The Circle of Friends intervention: a research synthesis

ABSTRACT

This research synthesis investigates studies that evaluated a well-known intervention, the Circle of Friends (CoF) – which seeks to promote the inclusion of socially neglected students. The aim of the review is two-fold; first, to identify the intervention’s variant applications and, second, to record what evidence-based outcomes have been reported. A systematic literature search including seven databases was conducted and based on inclusion and exclusion criteria, a total of nine articles were identified. The primary outcomes indicated that the intervention was mainly used with younger students (primary education) and only rarely used in secondary schools. Additionally, the results show that the intervention was set up significantly less often for girls. The quality of the majority of evaluations was lacking which leads to recommendations for further research on this intervention to include the utilisation of more rigorous and more rounded research designs and proper trialling. Critical issues surrounding the implementation of the intervention are also outlined.

KEYWORDS:

• Circle of Friends
• social participation
• special educational needs
• friendship
• peer group
• social skills intervention
• research synthesis

Introduction

Since the Salamanca Statement in 1994, the concept and practice of inclusive education have been widely debated, following the efforts of national governments throughout the world to develop inclusive education policies and secure equal educational rights for all children (Ainscow, Slee, and Best 2019). In the last few years, the special-educational-needs discourse about inclusion has moved to a whole-school-discourse based on the acknowledgement that inclusion is concerned with diversity more generally (Haug 2017). This change from a narrow to a broad perspective means that the focus of inclusive education is not only on guaranteeing the full participation of students with disabilities but rather of all students experiencing difficulties, whether arising from disability, migration, socio-economic background, etc. (UNESCO 2018). Accordingly, inclusive education refers to more than a shared physical space, relying on positive social participation and students’ subjective perceptions of school belonging (Juvonen et al. 2019). New policies and research on inclusive education address all learners’ needs with regard to their academic and their social-emotional development and their full participation (for an overview see Schwab 2020).

Koster et al. (2009) operationalised social participation as an umbrella term to refer to the social dimension of inclusion including four key themes: friendships/relationships; contacts/interactions; self-perception of social acceptance and acceptance by classmates. In this respect, a number of students repeatedly experience situations of social exclusion within the school, namely students stigmatised due to different reasons (e.g., students with disabilities, students with experiences of forced migration, Black students and students of Colour),¹ (Schwab 2018; Raabe 2019). Being isolated systematically can result in negative consequences for students, hindering access to social experiences and reducing their sense of belonging in school. Previous studies have indicated that experiences of social exclusion can produce negative effects on students’ self-esteem (Harris, Qualter, and Robinson 2013). Students who feel isolated describe greater anxiety, frustration and sadness during their stay in school (Nelson et al. 2016) and run a higher risk of developing depressive symptoms (Schwab and Rossmann 2020).

In this regard, fostering social participation of students has been highlighted as a major aim of inclusive education (Avramidis, Avgeri, and Strogilos 2018; Garrote and Dessemontet 2015; Hassani et al. 2020; Sanches-Ferreira et al. 2019). The educational need to establish effective and efficient strategies to promote social participation of all students has driven increased interest in the exploration of social participation enhancement strategies. The literature on interventions to support students at risk for social exclusion is vast, mainly targeting students with disabilities (e.g., Kretzmann, Shih, and Kasari 2015). In a systematic review of interventions aiming to improve the social participation of students with SEN, Garrote, Dessemontet, and Opitz (2017) classified the identified interventions into three groups: those aiming at cultivating interaction strategies to targeted students, those involving group activities in the academic or social context, and those delivering training of professionals to facilitate social interactions. Improving peer support is the common denominator of these interventions, with a substantial body of research demonstrating their effectiveness as a method to increase social acceptance and interactions (Garrote, Dessemontet, and Opitz 2017). Hassani et al. (2020) point out that to foster social participation and sustainable development of social competencies the most promising interventions are those which are: a) multicomponent, b) long term, c) universal, d) promoting collaborative work in learning groups, e) carried out regularly and comprehensively and f) integrated into regular lessons. Additionally, Humphrey, Barlow, and Lendrum (2018) emphasise that the quality of intervention delivery is crucial for achieving positive effects which can be supported by providing pre-intervention training as well as assistance during the implementation.

The Circle of Friends intervention

Several interventions meeting the above requirements have been developed, albeit of variable popularity. One well-known intervention is the Circle of Friends, a peer support intervention, first developed by Pearpoint, Forest, and Snow (1992), and then described by Newton, Taylor, and Wilson (1996). The intervention involves the formulation of a network of social support for these students so that they feel accepted by their peers (Shotton 1998). It is based on the recognition that peers can play an important role in the individual’s behavioural change (Newton, Taylor, and Wilson 1996). Further, it is believed that the peer network not only helps the neglected child to improve his/her social skills but also has the potential to change peers’ behaviour and attitude towards the child experiencing social difficulties.

Newton and Wilson (2016) describe a whole class meeting as the starting point of the intervention, where the so-called ‘focus child’ is not present but has given permission to hold it. Here, students are encouraged to talk about positive but also challenging aspects of the ‘focus child’ and receive at the same time information on how peer help is key to overcome prevailing difficulties. Further aims of this session are to create empathy and collect ideas on how to be supportive as peers after receiving information about different types of relationships. The session ends after six to eight volunteers are selected to participate in the circle meeting. In the second session after the circle is set up, the ‘focus child’ receives feedback from her/his classmates on aspects they appreciate but also on those aspects, which make it difficult for them to build relationships with her/him. Following this feedback, the group as a whole comes up with ideas on how to support the ‘focus child’ and agrees on the next steps. In the following weeks, circle meetings are held to track changes as well as identify obstacles within the group.

Notwithstanding the great popularity this intervention has enjoyed over the last two decades, there has been no systematic effort to date to record the various ways it has been applied or its impact on the students’ social development. Consequently, this synthesis reviews the research on the use of the Circle of Friend in educational settings conducted between 1995 and 2019. Hence, this research synthesis aims to (a) identify and describe the programme’s variant applications and (b) summarise effects presented in the studies regarding the ‘focus child’ as well as other participants.

Method

Search procedure

This research synthesis followed the methods described in the PRISMA Statement (Liberati et al. 2009). The databases Scopus, PsychInfo, ERIC, Proquest, JSTOR, Web of Science as well as Taylor and Francis (Database: Education Research Abstracts Online) were used to search for relevant publications. The key words applied for all databases were circle* AND friend* in studies title and/or abstracts. In databases containing that option, the limitation was set to the subject areas of psychology and education (or social sciences). For example, the following syntax was used for the database Scopus:

circle* AND friend* AND SUBJAREA (psyc) SUBJAREA (soci) AND PUBYEAR > 1994 AND (LIMIT-TO (SRCTYPE,”j”)) AND (LIMIT-TO (DOCTYPE,”ar”) OR LIMIT-TO (DOCTYPE,”re”)) AND (LIMIT-TO (LANGUAGE, “English”)) AND (EXCLUDE (PUBYEAR,2020))

Inclusion criteria

Studies were considered eligible if they fulfilled the following criteria:

The main focus of the study was on the intervention of the Circle of Friends.

Outcomes were reported at least at the student level.

The study reports the use of empirical research (qualitative, quantitative or mixed-methods).

The study was published in a journal.

The study was published from 1995 to 2019.

The study was published in English.

Screening and selection of studies

The database search was conducted by three of the authors to ensure an independent selection process. As shown in Figure 1, the authors identified a total of 1723 articles through all seven databases at the first search stage by applying the above-mentioned inclusion criteria. The second step was the exclusion of duplicates, which reduced the total amount of articles to 1333. A title and abstract screening were conducted in the next step by all authors considering the inclusion criteria. Studies were marked as ‘unsure‘, if clear acceptance or rejection was not possible. These studies were discussed between all three authors and then a decision was made whether to include or exclude them based on the eligibility criteria. This phase resulted in the extraction of 23 articles after all three authors reached consensus. These articles were found eligible for the full-text screening which was conducted by all three authors independently. This selection led to the remaining nine articles to be included in the research synthesis. However, in one article (Barrett and Randall, 2004), two studies were discussed, which led to ten studies in nine articles to describe and analyse. The majority of the excluded articles were those that were solely theoretical, that simply used the term ‘circle of friend’ to refer to peer relationships, that described an intervention other than the Circle of Friends, or that had no student outcomes.

Data extraction

For data extraction, a protocol was designed by all authors which includes the topics listed in Table 1 (general intervention information – duration, variant of intervention, etc.) and in Table 2 (study information – design, measurement points, outcomes, etc.). For piloting the coding, two of the authors independently extracted information from the first two studies and then discussed the review protocol and resolved disagreements by consensus. After this step, the rest of the extraction was carried out independently by the two authors to avoid bias. The results were then compared by a third author to see if they corresponded. Additionally, another author conducted a final check of the data extraction sheets.

Figure 1. PRISMA search and selection diagram.

Table 1. General information on intervention.

Nr.	Article	Setting	Country carried out	Duration	Delivered by	‘Focus child(ren)’	Variant of intervention
1	Barrett and Randall (2004)	Mainstream primary school	UK (Scotland)	6 weeks/weekly à 30-min	Headteacher and Educational Psychologist (EP) in Training	1 student (boy) rejected by peers	Adapted version: ‘Focus child’ always present, circle for all circle members. 6 students in the circle – 4 boys, 2 girls.
2		Primary school	N/A	6 weeks/weekly meetings	Teacher, deputy class teacher and community education worker	3 students either rejected or neglected by their peers	Adapted version: 3 circles with 3 ‘focus children’, whole class involved. Circle for all circle members. ‘Focus child’ was not nominated officially and didn’t know that circles had been set up around them.
3	Frederickson and Turner (2003)	Mainstream primary schools	UK	6 weeks/weekly meetings	Intervention Group (IG) (Phase 1): EP student and teacherControl group (CG) (Phase 2):EP and teacher for the initial meeting and a staff member for the weekly meetings	20 students (19 boys, 1 girl, aged 6–12) with SEN (EBD and 30% with significant LD) in different schools and classrooms	Original version:2–5 peers in the circle, nominated by teachers.
4	Frederickson, Warren, and Turner (2005)	Mainstream primary schools	UK	6–10 weeks/weekly meetings	Assistant EP; Most of the staff and teachers participated as observers	Phase 1: 14 students (11 boys, 3 girls, aged 6.8–11.3) with SEN (7 with MLD, 6 with EBD and 1 with ASD) Phase 2: 7 students out of the 14 students from phase 1 (4 with MLD, 2 with EBD and 1 with ASD)	Original version: 6–8 students in the circle.
5	Kalyva and Avramidis (2005)	Preschool	UK	3 month/weekly à 30 min	Teacher	5 children (5 boys, aged 3.10–4.7) with ASD: 3 in the IG, 2 in the waiting CG	Original version: 25 peers/5 in each of the circles (15 girls and 10 boys).
6	Miller et al. (2003)	Mainstream elementary school	USA	Duration n/a/weekly à 30 min	N/A	3 students (boys, aged 11) with mild disabilities (EBD, developmental disability, academic difficulties) who were isolated/ignored	Original version: 3 circles; 4 peers (2 girls, 2 boys, aged 10–11) for each circle.
7	O’Connor (2016)	Mainstream	N/A	3 month/weekly à 30–40 min, 12 sessions	Teacher (usually)	Student (boy, aged 10) with Asperger Syndrome and learning needs	Adapted version: ‘Focus child’ always present
8	Shotton (1998) Educational	N/A	N/A	6 weeks/weekly à 30 min	EP (author of study)	Student (girl, aged 13) displaying anxiety and becoming school refuser	Adapted version: ‘Focus child’ always present, circle for 2 additional students (girls not having friends).4 peers (girls) in the circle
9	Smith and Cooke (2000)	Reception class	N/A	Approx. 7 weeks	EP and teacher	Student (boy, aged 5) at risk of exclusion	Adapted version: Whole class intervention combined with behavioural management framework and involvement of parents at planning meeting.
10	Whitaker et al. (1998)	Primary and secondary schools (Mainstream schools (6) and school for students with moderate LD (1))	UK	3–17 meetings at the time of evaluation/weekly à 20–30 min	Staff of the Autism outreach team and teacher	7 students with ASD	Original version: Peers for the circle nominated by teachers, other students with EBD and/or students, who would also profit included in the circle. 6–8 peers in the circle.

Table 2. Study information.

Nr.	Article	Sample	Methods	Measurement points	Outcomes for ‘focus child(ren)’/circle members/other class members/others (e.g. teacher)
1	Barrett and Randall (2004)	N n/a for whole class N = 6 (‘focus child’: N = 1; circle members: N = 5)	QuantitativeStudents Sociometric questionnaire (peer acceptance: how much children like to play with each member of the class; enjoyment rating of circle members)Staff and parents Evaluation questionnaires (unclear if open questions included)QualitativeStudents Whole class discussion	T1: pre-intervention T2: post-intervention	For ‘focus child’ QuantitativeNo effect on peer acceptance by the whole class.Increase in peer acceptance by the circle members. Enjoyment rating for the CoF: 5/10 QualitativeWould only remain friends with boys in the group. Staff reported limited effect on peer relationships. Parentsreported that children liked the CoF. Rating for the group 10/10 regarding help for their children. For circle members Quantitative Peer acceptance (no values available) increased for 2 circle members, no effect for 2 members and decreased for 1 circle member; enjoyment rating: 10/10 QualitativeWould stay friends with all of the circle members. Class members: Reported changes in the friendliness of the circle members For class members Qualitative Most of the non-circle members reported interest in taking part in a circle in the future.
2		N n/a for whole class ‘focus child’: N = 3	QuantitativeStudentsMy Class Inventory for actual and preferred class cohesion scores.Social skills questionnaireQualitativeStudentsClass discussion Interviews TeachersInterviews	T1: pre-intervention T2: mid-point of intervention T3: post-intervention	For ‘focus child’ Self-report: Increase in friendships for 2/3 children. More playtime with other circle members for 2/3 children. More awareness with whom from their circle playing is possible for all 3 children. Increase of 51% in scores on self-reported social skills questionnaire for one child. For circle members N/A – whole class involved For class members QuantitativeHigh increase in scores on self-report social skills. Enhanced class cohesion scores (MCI) decrease of the gap between mean, actual and preferred scores at T3. Enhanced communication and relationships between adults and students stated. Teachers reported a positive change in the class.
3	Frederickson and Turner (2003)	N/a for whole class and circle members ‘Focus child’: N = 20	IG (Phase 1) and wait-list CG (Phase 2)QuantitativeStudentsSociometric Rating Scale Self-Perception Profile for ChildrenMy Class Inventory (MCI-SF) TeachersTeacher’s Rating Scale of Child’s Actual BehaviourMy Class Inventory (MCI-SF)	IG (Phase 1) T1: pre-intervention T2: post-interventionCG (Phase 2) within-subject designT1–T2–intervention–T3	For ‘focus child’ IG (Phase 1 – between-group analyses)Positive effects on social acceptance by class (η^2 = .35) and circle members (η^2 = .31) from T1 to T2. No significant effects (students’ Self-Perception Profile and teachers’ ratings) for Child’s Actual Behaviour and My Class Inventory. CG (Phase 2 – within group analyses)Measured at three times: positive effects on social acceptance by class members (η^2 = .49) Significant increase of student self-perception on the Global Self-Worth scores (η^2 = .42). Significant increase in social acceptance scores (η^2 = .48) rated by the teacher. No changes in MCI-SF scores. No effect on the perception of social acceptance or behavioural conduct. For circle members N/A For class members CG (Phase 2) Decrease in competitiveness and experienced difficulty.
4	Frederickson, Warren, and Turner (2005)	N n/a for whole class and circle members ‘Focus child’: N = 14/N = 7 for follow-up (T4) drop out due to several reasons	QuantitativePhase 1Sociometric ratings: LITOP Questionnaire from the Social Inclusion Survey (SIS)Guess Who peer assessment–adapted version	Phase1T1: pre–interventionT2: 3-5 days post whole class meetingT3: post-interventionPhase 2T4: follow up – approx. 18 weeks post T3	For ‘focus child’Phase 1Whole class acceptance increased (η^2 = .30) and rejection decreased (η^2 = .40) after the whole class meeting from T1 to T2. The significant decrease (η^2 = .36) in rejection of the ‘focus child’ was only maintained at T3 for circle members Phase 2 Rejecting ratings decreases for circle members at T4 (η^2 = .29) and acceptance ratings return to baseline. No follow-up-effects for whole class, ratings return to baseline. For circle members N/A For class members N/A
5	Kalyva and Avramidis (2005)	N = 30 (‘Focus child’: N = 5, out of them, 2 in the waiting CG; circle members N = 25)	IG = 3/CG = 2 Quantitative Observations: 3 h for each child/1 h at T1, T2 and T3. Nr. of responses ‘focus child’ to peers’ contact initiatives.Nr. of initiation attempts of ‘focus child’.	T1: pre-intervention T2: post-intervention T3: follow-up (two months after T2)	For ‘focus child’ Significant increase in the nr. of successful responses and initiations and a significant decrease in the nr. of unsuccessful responses and initiations of ‘focus child’ in IG compared to ‘focus child’ in the CG – maintained at T3. No significant difference in responses and initiations rates between IG and CG at T1. For circle members N/A For class members N/A
6	Miller et al. (2003)	N = 30 (‘Focus child’: N = 3; circle members: N = 12/4 per circle; CG: N = 15)	Quantitative Observations: approx. 9 min during circle and recess time on % and form of interaction CoF Sociogram	Multiple measurements during: T1: pre-intervention T2: intervention T3: maintenance	For ‘focus child’Increase in appropriate and decrease in inappropriate interactions as well as no-interactions at lunch for all 3 students. Increases in friendly play at recess for 2 students For circle members Circle 1 peers T2 and T3: increase in appropriate interactions and decrease in no-interactions. Inappropriate interactions increased in T2 and decreased in T3. Circle 2 peers T2 and T3: increase in appropriate interactions and decrease in inappropriate and no interactions. Circle 3 peers T2: increase in appropriate interactions and decrease in inappropriate and no interactions. T3: decrease in appropriate interactions and remain of inappropriate and increase in no interactions For class members N/A
7	O’Connor (2016)	N/A for circle members ‘Focus child’: N = 1	Quantitative Belonging scales–adapted version of Psychological Sense of School Membership ScaleSocio-metric measure: Social Inclusion Survey Qualitative Observations–STAR approach	T1: baseline (over three months) T2: post-intervention	For ‘focus child’ Experiences higher global self-esteem scores and feeling socially accepted.Growth of peers’ willingness to make social initiations and relationships with ‘focus child’.Increase of likelihood to accept ‘focus child’s’ behaviour (empathy). Increase in the frequency and quality of social contacts between ‘focus child’s’ peer groups. For circle members N/A For class members N/A
8	Shotton (1998)	N n/a for the whole class N = 5 (‘Focus child’: N = 1 circle members N = 4)	QuantitativeStudentsSociometric questionnaire Qualitative Students Talks, discussions (not clearly described) Teachers, parents Talks, discussions (not clear).	T1: pre-intervention T2: post-intervention	For ‘focus child’ Happier at school (teacher and mother’s feedback, but not on the ‘focus child’s perception). More talkative and more confident (teacher feedback); Gained peers with whom she would talk to. For circle members Felt less isolated and much happier at school. 2 other students having no friends had more friends (self-report). For class members N/A
9	Smith and Cooke (2000)	N n/a for whole class ‘Focus child’: N = 1	Quantitative Observations Qualitative Classroom staff, mother Reports	T1: pre-intervention T2: post-intervention	For ‘focus child’ Improved social skills and less antisocial behaviour at T2. For circle members N/A–whole class involved For class members Motivated to improve their skills and focus on positive aspects.
10	Whitaker et al. (1998)	N n/a for whole class N = 59 (Circle members N = 52 and ‘focus child’ N = 7) N = 7 (circle facilitators)	Qualitative Interviews with: school staff involved in the circle,the focus, his/her parents, and thecircle members. QuantitativeB/G Steem for the circle members (N = 3 classes, out of 6). For other school staff no information on the instrument.	no clear pre/post design	For ‘focus child’TeachersReported enhanced quality of contacts and higher nr. of new friends (circle and non-circle members).Less adult support needed.Decrease of anxiety (4/7 teachers).Improved behaviour due to problem-solving focus. Negative: ‘increased egocentricity’. ParentsReported that children were more extroverted.More playtime with same-aged students.For circle membersTeachersReported development of empathy (all teachers).Greater self-esteem (4/7 teachers).Enhanced skills in joining discussions (3/7).Distress with ‘focus child’s’ behaviour (2/7).Self-reportAbility to offer help (32/52).Growth of self-expression and empathy.Would continue with the circle (48/52).No reciprocal relationships between them and ‘focus child’ B/G Steem: higher self-esteem for 70% compared to less than 50% of their class members (non-circle). For class membersN/A

Results

Data of the final studies identified through the database search are presented in two sections. The first section describes the different variations of the intervention and gives information on the participants and the context the intervention was carried out (Table 1). The second section reports on the evaluation tools and outcomes of the studies carried out (Table 2).

Circle of Friend intervention’s variant applications

Half of the studies applied an adapted version of the intervention while the other half carried out the original version. The main difference of the adapted versions compared to the original version was that the ‘focus child’ was always present and that the circle was either for all children or for more than one child. One study applied a whole class intervention combined with a behavioural management framework where parents were involved at the planning meeting, another study also involved the whole class but set up three circles. The interventions were carried out for six weeks (n = 4), six to ten weeks (n = 1), approximately seven weeks (n = 1), three months (n = 2) while one study did not provide any information on the total length and one study only gave information on the number of meetings conducted (3 to 17), which is 1 to 17 weeks at the time of evaluation. Nearly all circle sessions were conducted weekly. However, one study did not provide precise information on this. The duration of the meetings was between 30 and 40 minutes in five studies, 20 to 30 minutes in one study and no information was provided for four interventions. (Student) Teachers as well as educational psychologists were in charge of delivering the intervention most of the time, while the information is missing for one study. The majority of the interventions were carried out in primary schools (n = 5); the remaining were carried out in preschool and reception class (n = 2) and primary and secondary (n = 1). One study did not provide any information on the school level but did include information on student age, 13. Overall, it is difficult to cluster school-level information due to the different school systems and ages of children enrolled.

‘Focus children’, aged between 3.10 and 13, were 41 boys and five girls in total in nine studies whereas this data was missing for one study. Students with different special educational needs (e.g., Autism Spectrum Disorder (ASD), Emotional and Behavioural Disorder (EBD), Learning Disability (LD)) were the ‘focus children’ in seven studies, while three studies focused on students who were rejected or at risk of exclusion and one study was on a student displaying anxiety.

Evaluation methods and impacts of the Circle of Friends

The exact sample size of participants, more precisely class members, circle members, teachers/school staff and/or parents, involved in the evaluation was either completely or partly missing in eight studies although outcomes were reported either for the respective person her-/himself or for the ‘focus child’. For the ‘focus child’ and the circle members, sample size ranged from 5 to 59, for those studies that reported sample size for these groups. When considering only the ‘focus child’ in four studies, the circle was set up for one, in two studies for three, in one study for five, in one study for seven, in one study for 14 and in one study for 20 ‘focus children’. Six studies combined quantitative and qualitative methods, four studies used quantitative methods. Ten studies were designed with at least two measurement points. A follow-up assessment after the completion of the intervention was conducted in three studies. Six studies used some kind of a sociometric nomination methodology.

Outcomes for the ‘focus child’ were available in all ten studies, for the circle members in six studies (two studies applied a whole class intervention) and for the other class members, who were not part of the circle, in four studies. Nevertheless, it has to be stated that there are huge differences in qualitative and quantitative data available, as well as in data deriving from self-reports, reports by circle members, class members and/or teachers. The main outcomes described concerned an increase in acceptance by the circle members and/or other class members for the ‘focus child’ as well as an improvement of social skills (reported in form of, e.g., higher social skills scores, improved behaviour, increase of the number of friends or quantity and quality of social contacts). The growth of social skills was also reported for the circle members in five studies. Two studies included information about the effect size, while eight studies did not report any effect sizes.

Discussion

Over the past years, several attempts have been made to improve students’ social participation, especially within the subsample of students with SEN (see, e.g., Hassani et al. 2020). A frequently used intervention to support inclusion is the Circle of Friends. However, despite its widespread popularity, no systematic review on its effects has been conducted to date. Therefore, the present paper took a closer look at the few available studies evaluating the intervention.

Generally, summarising the results of the included studies proved to be a particularly challenging task as they represent different research designs employing a broad mix of research methods. Some explanations can be put forward to account for this trend; for example, mixed-method designs have increasingly become more common; and in the field of inclusive education, in particular, it is often suggested to use qualitative approaches that allow the students’ voices to be heard (Brantlinger et al. 2005). Nevertheless, a substantial number of the included studies utilised a sociometric nomination technique, which indicated that, the number of friends or peer acceptance increased during the intervention application as well as directly afterwards (e.g., Frederickson, Warren, and Turner 2005; Barrett and Randall 2004). Similarly, positive results emerged from the studies that examined improvements in the social skills of the participating students suggesting positive gains for ‘focus children’ (e.g., Shotton 1998; Smith and Cooke 2000; Kalyva and Avramidis 2005). Another observation concerns the reported positive outcomes for the other circle members for whom social participation and especially social skills are also increasing albeit to a lesser extent. In contrast, for class members outside of the circle, limited effects were stated.

Limited as well as positive effects were reported in both, studies evaluating the original as well as an adapted version of the Circle of Friends. However, Barret et al. (2004) compared two adapted versions and reported that the whole class version of the intervention seemed to be more effective, although effects were limited in both versions. Also, Frederickson, Warren, and Turner (2005), who applied the original version, reported that the first whole class meeting had the largest impact on social acceptance. Cautiously stated, these findings might provide a small indication that including the whole class may be more promising when using the Circle of Friends. However, limited evidence exists for longitudinal effects on both accounts, social participation and social skills development.

Comparing results in the present paper was furthermore difficult for several reasons. First, the majority of the studies did not report the exact sample size, the methodology used nor effect sizes. It has also to be noted that the sample size was very small in general. Although the original version of the intervention suggests establishing one circle for one child, future research should aim for a larger sample size (e.g., by setting up more circles) in order to present more reliable results. Secondly, the different studies were interested in varying outcomes (e.g., social skills, peer acceptance, self-esteem). Results were also often presented vaguely which made it demanding to understand. Nevertheless, not only did these shortcomings make it difficult to discuss the effects of the Circle of Friends but also the fact that it can’t be told for sure if the results were due to the intervention or due to other factors. Only three studies involved a control group, and of these studies, Miller et al. (2003) compared data of students with SEN with data of students without SEN. Furthermore, it is not possible to say if the intervention works for all target groups. The majority of the studies addressed young children with SEN in primary education. While speaking of children, it is worth pointing out that it was primarily boys (n = 41) and only five girls for whom a circle was set up. Consequently, it is unclear if the intervention works equally for all genders, youth and children with other diversity dimensions and/or challenges as well.

Recommendation for future research

Specifically, some topics for the Circle of Friends could be thought through before using this intervention. For example, the selection of the ‘focus child’. While the ‘focus child’ in most studies is described as socially isolated, no clear criteria for determining such status were given. Furthermore, the question arises if negative effects could appear within the approach of selecting a ‘focus child’. There is always a danger that students may feel overwhelmed, stigmatised etc. by being in the spotlight. Further, if only some students out of the class are included in the circle, which is the case for the original version, what happens with all the other peer members in class? Previous intervention studies (e.g., Vaughn, Denton, and Fletcher 2010) indicated that intervention effects are long-lasting if interventions are implemented within the daily classroom rather than implementing an intensive intervention for a short period of time by an external implementer.

Apart from these questions, for further research on intervention studies, it is recommended to describe in detail the exact implementation of the intervention. Without knowing precisely how the intervention was implemented, evaluating its effects is rather insignificant. In addition, the importance of the implementation quality (and if the implementers have been trained) needs to be taken into account. In the best case, a mixed-method design which includes a teacher diary (to note how the intervention was implemented, which aspects worked well and which ones did not), an evaluation tool (including scales, interview guides) for all participants as well as information on how to conduct the evaluation are required to get comparable evidence on the intervention effects. The field would benefit from the utilisation of randomised control trials to confirm the causal relationship between the Circle of Friends and social participation coupled with the employment of external observers to assess intervention effects. Moreover, the long-term and follow-up effects should be taken into account more intensively.

Limitations

While considering the results of the review reported here, readers should be aware of some limitations. First, as in all systematic literature reviews, the focus on a limited number of articles, the use of specific search terms and the application of strict criteria for selecting the identified studies might have resulted in an incomplete account of the literature. Second, the well-documented publication bias in academic journals (see, e.g., Scargle 2000) may have prevented many studies that have reported zero effects (as well as negative effects in the case of intervention studies) from getting published thus resulting in their exclusion from the present review. Another limitation is that some of the included studies are not providing all information, which would be relevant for the present research synopsis. For example, information on samples, the variant of Circle of Friends applied, and the methodology were missing in some cases. Most importantly, a meta-analysis of the effects on students (with SEN) and their peers was not possible as only a very limited number of articles provided enough information for such calculation to be performed. Consequently, the evaluation of the research outcomes of the selected studies was based on the reviewers’ judgement rather than strict statistical criteria. Recognising these limitations, the present review has revealed the lack of adequate empirical evidence supporting the effectiveness of the Circle of Friends, thus highlighting the need for its rigorous trialling in future research efforts.

Conclusion

Around 25 years after the Circle of Friends was outlined, this review showed that the evidence of its effect on social participation of students (with SEN) is still rare. Moreover, the existing studies do not imply a very strong impact. However, the at-risk-situation of students with different diversity dimensions concerning their social participation still shows a need for change. Within the next few years, more effort should be put into this area to ensure that inclusive education is not just about educating students within the same classrooms but also supports them to achieve full social participation. Far from solely emphasising academic standards, schools should actively implement programmes aiming to enhance the social-emotional development of all students and especially those who are socially neglected. Researchers also have an important role to play in this, that of evaluating such programmes and disseminating evidence-based practices.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Sepideh Hassani

Sepideh Hassani is a PhD candidate at the Department of Education, University of Vienna. She works as a researcher in several projects as well as a lecturer at the Center for Teacher Education. Her present research focuses on social participation, social emotional learning, equality, diversity as well as anti-discrimination.

Silvia Alves

Silvia Alves is an invited Adjunct Professor at the School of Education, Porto Polytechnic Institute and a researcher at the Centre for Research and Innovation in Education (inED). Her research focuses on inclusive education, peers’ attitudes towards the inclusion of children with disabilities, and the implementation of disability awareness interventions.

Elias Avramidis

Elias Avramidis is an associate Professor at the Department of Special Education, University of Thessaly (Greece). His research mainly focuses on examining the theory and practice of inclusive education and the barriers to its implementation.

Susanne Schwab

Susanne Schwab is a full Professor for School Pedagogy at the University of Vienna, Center for Teacher Education and Educational Science and extraordinary professor at the North-West University in Vanderbijlpark, South-Africa. Her current research focuses on inclusive education and on learning support and learning intervention, as well as on students' achievement and social and emotional development of students from minority groups.

Notes

1 The terms Black students and students of Colour are used as political terms which indicate a social position. They refer to shared experiences of discrimination and not to the colour of skin.