• Pyramid clubs;
  • peer exclusion;
  • anxious solitary;
  • shy withdrawn;
  • internalization;
  • resilience


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


Social withdrawal is known to precede and contribute to the development of internalization problems in shy children. This study examined the efficacy of a selective primary school-based psychosocial intervention for withdrawn children.


The sample comprised 82 Primary 4 children aged between 7 and 8 years from 7 schools across Northern Ireland. There were 31 children in the comparison group (50.6% male) and 57 children received the intervention (41.7% male). A 2 × 2 mixed-model design was used: group (intervention group vs. comparison group) × 3 time points (pre- vs. 10 postintervention vs. 12-week follow-up) with repeated measures on the time factor. Teachers completed the SDQ at the three time points to assess participants' socio-emotional health status.


Pyramid participants showed greater reductions than the comparison group on the Emotional Symptom and Peer Problem factors following the intervention and their scores did not return to baseline levels at follow-up. Pyramid attendance accounted for approximately 12% and 9% of the variance in the Emotional Symptom and Peer Problems decreases respectively.


Pyramid Plus helped to alleviate internalization problems in participants and initially had a positive impact on social re-integration. The incorporation of cognitive restructuring techniques may be a necessary intervention modification to affect participants' emotion regulation strategies.

Key Practitioner Message

  • Evaluations of Pyramid Plus from the UK found support for the efficacy of Pyramid in reducing social withdrawal and depressive symptoms in shy children
  • This current research marks the first evaluation of Pyramid Plus in NI
  • Findings suggest that attending Pyramid led to significant reductions in Emotional Symptoms and Peer Problems for participants
  • At follow-up, Pyramid participants were once again more likely to be reporting somatic symptoms, worries and unhappiness than the comparison group
  • It is suggested that cognitive restructuring techniques may need to be incorporated into the intervention strategy to address the ‘within’ child vulnerabilities that are known to underlie the development of social anxiety


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Shyness is a form of social withdrawal that encompasses wariness and anxiety in the face of social novelty and perceived social evaluation (Rubin & Coplan, 2004). Results from an increasing amount of studies have linked shyness to maladjustment across the lifespan, particularly internalizing issues, loneliness, social anxiety, lower self-worth and a low repertoire of coping strategies. (Bohlin, Hagekull & Andersson, 2005; Coplan & Prakash, 2003; Coplan, Prakash, O'Neil & Armer, 2004). Greco and Morris (2001) have highlighted the need for early intervention and prevention programmes that are particularly tailored to the needs of the shy child.

The Pyramid Plus model of intervention was developed during the late 1970s in the United Kingdom. During latency, all children are postulated to experience a social hunger for acceptance by their peers. Gazelle (2010) has highlighted that peer exclusion is the key mediator of pathological anxiety and depression for shy children. Shy children desire interaction with their peers (normal levels of social approach motivation), but they avoid social interaction because they are worried that they will be rejected (high in social avoidance motivation). Gazelle (2006) has shown that these children are particularly ‘at risk’ of developing internalizing problems in classrooms with negative emotional climates characterized by disruptive behaviour and infrequent pro-social peer interaction. Gazelle and Ladd (2003) proposed a diathesis-stress model where peer problems reinforce social anxiety, which in turn trigger depressive symptoms.

The Pyramid approach aims to provide a safe environment for shy children where their sense of social belongingness can be nurtured. This encompasses the Pyramid rationale for providing targeted as opposed to universal interventions. Externalizing aggressive behaviours would immediately pose a social threat for these children and interfere with the de-sensitization process. Targeted interventions constitute a specialized health intervention for a specific subgroup of a population, as opposed to a universal approach where all members of a population receive the intervention. A recent systematic review on preventive interventions for internalizers revealed that targeted interventions run by trained interventionists were consistently associated with larger effect sizes than universal approaches delivered by endogenous providers such as teachers (Stice, Shaw, Bohon, Marti & Rohde, 2009).

Each individual child is postulated to have two distinct belongingness orientations, a growth orientation and a deficit reduction orientation. Belongingness orientation influences how they are perceived and treated by others (Lavigne, Vallerand & Crevier-Braud, 2011). A growth orientation is associated with genuine interest and compassion for others and has been found to be associated with emotional competence, subjective well-being, peer acceptance and resilience (Diener, 2000). The deficit reduction orientation, on the other hand, is associated with social interactions that are motivated by a need to fill a void. This is associated with loneliness, social anxiety and low self-esteem (Lavigne et al. 2011; Figure 1).


Figure 1. Pathways to depression and anxiety in anxious solitary children as proposed by the Pyramid Plus model

Download figure to PowerPoint

There have been two informal evaluations of Pyramid reported in the ‘grey literature’; these both found reductions in depressive symptoms and social withdrawal in attendees in comparison with nonattendees (Davies, 1999; Skinner, 1996). It was, however, the research carried out by Dr Ohl that has been instrumental in moving the efficacy investigation process towards a more empirical and objective grounding. These studies have been published in peer-reviewed journals (Ohl, Fox & Mitchell, 2012; Ohl, Mitchell, Cassidy & Fox, 2008) and the analyses revealed a significant group x time interaction for total difficulty scores [F (1,92) = 4.25, p = .04) in 2008 and significant group x time interactions for Emotional Symptoms (F (1,371) = 25.69, p < .001] and Pro-social behaviour (F (1,370) = 12.25, p < .01) in 2012 (Figure 2).


Figure 2. The Pyramid Intervention Model. Phase 1 Whole Class Screening (SDQ, Goodman, 1997)

Download figure to PowerPoint

This evaluation was set up by Barnardo's based on the UK evaluation process. This is currently the only primary school-based intervention in NI that aims to address the vulnerabilities of shy children during latency. In NI, clubs are run during school hours. Anxiety has been shown to produce a state of physiological arousal that narrows the focus of attention on perceived threat; this is associated with poor class engagement, school refusal and reduced short-term memory capacity. Anxiety has consistently been shown to be associated with underachievement at school in cross-sectional research (Langley, Bergman, McCracken & Piacentini, 2004). Longitudinal evaluations of CBT-based anxiety reduction programmes have shown that treatment-associated reductions in anxiety were followed by enhanced academic performance (Wood, 2006).

NI suffered civil conflict for over 30 years and although there has been a paramilitary ceasefire since 1994, there are still outbreaks of sporadic sectarian violence. The findings of Muldoon and Trew (2000) found that the experience of conflict for NI children was associated with lower behavioural adjustment associated with aggression and externalizing behaviours. More recent research has elucidated that sectarian violence was associated with mothers' perceptions of children's emotional security about community (Goeke-Morey et al., 2009). This study examined the impact of community conflict on child adjustment, including emotional and conduct problems, via mothers' perceptions. They found that sectarian and non-sectarian violence exposure were significant predictors of children's internalizing and externalizing symptoms and lower pro-social behaviour. The highlighted evidence suggests that witnessing such violence predicts both internalizing and externalizing problems in children. Shy children are known to be at increased risk of developing disorders in classrooms with negative emotional climates, which highlights the need for targeted interventions for shy children in school areas where there are high levels of externalizing behaviours and low levels of pro-social orientation.

The main aim of this research was to examine the effectiveness of Pyramid clubs on a sample of Primary four NI children. It was predicted that:

  • The pre-intervention Emotional Symptom, Peer Problem and Pro-social Skill scores of the Pyramid attendee children would differ significantly from those of the comparison group.
  • Pyramid participants' internalization scores would have decreased; to be approximately in line with those of the comparison group at time 2 and these changes would be maintained at follow-up.
  • Pyramid participants' Pro-social scores would have increased, to be approximately in line with those of the comparison group at time 2, and these changes would be maintained at follow-up.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


The original sample comprised 208 Primary 4 children, aged between 7 and 8 years (45.2% male), who were screened for the 2009–2010 and 2010–2011 interventions. Teacher-rated SDQ data were collected by Barnardo's at two time points (pre and post). Following the 2010–2011 interventions, the researcher collected data at the 12-week follow-up. In the overall 2009–2010 dataset, 57% of the overall sample, 56% of Pyramid sample and 58% of the comparison group were female. In the 2010–2011 dataset, 53% of the overall sample, 60% of the Pyramid sample and 45% from the comparison group were female. Table 1 contains details of the number of children who were screened at baseline, those allocated to either the Pyramid group or the Comparison group and the gender percentage in each group for each school.

Table 1. Data collection for the evaluation of pyramid plus Primary 4 evaluation by school type and gender at baseline
SchoolSchool typeAcademic yearNumber sampledPyramid (% female)Comparison (% female)
1Controlled Mixed2009–20101010 (60%) 
2Controlled Mixed2009–201077 (57%) 
3Controlled Mixed2009–201088 (37.5%) 
4Maintained Mixed2009–20102011 (73%)9 (67%)
5Controlled Mixed2009–20102710 (50%)17 (53%)
6Controlled Mixed2010–20112412 (33%)12 (42%)
7Maintained Mixed2010–20111911 (55%)8 (25%)
8Controlled Mixed2010–20111812 (50%)6 (100%)
9Controlled Mixed2010–20111311 (55%)2 (50%)
10Controlled Mixed2010–2011229 (78%)13 (46%)
11Controlled Mixed2010–20111510 (40%)5 (40%)
12Maintained Mixed2010–2011259 (67%)16 (37.5%)

Selection process for pyramid plus participants

The Pyramid Intervention is both a targeted and indicated intervention programme. Children were selected for Pyramid Plus based both on their SDQ scores and or on the basis of the involved professional's knowledge of the risk factors for internalization problems that may have been ongoing in their lives, such as parental psychopathology, separation, divorce or witnessing domestic or sectarian violence. The professionals involved included the class teacher, the head teacher and social workers. These professionals came together following the screening procedure for a multidisciplinary meeting with the Pyramid staff to facilitate the allocation of participants and to plan the intervention strategy needed for each child. This method was adopted in line with the selection procedures recommended in the Pyramid intervention manual (Pyramid, 2007). In this way, children whose SDQ scores were in the normal SDQ range (no clinical significance) were allocated a place in club, if they were displaying subtle changes in behaviour such as withdrawal and were known to be experiencing difficulties in their home or community life. Children who scored in the borderline (at risk) and abnormal (clinical diagnosis) SDQ ranges for Emotional Symptoms, Peer Problems who were not displaying co-morbid externalizing problems, were also offered a place in the club. As already outlined in the introduction, aggressive behaviour associated with externalizing symptoms would undermine the desensitization process that underlies the Pyramid intervention model.

Comparison group

A nonequivalent groups design was used, as the Pyramid Primary 4 intervention is a selective one, a randomized method of allocating children to either treatment or control group was not deemed suitable. Similarly, the use of a ‘waiting list’ control group (Webster-Stratton & Taylor, 2001) was not deemed ethical. The control group therefore comprised the class children who were screened for the intervention and who were deemed not to need it or those who were deemed to be presenting with externalizing behaviours.

Participants included in the ANOVA analysis

Analysis of variance performed in SPSS could only be performed on the cases, which had a complete set of data. Missing data patterns resulted in the exclusion of 126 cases from the analysis. The analysed sample comprised 82 Primary 4 children aged between 7 and 8 years from 7 schools across NI. There were 31 children in the comparison group (50.6% male) and 57 children received the Pyramid Plus intervention (41.7% male).

The intervention

The intervention is a therapeutic school-based intervention that is implemented during school hours, over 10 weekly sessions. Each weekly session runs for duration of 90 min (see Figure 3). Pyramid Clubs are offered to between 10 and 12 of the screened class children and the staff-to-child ratio is 1:4. The Pyramid Primary 4 intervention comprises a three-stage model and the procedural aspects are illustrated in Figure 3.


Figure 3. The Three Phases of the Pyramid Intervention Model

Download figure to PowerPoint

Parental permission and ethical approval

Ethical approval for the study was granted by the University's Ethics Committee. Parental permission was obtained for all children. No children were withdrawn by parents after the study started.


The Teacher-rated Strengths and Difficulties Questionnaire (SDQ, Goodman, 1997) was used in this study; it is one of the most widely used brief questionnaires for assessing child mental health problems (Goodman, Lampling & Ploubidis, 2010). It comprises 25 items, which are divided into 5 scales, each with 5 items. Three of these scales were utilized in the current evaluation: Emotional Symptoms, Peer Problems and Pro-social Behaviour. Each item on these scales is rated with a 3-point Likert-type scale of not true (score 0), somewhat true (score 1) and certainly true (score 2).

SDQ UK Norms

The UK normative data (norms) for the informant-rated SDQ, used in this study, are based on a sample of 10,438 British children who took part in a survey on child mental health commissioned by the Office for National Statistics in 1999 (Meltzer, Gatward, Goodman & Ford, 2000).


A mixed-model design was used; group (intervention group vs. comparison group) × 3 time points (pre- vs. 10-week postintervention vs. 12-week follow-up).

Analyses of outcome and improvement

Chi square tests of independence were used to examine whether there were significant group differences on the 10 SDQ Peer and Emotion items at baseline, postintervention and at 12-week follow-up. This approach meant that all available data could be analysed because tests of independence were conducted at each time point separately; so SPSS did not apply list-wise deletion because it was not a repeated-measures analysis.

Mixed-model analysis of variance

The data from the study were then analysed using analysis of variance (ANOVA) of each factor variable to assess main and interaction effects.

Shifts in Scoring Bands

A further analysis of changes in socioemotional behaviour was conducted by examining shifts in SDQ scoring bands generated by changes in both Pyramid Plus Attendee and Comparison group children's mean Emotional symptoms, Peer Problems and Pro-social Scores (Table 2).

Table 2. Banding ranges for SDQ subscales for teacher-rated 5- to 11-year olds
Emotional symptoms score0–456–10
Peer problems score0–345–10
Prosocial score6–1050–4

Frequencies for Peer Exclusion

Frequencies for teacher-rated peer exclusion were run for both groups at the three time points. Item 14 on the SDQ: Generally liked by other children, was used to estimate levels of social integration. This Peer Problem item is reverse-scored and was thus used as an indicator of peer exclusion or integration. The researcher did not have the freedom to introduce additional questionnaires, but deemed it as more important, to investigate this issue using the one item available, than to neglect assessing this aspect of the outcome evaluation considering its semantic significance.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Descriptive statistics from this current sample were compared with UK norms and the differences were tested using independent t-tests, gender differences were also assessed. The findings indicated that the mean Emotional Symptom score from this sample 2.33(2.51) was significantly higher (t = 4.80, p < .0001) than the UK norm 1.50 (1.90). There were no significant gender differences on Emotional Symptoms (t = −1.27, p > .05) or Peer Problems (t = −1.33, p > .05) (Table 3).

Table 3. UK norm and current NI Primary 4 descriptive statistics for pertinent SDQ subscales
SDQ subscaleUK norm (N = 4801)NI sample (N = 199) t
  1. *p < .05; ***p < .0001

Emotional1.50 (1.90)2.33 (2.51)4.80***
Peer1.40 (1.80)1.44 (1.90).25
Pro-social7.30 (2.40)7.41 (2.46).51
 Male (N = 88)Female (N = 111) t
Emotional2.08 (2.46)2.53 (2.54)−1.27
Peer1.24 (1.79)1.60 (1.98)−1.33
Pro-social6.97 (2.40)7.75 (2.45)−2.29*

The results from the Chi square tests of independence are reported in Table 4. They suggest that children were allocated to Pyramid based on their presentation of somatic symptoms, worry, unhappiness, peer exclusion and victimization. At time 2, the only significant group differences were in relation to peer exclusion and peer victimization. At time 3, significant group differences re-emerged, namely somatic, worry and unhappiness.

Table 4. Chi square tests of Independence for the SDQ items at the three time points
  1. *p < .05; **p < .001; ***p < .0001

Nervous clingy9.00**2.63.03
No good friend6.64**5.13*.74
Best with adult3.25.80.25

To assess violations of the assumption of sphericity, which are quite common in repeated-measures ANOVA, Mauchly's test of sphericity was examined. This suggested that the underlying assumption of homogeneity associated with the covariance matrix was violated for the Emotional Symptom (MW = .83, p < .05), Peer Problem (MW = .93, p > .05) and Pro-social subscales (MW = .91, p > .05). To adjust for this violation, Girden (1992) recommends that when epsilon (Greenhouse-Geisser estimate) is >.75, then the correction according to Huynh and Feldt should be used. Epsilon values for the three subscales used in the current analysis were >.75 and this suggested that the Huynh and Feldt adjustment is preferred (Table 5).

Table 5. Mean and mean differences for both groups at the three time points
  1. *p < .05; **p < .01

EmotionalPyramid3.04 (2.75)1.20 (1.65)1.43 (1.95)
 Comparison1.37 (1.92).70 (1.65).80 (1.54)
t  −5.24**−1.66−2.22*
Peer problemsPyramid1.85 (2.09).74 (1.35).80 (1.54)
 Comparison.84 (1.44).30 (.67).64 (1.0)
t  −4.04**−2.57**−.70
Pro-socialPyramid7.03 (2.65)7.87 (1.95)7.78 (2.10)
 Comparison7.93 (1.09)8.50 (1.97)8.10 (1.89)
t  2.66*1.76.98

The main phase of the analysis involved conducting the mixed between within ANOVA. The results indicate that longitudinal changes in Emotional Symptoms (F (1.75,85) = 9.05, p < .001) and Peer Problems (F (1.92, 85) = 7.35, p < .001) were dependent on group membership. There was no significant interaction effect for Pro-social Skills (F (1.90, 85) = 1.31, p = .28).

Shifts in scoring bands revealed that 33.3% of Pyramid children were experiencing borderline to abnormal levels of Emotional Symptoms at time 1; this decreased to 6.3% at time 2 and showed a slight increase to 10% at time 3. A similar trend was obvious for Peer Problems, with 22.8% experiencing borderline and abnormal indicators at baseline. This decreased to 3.2% at time 2 and increased only slightly to 5.8% at time 3. It can be concluded therefore that for the majority of participants, attending Pyramid helped to alleviate their Emotional Symptoms and Peer Problems (Table 6).

Table 6. Mixed analysis of variance results showing main and interaction effects
OutcomeVariableF df SigPES
  1. PES, Partial Eta Squared

Emotional symptomsGroup16.51(1,86).00.16
Group * Time9.05(1.75,85).00.12
Peer problemsGroup40.94(1,86).00.09
Group * Time7.35(1.92,85).00.09
Pro-social skillsGroup2.94(1,86).09.03
Group * Time1.31(1.90,85).28.02

The teacher-rated frequencies for the peer exclusion are reported in Table 7. This showed that 35.6% of Pyramid participants were experiencing peer exclusion before the intervention; this reduced to 13.7% at time 2, and increased, although not to baseline levels, to 24.3% at time 3.

Table 7. Shifts in SDQ scoring bands over 3 time points
EmotionalNorm 80%Borderline 10%Abnormal 10%
Pyramid T166.7%8.7%24.6%
Comparison T193.1%1.7%5.2%
Pyramid T293.7%4.7%1.6%
Comparison T294.1%   5.9%
Pyramid T390%4.3%5.7%
Comparison T396.6%1.7%1.7%
Pyramid T177.1%7.1%15.7%
Comparison T193.1%3.4%3.4%
Pyramid T296.9%1.6%1.6%
Comparison T2100%  
Pyramid T394.3%2.9%2.9%
Comparison T398.3%1.7% 


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The overall Emotional Symptom mean score for this sample was significantly higher than the UK mean. Goeke-Morey et al. (2009) found that sectarian and non-sectarian violence exposure were significant predictors of both children's' internalizing and externalizing symptoms in a recent NI sample. Prior to this, the impact of the troubles in NI on children was more focused on behavioural adjustment and externalizing psychopathology (Muldoon & Trew, 2000). Sectarian violence is still a fairly regular occurrence in both Belfast and Derry/Londonderry and the impact of the past and present ‘Troubles’ may still both directly and indirectly be influencing suffering in the NI child population (Bunting, Murphy, O' Neill & Ferry, 2011).

The results from the Chi square tests of independence showed clear distinctions between the comparison and intervention group at baseline. These highlighted the need for the intervention. There were no Emotional Symptom differences at time 2, although Pyramid children were still more likely to be friendless and to be experiencing bullying. At time three, group differences re-emerged on somatic symptoms, worry and unhappiness.

Examination of the Pyramid intervention model reveals very little in the line of therapeutic strategies to help anxious children with their cognitive vulnerabilities. The key variable associated with pathological levels of anxiety in shy children is the inability to inhibit the perception of threat in social situations. Evidence has consistently shown that anxious children overestimate the probability of danger, neglect rescue factors and ignore their own coping skills. The introduction of cognitive restructuring techniques into the Pyramid model may help the participants maintain the ground they have gained during the intervention itself. Cognitive restructuring techniques are seen as a secondary control strategy, where children are thought to recognize and change negative thoughts. Myles (2003) offers a variety of self-talk-based games and exercises that include ‘Spot the Message’ ‘Guess the Message’ ‘Fix the Feeling’ and ‘Emotional Toolbox’ that could be incorporated into the circle time element of the intervention model.

The current evaluation was constrained by the research design; future evaluations need to be mindful of the balance between a flexible design and the complex issues that then arise when analysing data where a randomization procedure has not been employed. Evaluations of this nature require the least disturbance of the up and running evaluation as possible, balanced with an understanding of the kind of evidence that is required to understand the dynamic components of the intervention process. For practical reasons, Randomized Control Trials will never be an option for these kinds of intervention evaluations, however; the quality of the evaluations could be sufficiently improved if the design incorporated measures of the proposed mediating mechanisms.

Future evaluations need to consider whether the exclusive use of SDQ (Goodman, 1997) is the most appropriate method for evaluating this intervention. The Social Anxiety Scale for Children (SAC, La Greca & Stone, 1993) includes three subscales that relate directly to the empirically identified issues associated with anxious solitude. Items on this measure include fear of negative evaluation, social avoidance and distress in situations. The need to assess social integration and loneliness is also of the utmost importance. The Social Inclusion Survey (Frederickson & Graham, 1999) would also be important as the teacher-rated measure of social exclusion used here is really only a very rough indicator of classroom dynamics. In this way, future research could incorporate ‘the voice of child’. It is acknowledged that to Barnardo's credit, they were very mindful that the teachers or children were not to be subjected to batteries of measurement instruments.

It is concluded that strongest aspect of the current intervention model is the targeted approach to early interventions for shy children during latency. The aims of the intervention are also congruent with empirical findings in relation to the buffering effects of friendship; however modifications may be required to treat the ‘within’ child factors that contribute to their social anxiety. In addition to this, there was no way of uncovering whether Pyramid participants felt stigmatized by attending club and that is an issue that may very well merit some qualitative research in the future.

Overall, the current findings offer preliminary support for the efficacy of Pyramid Plus but also highlight the need for modifications to the intervention and a more systematic and rigorous evaluation process. This investigation was carried out under the auspices of the CHILD (Child Health Initiatives in Life Development) programme at the University of Ulster (Table 8).

Table 8. Frequencies for peer exclusion
GroupNot trueSomewhat trueTrue


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

There was no external funding for this study. The authors acknowledge the help and support of Joan McGovern, Margaret McDonald and Angela McDonald at Barnardos NI in carrying out the study. The authors have declared that they have no competing or potential conflicts of interest.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Bohlin, G., Hagekull, B., & Andersson, K. (2005). Behavioral inhibition as a precursor of peer social competence in early school age: The interplay with attachment and nonparental are. Merrill-Palmer Quarterly, 51, 119.
  • Bunting, B.P., Murphy, S.D., O' Neill, S.M., & Ferry, F.R. (2011). Lifetime prevalence of mental health disorders and delay in treatment following initial onset: Evidence from the Northern Ireland Study of Health and Stress. Psychological Medicine, 42, 17271739.
  • Coplan, R.J., & Prakash, K. (2003). Spending time with teacher: Characteristics of preschoolers who frequently elicit versus initiate interactions with teachers. Early Childhood Research Quarterly, 18, 143158.
  • Coplan, R.J., Prakash, K., O'Neil, K., & Armer, M. (2004). Do you'want' to play? Distinguishing between conflicted-shyness and social disinterest in early childhood. Developmental Psychology, 40, 244258.
  • Davies, J.H. (1999). Children's writing improvements following participation in the Pyramid scheme. Unpublished report, University of Surrey, Surrey.
  • Diener, E. (2000). Subjective well-being. The science of happiness and a proposal for a national index. American Psychologist, 55, 3443.
  • Frederickson, N.L., & Graham, B. (1999). Social skills and emotional intelligence. Windsor: NFER Nelson.
  • Gazelle, H. (2006). Class climate moderates peer relations and emotional adjustment in children with an early history of anxious solitude: a Child X Environment model. Developmental Psychology, 42, 11791192.
  • Gazelle, H. (2010). Anxious solitude/withdrawal and anxiety disorders: Conceptualization, co-occurrence, and peer processes leading toward and away from disorder in childhood. In H. Gazelle & K.H. Rubin (Eds.), Social anxiety in childhood: Bridging developmental and clinical perspectives. New directions in child and adolescent development (pp. 6778). San Francisco: Jossey-Bass.
  • Gazelle, H., & Ladd, G.W. (2003). Anxious solitude and peer exclusion: A diathesis-stress model of internalizing trajectories in childhood. Child Development, 74, 257278.
  • Girden, E.R. (1992). ANOVA: repeated measures sage university papers series on quantitative applications in the social sciences, 84. Thousand Oaks, CA: Sage.
  • Goeke-Morey, M.C., Cummings, E.M., Ellis, K., Merrilees, C.E., Schermerhorn, A.C., Shirlow, P., & Cairns, E. (2009). The differential impact on children of inter- and intra-community violence in Northern Ireland. Peace and Conflict: Journal of Peace Psychology, 15, 367383.
  • Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581586.
  • Goodman, A., Lampling, D.L., & Ploubidis, G.B. (2010). When to use broader Internalising and Externalising subscales instead of the hypothesised five subscales on the Strengths and Difficulties Questionnaire (SDQ): Data from British parents, teachers and children. Journal of Abnormal Child Psychology, 38, 11791191.
  • Greco, L.A., & Morris, T.L. (2001). Treating childhood shyness and related behavior: Empirically evaluated approaches to promote positive social interactions. Clinical Child and Family Psychology Review, 4, 299318.
  • La Greca, A., & Stone, W. (1993). Social anxiety scale for children-revised: Factor structure and concurrent validity. Journal of Clinical Child Psychology, 1, 1727.
  • Langley, A., Bergman, R., McCracken, J., & Piacentini, J. (2004). Impairment in childhood anxiety disorders: preliminary examination of the child anxiety impact scale parent version. Journal of Child and Adolescent Psychopharmacology, 14, 105114.
  • Lavigne, G.L., Vallerand, R.J., & Crevier-Braud, L. (2011). The fundamental need to belong: on the distinction between growth and deficit-reduction orientations. Personality and Social Psychology Bulletin, 37, 11851201.
  • Meltzer, H., Gatward, R., Goodman, R., & Ford, T. (2000). Mental health of children and adolescents in Great Britain. London: The Stationery Office.
  • Muldoon, O.T., & Trew, K. (2000). Children's experience and adjustment to conflict related events in Northern Ireland. Peace Psychology: Journal of Peace and Conflict, 6, 157176.
  • Myles, B.S. (2003). Behavioral forms of stress management for individuals with Asperger syndrome. Child and Adolescent Psychiatric Clinics of North America, 12, 123141.
  • Ohl, M., Fox, P., & Mitchell, K. (2012). Strengthening socio-emotional competencies in a school setting: Data from the Pyramid Project. British Journal of Educational Psychology, doi: 10.1111/j.2044-8279.2012.02074.x
  • Ohl, M., Mitchell, K., Cassidy, T., & Fox, P. (2008). The pyramid club primary school-based intervention: Evaluating the impact on children's socio-emotional health. Child and Adolescent Mental Health, 13, 115221.
  • Pyramid. (2007). Club leader training manual. London: Pyramid.
  • Rubin, K.H., & Coplan, R.J. (2004). Paying attention to and not neglecting social withdrawal and social isolation. Merrill-Palmer Quarterly, 50, 506534.
  • Skinner, C. (1996). Evaluation of the effectiveness of Pyramid Clubs held in 1995–6. Unpublished report, University of Surrey.
  • Stice, E., Shaw, H., Bohon, C., Marti, C.N., & Rohde, P. (2009). A meta-analytic review of depression prevention programs for children and adolescents: factors that predict magnitude of intervention effects. Journal of Consulting and Clinical Psychology, 77, 486503.
  • Webster-Stratton, C., & Taylor, T. (2001). Nipping early risk factors in the bud: Preventing substance abuse, delinquency, and violence in adolescence through interventions targeted at young children (ages 0–8 years). Prevention Science, 2, 165192.
  • Wood, J.J. (2006). Effect of anxiety reduction on children's school performance and social adjustment. Developmental Psychology, 42, 345349.