadest evidence trawl to date.

Systematic reviews are generally acknowl-edged as powerful tools in the evaluation of evi-dence of effectiveness, providing quantitativeestimates of the average impact of interventions


by guest on Decem


Dow

nloaded from


and reducing bias through their exhaustive strat-egies. Reviews of reviews take this further and,with a large number of studies to analyse, as inthis case, here, can produce some robust andreliable evidence.

However, it is important to note what thisdesign can miss. As a review of reviews thisreview could not, by definition, identify newstudies of primary approaches that may bepromising. It is also restricted to work thatpasses the quality criteria of systematic review.This means that this review only includes whathas been evaluated through controlled trialsand which may miss promising interventionswhich have been evaluated in other ways,through before and after studies or qualitativeapproaches, for example.

Nearly all authors of the reviews commentedon the lack of methodological rigour in many ofthe studies they analysed that made coming tofirm conclusions difficult. The problems encoun-tered included: lack of control groups; lack ofrandomization; small numbers; short duration;poor reporting quality and, in particular,inadequate description of methodological pro-cedures; lack of assessment of interventionimplementation; missing data; and failure toreport all outcomes. However, the methodologi-cal weaknesses may not greatly affect the val-idity and reliability of the conclusions. Wilsonand Lipsey (Wilson and Lipsey, 2006a), review-ing universal school-based social informationprocessing interventions on aggressive behav-iour, found that studies with higher qualitymethods (e.g. those with random assignment orlow attrition) did not produce better (or worse)outcomes than studies using less rigorousmethods. The same authors also found thatthere were no significant differences in terms ofoutcome between experimental andquasi-experimental studies (Wilson and Lipsey,2006b). The lack of difference made by meth-odological rigour lends support to the idea thatthe findings of this review may be applicableacross a broad range of work, including thatwhich falls outside the strict parameters of thisreview.

ACKNOWLEDGEMENTS

The authors would like to thank DataPrev col-leagues: Peter Anderson, Eva Jane-Llopis,Miquel Codony Bodas, Fleur Braddick and

Rebecca Gordon and Clemens Hosman; theleaders of the other work packages, in particu-lar Sarah Stewart-Brown for her invaluableadvice and the European Union for funding theproject. The opinions voiced in this review arethe authors’ own.

REFERENCES

** ¼ review included in the analysis*** ¼ intervention to be found in Europe**Adi, Y., Killoran, A., Janmohamed, K. andStewart-Brown, S. (2007a) Systematic Review of TheEffectiveness of Interventions to Promote MentalWellbeing in Primary Schools: Universal ApproachesWhich Do not Focus on Violence or Bullying. NationalInstitute for Clinical Excellence, London.

**Adi, Y., Schrader McMillan, A., Kiloran, A. andStewart-Brown, S. (2007b) Systematic Review of TheEffectiveness of Interventions to Promote MentalWellbeing in Primary Schools: Universal Approacheswith Focus on Prevention of Violence and Bullying.National Institute for Clinical Excellence, London.

**Bayer, J., Hiscock, H., Scalzo, K., Mathers, M.,McDonald, M., Morris, A. et al. (2009) Systematicreview of preventative interventions for children’smental health: what would work in Australian contexts?Australian and New Zealand Journal of Psychiatry, 43,695–710.

**Beelmann, A. and Losel, F. (2006) Child social skillstraining in developmental crime prevention: effects onanti-social behaviour and social competences.Psciothema, 18, 603–610.

**Beelmann, A., Pfingsten, U. and Losel, F. (1994) Effectsof training social competence in children: ameta-analysis of recent evaluation studies. Journal ofClinical Child Psychology, 23, 260–271.

**Berkowitz, M. W. and Bier, M. C. (2007) What works incharacter education? Journal of Research in CharacterEducation, 5, 29–48.

**Blank, L., Baxter, S., Goyder, L., Guillaume, L.,Wilkinson, A., Hummel, S. et al. (2009) SystematicReview of The Effectiveness of Universal InterventionsWhich Aim to Promote Emotional and Social Wellbeingin Secondary Schools. National Institute for ClinicalExcellence, London, UK review.

**Browne, G., Gafni, A., Roberts, J., Byrne, C. andMajumdar, G. (2004) Effective/efficient mental healthprograms for school-age children: a synthesis of reviews.Social Science and Medicine, 58, 1367–1384.

CASEL (2010) CASEL select programmes. http://www.casel.org/programs/selecting.php (accessed 30 November2010).

**Catalano, R., Berglund, M. L., Ryan, G. A. M.,Lonczak, H. S. and Hawkins, J. D. (2002) Positive youthdevelopment in the United States: Research findings onevaluations of positive youth development programs.Prevention and Treatment, 5, Article 15. 1–111.

**Clayton, C. J., Ballif-Spanvill, B. and Hunsaker, M.(2001) Preventing violence and teaching peace: a reviewof promising and effective antiviolence,conflict-resolution, and peace programs for elementary


by guest on Decem


Dow

nloaded from

http://www.casel.org/programs/selecting.php








school children. Applied and Preventive Psychology, 10,1–35.

DES (Department for Education) (2010) Social andemotional aspects of learning. http://nationalstrategies.standards.dcsf.gov.uk/inclusion/behaviourattendanceandseal (accessed 30 November 2010).

**Diekstra, R. (2008a) Effectiveness of school-based socialand emotional education programmes worldwide—partone, a review of meta-analytic literature. In Social andmotional education: an international analysis. Santander:Fundacion Marcelino Botin, Fundacion Marcelino Botin,Santander, 255–284.

**Diekstra, R. (2008b) Effectiveness of school-based socialand emotional education programmes worldwide—parttwo, teaching social and emotional skills worldwide, Ameta-analytic review of effectiveness. In Social andemotional education: an international analysis.Santander: Fundacion Marcelino Botin, FundacionMarcelino Botin, Santander, 285–312.

**Durlak, J. A. and Weissberg, R. P. (2007) The Impact ofAfter-School Programs that Promote Personal and SocialSkills. Collaborative for Academic, Social, andEmotional Learning, Chicago, IL.

**Durlak, J. and Wells, A. (1997) Primary preventionmental health programs for children and adolescents: ameta-analytic review. American Journal of CommunityPsychology, 25, 115–152.

**Durlak, J. A., Taylor, R. D., Kawashima, K., Pachan,M. K., DuPre, E. P., Celio, C. L. et al. (2007) Effects ofpositive youth development programs on school, family,and community systems. American Journal ofCommunity Psychology, 39, 269–286.

**Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor,R. D. and Schellinger, K. (2011) The impact of enhan-cing students’ social and emotional learning: ameta-analysis of school-based universal interventions.Child Development, 82, 474–501.

Ecclestone, K. and Hayes, D. (2009) The Dangerous Riseof Therapeutic Education. Routledge, London.

**Ekeland, E., Heian, F., Hagen, K. B., Abbott, J. andNordheim, L. (2004) Exercise to improve self-esteem inchildren and young people. Cochrane DatabaseSystematic Review, 1, CD003683.

**Farrington, D. P. and Ttofi, M. M. (2009) School-basedprograms to reduce bullying and victimization. CampbellSystematic Reviews, 2009, 6, 1–147.

**Gansle, K. A. (2005) The effectiveness of school basedanger interventions and programmes: a meta-analysis.Journal of School Psychology, 43, 321–341.

**Garrard, W. and Lipsey, M. (2007) Conflict resolutioneducation and anti-social behavior in US schools. Ameta-analysis. Conflict Resolution Quarterly, 25, 9–37.

**Green, J., Howes, F., Waters, E., Maher, E. andOberklaid, F. (2005) Promoting the social and emotionalhealth of primary school aged children: reviewing theevidence base for school based interventions.International Journal of Mental Health Promotion, 7,30–36.

**Greenberg, M. T., Domitrovich, C. and Bumbarger, B.(2001) Preventing Mental Disorders in School AgedChildren. A Review of the Effectiveness of PreventionProgrammes. Prevention Research Center for thePromotion of Human Development, College of Healthand Human Development Pennsylvania State University.

Gross, J. (ed.) (2008) Getting in early. London: SmithInstitute and the Centre for Social Justice.

**Hahn, R. (2007) Task force on community preventiveservices: effectiveness of universal school-based pro-grams to prevent violent and aggressive behaviour: a sys-tematic review. American Journal of PreventiveMedicine, 33, 114–129.

**Haney, P. and Durlak, J. (1998) Changing self-esteem inchildren and adolescents: a meta-analytical review.Journal of Clinical Child and Adolescent Psychology, 27,423–433.

**Harden, A., Rees, R., Shepherd, J., Ginny, B., Oliver, S.and Oakley, A. (2001) Young People and Mental Health:A Systematic Review of Research on Barriers andFacilitators. Institute of Education, University ofLondon EPPI-Centre, London.

Healthy Schools (2011) http://home.healthyschools.gov.uk/(accessed 30 November 2010).

**Hoagwood, K. and Erwin, H. D. (1997) Effectiveness ofschool-based mental health services for children: a10-year research review. Journal of Child and FamilyStudies, 6, 435–451.

**Horowitz, J. L. and Garber, J. (2006) The prevention ofdepressive symptoms in children and adolescents: ameta-analytic review. Journal of Consulting and ClinicalPsychology, 74, 401–415.

Jenkins, R. and Barry, M. M. (2007) Implementing MentalHealth Promotion. Churchill Livingstone, Elsevier,London.

Kidsmatter (2009) http://www.kidsmatter.edu.au) (accessed30 November 2010).

**Kragg, G., Zeegers, M. P., Kok, G., Hosman, C. andAbu-Saad, H. H. (2006) School programs targetingstress management in children and adolescents: ameta-analysis. Journal of School Psychology, 44, 449–472.

**Maxwell, C., Aggleton, P., Warwick, I., Yankah, E., Hill,V. and Mehmedbegovic, D. (2008) Supporting children’semotional wellbeing and mental health in England: areview. Health Education, 108, 272–286.

**McCarthy, O. and Carr, A. (2002) Prevention of bully-ing. In Carr, A. (ed), Prevention: What Works withChildren and Adolescents? A Critical Review ofPsychological Prevention Programmes for Children,Adolescents, and Their Families, Brunner-Routledge,Hove, pp. 205–221.

**Merry, S. N., McDowell, H. H., Hetrick, S. E., Bir, J. J.and Muller, N. (2004) Psychological and/or educationalinterventions for the prevention of depression in chil-dren and adolescents. Cochrane Database of SystematicReviews, Issue 2, 1–124, Art. No.: CD003380. DOI:10.1002/14651858.CD003380.pub2. New Zealand.

Mind Matters (2009) www.mindmatters.edu.au (accessed30 November 2010).

**Mytton, J. A., DiGuiseppi, C., Gough, D., Taylor, R. S.and Logan, S. (2002) School-based secondary preventionprogrammes for preventing violence. Cochrane Databaseof Systematic Reviews 2006, Issue 3, 1–93, Art. No.:CD004606. DOI: 0.1002/14651858.CD004606.pub2

Nederlands Jeugd Institute (2010) Effective youth inter-ventions http://www.nji.nl/eCache/DEF/1/08/320.htmll(accessed 12 February 2010).

**Neil, A. L. and Christensen, H. (2007) Australian schoolbased prevention and early intervention programs for


by guest on Decem


Dow

nloaded from

http://nationalstrategies.standards.dcsf.gov.uk/inclusion/behaviourattendanceandseal









http://home.healthyschools.gov.uk/






http://www.kidsmatter.edu.au






www.mindmatters.edu.au




http://www.nji.nl/eCache/DEF/1/08/320.htmll







anxiety and depression: a systematic review. MedicalJournal of Australia, 186, 305–308.

**O’Mara, A. J., Marsh, H. W., Craven, R. G. and Debus,R. L. (2006) Do self-concept interventions make adifference? A synergistic blend of construct validationand meta-analysis. Educational Psychologist, 41, 181–206.

Oxman, A. D., Cook, D. J. and Guyatt, G. H. (1994)Users’ guides to the medical literature: VI. How to usean overview. JAMA, 272, 1367–1371.

**Park-Higgerson, H.-K., Perumean-Chaney, S. E.,Bartolucci, A. A., Grimley, D. M. and Singh, K. P.(2008) The evaluation of school-based violence preven-tion programs: a meta-analysis. Journal of SchoolHealth, 78, 465–479.

Paulus, P. (2009) Mental health—backbone of the soul.Health Education, 109, 289–298.

**Payton, J., Weissberg, R. P., Durlak, J. A., Dymnicki,A. B., Taylor, R. D., Schellinger, K. B. et al. (2008) ThePositive Impact of Social and Emotional Learning forKindergarten to Eighth-Grade Students: Findings fromThree Scientific Reviews. Collaborative for Academic,Social, and Emotional Learning, Chicago, IL.

**Reddy, L. A., Newman, E., DeThomas Courtney, A.and Chun, V. (2009) Effectiveness of school–based pre-vention and intervention programs for children and ado-lescents with emotional disturbance: a meta-analysis.Journal of School Psychology, 47, 77–99.

**Rones, M. and Hoagwood, K. (2000) School-basedmental health services: a research review. Clinical Childand Family Psychological Review, 3, 223–241.

**Schachter, H. M., Girardi, A., Ly, M., Lacroix, D.,Lumb, A. B., van Berkom, J. V. et al. (2008) Effects ofschool-based interventions on mental health stigmatiza-tion: a systematic review. Child and AdolescentPsychiatry and Mental Health, 2, 1753–2000.

**Scheckner, S., Rollin, S., Kaiser-Ulrey, C. and Wagner,R. (2002) School violence in children and adolescents: ameta-analysis of the effectiveness of current interven-tions. Journal of School Violence, 1, 5–33.

Schools for Health in Europe (2010) http://ws10.e-vision.nl/she_network/index.cfm (accessed 30 November 2010).

**Shucksmith, J., Summerbell, C., Jones, S. andWhittaker, V. (2007) Mental Wellbeing of Children inPrimary Education (targeted/indicated activities).National Institute of Clinical Excellence, London.

**Sklad, M., Diekstra, R., Gravesteijn, C., de Ritter, M.and Ben, J. (2010) Effects of Socio-Emotional LearningPrograms. Paper presented at the 6th World MentalHealth Conference, November 2010, Washington DC.

**Stage, S. A. and Quiroz, D. R. (1997) A meta-analysisof interventions to decrease disruptive classroom behav-iour in public education settings. School PsychologyReview, 26, 333–368.

**Tennant, R., Goens, C., Barlow, J., Day, C. andStewart-Brown, S. (2007) A systematic review of reviews

of interventions to promote mental health and preventmental health problems in children and young people.Journal of Public Mental Health, 6, 25–33.

**Tilford, S., Delaney, F. and Vogels, M. (1997)Effectiveness of Mental Health Promotion Interventions:A Review. Health Education Authority, London.

**Vreeman, R. C. and Carroll, A. E. (2007) A systematicreview of school-based interventions to preventbullying. Archives of Pediatric Adolescent Medicine, 161,78–88.

**Waddell, C., Peters, R. V., Hua, R. M. and McEwan, K.(2007) Preventing mental disorders in children: a sys-tematic review to inform policy-making. CanadianReview of Public Health, 98, 166–173.

Weare, K. (2000) Promoting Mental, Emotional and SocialHealth: A Whole School Approach. Routledge, London.

Weare, K. (2010) Mental Health and social and emotionallearning: evidence, principles, tensions, balances.Advances in School Mental Health Promotion, 3, 5–17.

**Wells, J., Barlow, J. and Stewart-Brown, S. (2003)A systematic review of universal approaches to mentalhealth promotion in schools. Health Education, 103,197–220.

WHO (1997) The Health Promoting School: AnInvestment in Education, Health and Democracy.Conference report on the first conference of the EuropeanNetwork of Health Promoting Schools, Thessaloniki,Greece, Copenhagen, .

**Wilson, S. J. and Lipsey, M. W. (2006a) The Effects ofSchool-Based Social Information ProcessingInterventions on Aggressive Behaviour: Part I: UniversalPrograms. Nashville, Tennessee, Center for EvaluationResearch and Methodology, Vanderbilt Institute forPublic Policy Studies, Vanderbilt University.

**Wilson, S. J. and Lipsey, M. W. (2006b) The Effects ofSchool-Based Social Information ProcessingInterventions on Aggressive Behaviour: Part 2: Selected/Indicated Pull out Programmes. Nashville, Tennessee,Center for Evaluation Research and Methodology,Vanderbilt Institute for Public Policy Studies,Vanderbilt University.

**Wilson, S. J. and Lipsey, M. W. (2007) School-basedinterventions for aggressive and disruptive behavior:update of a meta-analysis. American Journal ofPreventive Medicine, 33, 130–143.

**Wilson, S. J., Lipsey, M. W. and Derzon, J. H. (2003)The effects of school-based intervention programs onaggressive behavior: a meta-analysis. Journal ofConsulting and Clinical Psychology, 71, 136–149.

***Witt, C., Becker, M., Bandelin, K., Soellner, R. andWillich, S. N. (2005) Qigong for schoolchildren: a pilotstudy. Journal of Alternative and ComplementaryMedicine, 11, 41–47.

Zins, J. E., Weissberg, R. P., Wang, M. C. and Walberg, H.(2004) Building Academic Success on Social and EmotionalLearning. Columbia Teachers College, New York.


by guest on Decem


Dow

nloaded from

http://ws10.e-vision.nl/she_network/index.cfm









Documents

Mental health promotion and problem prevention in schools: what … · 2015-10-19 · EMBASE, ERIC, CINALH, Sociological Abstracts, ASSIA, Psycinfo, the Cochrane Database of Systematic