A Systematic Review of Behavioral Interventions for Elementary School Children with Social, Emotional and Behavioral Difficulties: Contributions from Single-Case Research Studies

Challenges arising from the classroom behavioral management of students with social, emotional and behavioral difficulties are a concern for educational professionals. The purpose of this study is to review common elements of behavior interventions for the disruptive behaviors of children with social, emotional and behavioral difficulties. A systematic review was conducted through an electronic search of studies (from 2000 to 2017) on ERIC, Web of Science, FRANCIS, and MEDLINE databases. The inclusion criteria involved: (i) an intervention improving behaviors at school of children with disruptive behaviors; (ii) elementary school children with the majority of the sample or average age between 6-11 years old; (iii) at least one measurable outcome focusing on social/emotional/behavioral outcomes; (iv) single-case designs. Of the 5339 articles that were identified in the initial screening, 27 met the criteria to be included in the review. Common characteristics of successful interventions are discussed to make recommendations for future implementation.


Introduction
Social, emotional, and behavioural difficulties (SEBDs) have a prevalence of 2%-16% in the scholastic population, and the international literature reveals that there is a growing concern about the impact of these difficulties in classrooms (Cefai et al., 2008;Cooper & Cefai, 2013; Organisation for Economic Co-operation and Development [OECD], 2019). SEBDs are chronic, and include disturbing and/or disruptive behaviour such as persistent rule breaking, bullying of others or being a victim of bullying, social isolation, and refusal to engage in or failure to complete learning tasks (Cooper & Cefai, 2013), and generally interfere with the social functioning of the individual and with that of his/her significant others, reducing academic engagement, and negatively affecting classroom climate. These behaviours range from infrequent but extreme problems to less severe ones, but occur at high frequencies (Clunies-Ross et al., 2008).
Behaviour problems within a classroom increase the stress levels of both teachers and students, disrupt the flow of lessons, impede learning objectives and the processes of learning, and negatively affect school climate (Parsonson, 2012). Students with SEBDs also tend to internalise problems such as anxiety and/or withdrawal, which in turn places them at risk of academic failure and dropout (Arici-Ozcan et al., 2019;Gunter et al., 2002;Riney & Bullock, 2012). Challenges in the classroom behavioural management of students with SEBD are usually a great concern for educational professionals; these challenges impact both the student and the classroom as a whole (Chafouleas et al., 2010;Clunies-Ross et al., 2008).
In defining intervention strategies, it is important to consider both the individual student's behavioural expressions and their consequences on overall classroom dynamics, including the effects on teachers, which must be addressed to effectively meet the challenge of educating students with SEBDs (Gunter et al., 2002). Educators are confronted daily with the choice of appropriate strategies to ensure the best learning experience. They can choose from a great number of interventions, and this choice has important implications. The risk of compromising teaching and learning effectiveness motivates teachers and other professionals to seek effective behaviour management strategies (Sugai & Horner, 2002). Several studies have reported the need to equip teachers with proper training in practical and proven classroom management strategies that are supported by evidence (Clunies-Ross et al., 2008;Sadik, 2017;Trussell et al., 2016). Interventions on disruptive behaviours range from the use of preventive classroom management strategies to the implementation of individualised interventions for specific problem behaviours. The entire range has been recognised as important in educational psychology (Emmer & Stough, 2001). Teachers need to acquire new instructional and behaviour management skills and incorporate these skills into their teaching repertoire to support informed decisions and implement the best strategies to reduce disruptive and challenging behaviours in the classroom (Di Gennaro et al., 2007;Simonsen et al., 2008). The positive impact of providing teachers with training on students' behaviour and the general classroom climate (including the wellbeing of teachers) is widely acknowledged (Hieneman et al., 2005;Panacek & Dunlap, 2003).
The common approaches used to tackle behaviours of concern at school are focused primarily on behavioural and academic interventions (Mooney et al., 2003;Stage & Quiroz, 1997). According to DuPaul and Weyandt (2006), behavioural interventions for students with challenging/disruptive/problematic behaviours include antecedent-based strategies, consequence-based strategies, and self-management approaches. The optimal behavioural protocol includes components of all three approaches, aiming to anticipate and limit the effects of disruptive behaviours (proactive intervention) and encourage and support appropriate/alternative behaviours among students (reactive intervention). Briefly, antecedent strategies involve behaviour management procedures that reduce the likelihood of problematic behaviours occurring, whereas contingency management strategies encourage and support appropriate behaviours and/or manage disruptive behaviours when they occur.
Environmental changes such as increasing structure and predictability, reducing distractions, creating routines, establishing goals, and offering rewards represent interventions that may help to prevent inappropriate behaviours that generally have an escalation when treated only with reprimands and referrals (Scott et al., 2007). Moreover, because of their aggressive style and propensity towards opposing the rules, children with SEBDs often establish negative relationships and thus may seem outwardly sociable and outgoing. The literature demonstrates that social skills training programmes (Charlesbois et al., 2003;Maddern et al., 2004) for students with SEBDs, who often experience difficulties with teachers and peers, may help increase low self-esteem that arises from repeated frustrations, failures, and unsatisfactory interpersonal relationships. In the classroom, the teacher can encourage the integration of the child with SEBDs while tutoring small groups or in cooperative learning where the child has an active social part; this can begin the process of removing the negative label that has been given to the child.
The amount of research on interventions for disruptive behaviour in the classroom for students with SEBDs, succinctly sampled above, requires systematisation to allow individual decision-makers to unbiasedly assess primary research and to consequently make the most appropriate educational decisions, a common goal for systematic reviews (Lasserson et al., 2019). Evans et al. (2004) conducted a systematic review of research from the period 1975-1999, in order to evaluate the effectiveness of different strategies to support children with emotional and behavioural difficulties in mainstream classrooms within primary schools. The 27 studies that they identified included randomised controlled trials and reversal design experiments. Later, Cooper and Jacobs (2011) conducted an extensive review of international literature on educational models for children with SEBDs that demonstrated evidence-based outcomes for children. These authors analysed multiple dimensions (e.g. interventions in the classroom, whole-school approaches, working with parents, and multi-agency intervention), and collected studies using different methods (e.g. randomised clinical trials, prospective 'naturalistic studies' without control/comparison groups, retrospective studies, and case studies), although focusing, where possible, on randomised clinical trials.
Single-subject research, a rigorous scientific methodology, has been carried out to establish evidence-based practices (Horner et al., 2005), and has been the focus of systematic reviews in such areas as social communication skills for children with autism (Hansen et al., 2014) and training programmes for parents of children with autism spectrum disorders (Patterson et al., 2011).

Research Goal
Interventions for disruptive behaviour in the classroom can be grouped into two levels: (1) interventions implemented at the individual level, in which the unit of analysis is set at the student targeted for intervention, and (2) classroombased interventions, in which the unit of analysis corresponds to the classroom as a whole. This study aimed to determine the common elements of individual-level behaviour interventions for disruptive behaviours of children with social, emotional, and behavioural difficulties through a systematic review of single-case research studies.

Search Strategy
The search for this study was performed from October 2017 through November 2017, in multiple databases that index literature from the fields of health, psychology, and education: ERIC, Web of Science, FRANCIS, and MEDLINE. A wide range of terms for SEBDs were combined with terms for 'classroom strategies' and 'intervention studies' (Table 1).  Table 2 displays the inclusion and exclusion criteria for the studies. The population of interest in this review was school-aged children who receive their primary and elementary education in general classrooms and display SEBDs. Due to the cultural differences in educational systems between countries, the age range was limited, and the majority of the sample or average age was between 6 and 11 years. Students with SEBDs or learning disabilities were included, but students with cognitive or severe disabilities such as intellectual disability or autism and mental health problems were not considered.

Selection Criteria
Articles published as full texts in peer review journals between 2000 and 2017 and in the English language were included. A previous systematic review published by Evans et al. (2004) had already researched the literature published in 1999 or earlier. Systematic or literature reviews, case studies, abstracts, conference papers, theses, books, and other grey literature were excluded, in accordance with the study design criteria, which only included interventionbased studies with single-subject designs.
Regarding the intervention, the only criterion established was the description of the intervention. Since the goal of this systematic review was to understand and synthesise the common characteristics of effective disruptive behaviour interventions for children with SEBD, no more criteria were set.
To ensure that all the important information was addressed, included studies assessed at least one dependent variable measured as a directly observable behaviour related to problematic classroom performance.

Inclusion criteria Exclusion criteria Sample
Students demonstrating behavioral problems, including those with high incidence disabilities such as learning and behavioral issues.
Studies including students with cognitive or severe disabilities such as intellectual disability or autism.
Elementary school children and youth with the majority of the sample or average age between 6-11 years old Studies involving children who are taught in selfcontained classrooms Settings School or classroom setting in which typical classroom instruction is conducted.
Studies focusing on non-classroom settings such as the cafeteria or recess Outcome Authors explicitly described a dependent variable as a directly observable behavior related to problematic classroom performance such as poor social interactions, low academic engagement, or disruptive behaviors.
Studies not reporting outcomes for children

Data Abstraction and Classification Process
After the identification of potentially eligible papers through the database search, three researchers conducted a review of the papers based on titles and abstracts, after a random distribution of papers per each one. For the eligibility assessment, the papers were independently analysed and the study characteristics were extracted with a data extraction tool for data on authors, sample characteristics and selection criteria, design, description of the intervention, behaviours targeted, types of strategies, measurement tools, results, and limitations of the studies. Before this procedure, researchers performed an independent analysis of five studies. The coding scheme was discussed by researchers. Disagreements between reviewers were resolved by consensus. In more difficult cases, a review by an independent researcher was required to ensure that consensus was obtained. There were 5339 records identified in the initial search. A total of 1656 duplicates were removed. After a review of the titles and abstracts and applying the inclusion criteria, 364 records remained for full-text review. Full articles were then retrieved and evaluated for relevance. Most studies removed during the full-text analysis did not cover the age range defined for this study, did not analyse disruptive behaviours as a dependent variable or outcome measure of the study, did not focus on individual interventions, or did not include experimental design studies.

Findings / Results
A total of 27 studies reporting the effects of varying interventions to improve students' behaviour in the classroom were included in the review. Table 3 provides an overview of the study characteristics, including sample characteristics and selection criteria, design, description of the interventions, types of strategies, behaviours targeted, results, and limitations of the studies.

Direct classroom observation
Teacher: Functional Assessment and Motivation Interview (Janney, 2008) Social validity (teachers and students) Full intervention improved levels of ontask behavior On removal of the extinction procedure, on-task behavior dropped to lower levels Reinstatement of the full intervention improved on-task levels Acceptability ratings were highest for full intervention Small sample size and convenience sampling Need to use functional analysis in addition to the use of descriptive analysis to determine function of behavior

A-B-A-B singlecase
Function-based student intervention Academic engagement Disruptive behavior Self-monitoring for 2 students with attentionseeking behaviors; taskmodification for 2 students with taskavoidance k Differential reinforcement of other behaviors (DRO) for 3 students and DRO with preferred activity for 1 student

Small sample size
Limited data set The hypothesized function of behavior was not validated

Participants Characteristics
Analysis of the studies carried out between 2000 and the present showed that most studies included between one and four participants (n=23, 79.3%). One study included 20 participants, another study included 66 participants, and 4 studies included between 7 and 9 participants. There was a clear predominance of male participants in the studies.
Among the 28 studies with information about the participants' genders (n=153 participants), 84.3% (n=129) of the participants were male.
Despite the emphasis on identifying studies focused on students with disruptive behaviours, there was variation in the descriptions of the students' conditions. Students were described as displaying problematic behaviours (n=7), displaying disruptive behaviours (n=11), being at risk for EBD (n=6), and displaying off-task behaviours (3). Only one study used verified emotional and/or behavioural disorder as a selection criterion (Wilkinson, 2005). Instead, most students were selected from general education classrooms based on teachers' nominations for demonstrating high levels of inappropriate classroom behaviour (e.g. Cook et al., 2014;Hansen et al., 2014). In addition to teachers' nominations, eight studies (29.6%) used specific instruments, such as the Social Skills Rating System (Gresham & Elliot, 1990) and the Personality and Behaviour Scale (Lin et al., 1992), to select participants (e.g. Brunch-Crump & Lo, 2017;Gresham et al., 2016). Three studies (11.1%) added the performance of functional behaviour assessments to identify potential participants (e.g. Cook et al., 2014). In five studies (18.5%), eligibility for inclusion in the study was also determined by the number of discipline referrals of students to the school director and/or psychologist (e.g. Campbell & Anderson, 2011;Lien-Thorne et al., 2005;Trussell et al., 2016).
Most of the studies did not report information related to participants' clinical or psychological diagnoses. In fact, only four studies reported that students presented a diagnosis of attention and deficit and hyperactivity disorder (e.g. Campbell & Anderson, 2011). With regards to special education services provision, among the 21 studies that reported this information (n=150 participants), 28% (n=42) of participants received this support due to behavioural problems.

Study Design
One of the inclusion criteria of this systematic review was an experimental single-subject research design. Among the studies included, 11 used multiple-baseline designs (e.g. Allday et al., 2012), 11 used reversal designs (e.g. Campbell & Anderson, 2008), 2 used alternating treatment designs (Kilgus et al., 2016;Moore et al., 2001), 1 used an A-B design (Regan & Howe, 2017), and 2 studies combined reversal and multiple-baseline designs (Cook et al., 2014;De Martini-Scully et al., 2000). Seven of the eight studies that used screening tools in their sample selection included a pre-posttest measure of students' problem behaviour; that is, the data collected with these tools formed a pre-baseline, which was compared with data at the end of the intervention (e.g. Carter & Horner, 2009;De Martini-Scully et al., 2000).

Targeted Behaviours and Outcome Measures
Disruptive, challenging, or problem behaviours were the primary dependent variable in the studies (n=17; 63%). A variety of behaviours were targeted. Most studies treated problem behaviours as a broad category (e.g. Carter & Horner, 2013;Cook et al., 2014). In these cases, problem behaviours were often operationally defined as behaviours not related to the academic tasks and that interfere with the teacher's instruction or the learning of other students. Two studies focused on specific problem behaviours, such as noncompliance, disruption, negative verbal or physical interactions, and not staying in one's seat (Campbell & Anderson, 2008;Miller et al., 2015). Off-task behaviour was considered as the primary dependent variable (n=11; e.g. Allday et al., 2012;Janney et al., 2013), but also as a specific form of problem behaviour (n=1; e.g. Miller et al., 2015). Off-task behaviours were often operationally defined as engaging in any motor activity or audible vocalisation not permitted by the classroom rules or not related to the assigned work or activity.
Academic factors, such as engagement, achievement, competence, and productivity, were the main secondary dependent variables of the studies (n=11; 40.7%; e.g. Campbell & Anderson, 2011;Gresham et al., 2006;Moore et al., 2001). Social skills was also included as a secondary variable in three studies (Gresham et al., 2006;Lo et al., 2011;Wu et al., 2010). Finally, in three studies, the teachers' skills with regards to implementing specific intervention strategies was included as a dependent variable (e.g. Allday et al., 2012;Trussell et al., 2016).
In terms of the evaluation of intervention effects, a wide variety of measures were used. The 27 studies focused on measuring problem behaviours using observation. Most observational data were collected using time-sampling record methods. Thirteen studies used partial-interval recording techniques (e.g. Mong et al., 2011;Nolan & Filter, 2012;Wu et al., 2012), six used momentary time sampling (e.g. Allday et al., 2012;Moore et al., 2001), and two used wholeinterval recording techniques (Campbell & Anderson, 2008;Janney et al., 2013). Furthermore, one study used the rating record technique (e.g. Wilkinson, 2005), and the remaining studies did not explicitly report information on the data record method.
Intervention effects were often estimated using more than one measurement instrument (n=20, 74.1%). In addition to observational methods, rating scales related to problem behaviours that were administered to teachers, such as the Functional Assessment Checklist (March et al., 2009), the Behaviour Intervention Rating Scale (Von Brock & Elliott, 1987), the Impairment Rating Scale (Fabiano et al., 2006), the Social Skills Rating System (Gresham & Elliott, 1990), and the Systematic Screening for Behavior Disorders (Walker & Severson, 1992), were also used. Parents participated in the evaluation of students' behaviours in two studies (Carter & Horner, 2009;De Martini-Scully et al., 2000).
Nine of the studies that met the inclusion criteria included Information and Communication Technologies in the form of apps or audio/video recorders that allowed teachers (e.g. Allday et al., 2012) and students (Moore et al., 2001) to track and assess progress in the targeted behaviours of the intervention.

Interventions, Strategies and Effects
The authors of all 27 studies considered their interventions to be generally successful. Thus, all the interventions and strategies described in those studies will constitute the focus of this section.
The independent variables of interest were disruptive behaviour interventions. These varied in function-based level, type of intervention, behavioural techniques, and implementation strategies. Fourteen studies (51.9%) included function-based interventions to address problem behaviours. Functional analysis was used as the framework for conducting these studies, which included an analysis of the antecedent. In most cases, function-based interventions were used in combination with other methods. For example, Brunch et al. (2017) implemented the Check-In Check-Out (CICO) intervention combined with a function-based self-monitoring strategy. Carter and Horner (2009) implemented the First Step to Success Program, combined with specific adaptations developed from the functional behavioural assessment. Thirteen studies (48.1%) used non-function-based interventions. Vance et al. (2012) used two nonfunction-based interventions: differential reinforcement of other behaviour and self-monitoring.
The reviewed studies also differed according to the type of intervention. Interventions consisted of both structured existing programmes and interventions that had been specifically developed for the study. Thirteen studies (48.1%) carried out existing interventions, such as the Check-in/Check-out (CICO; Crone et al., 2003), the First Step to Success (Walker et al., 1997), the Class Pass Intervention (Cook et al., 2014), and the Social Skills Training Curriculum (Wu et al., 2010), with some modification. The CICO programme was implemented in five studies (18.5%). The 14 studies with interventions that had been specifically developed for the study (e.g. Hansen et al., 2014;Lambert et al., 2006) focused on the use of specific behavioural techniques, such as 'behaviour-specific praise' (Allday et al., 2012), or group contingency (Ling et al., 2011). Beyond behavioural intervention strategies, three studies also implemented training in social skills to improve students' behavioural problems (Gresham et al., 2006;Lo et al., 2011;Wu et al., 2010).
In addition to function-based level and type of intervention, a range of behavioural techniques were used across the studies. These included antecedent-and consequence-based strategies. Antecedent-based strategies were used in 23 studies (85.2%). These strategies were as follows:  Gresham et al., 2006;Lien-Thorne et al., 2005). Antecedent-based strategies also included setting behavioural goals with participants (n=8, 29.6%; Campbell & Anderson, 2008 and self-monitoring (n=8, 29.6%; Hansen et al., 2014;Moore et al., 2001). Strategies such as the reduction in task duration (Moore et al., 2005) and the provision of opportunities to ask for or take breaks (n=3; 11.1%; Cook et al., 2014;Hansen et al., 2014;Trussell et al., 2016) were also used, although less frequently.
In four studies (14.8%), the implementation of interventions on students' disruptive behaviour in the classroom was combined with teacher training aimed at improving their ability to implement behavioural and pedagogical strategies such as behaviour-specific praise (Allday et al., 2012) and universal teacher practices (Trussell et al., 2016).
Overall, the reviewed studies reported favourable results. Both function-and non-function-based methods were found to effectively reduce the occurrence of problem behaviours. Of the 13 studies that implemented existing interventions, 5 (38.5%) reported a slightly positive differential effect between these interventions and additive function-based intervention (e.g. Campbell & Anderson, 2008;Carter et al., 2009). One study reported differences in the effects of antecedent-and consequent-based intervention strategies, with the former being more effective at reducing disruptive behaviours (Restori et al., 2007).

Fidelity and Social Validity of Interventions
The fidelity of the intervention was measured in 20 studies (74.1%) to monitor whether the intervention was implemented as planned (e.g. Hansen et al., 2014;Trussell et al., 2016). Ensuring the integrity of interventions varied in terms of the procedures adopted. In the majority of the studies, the fidelity of implementation was evaluated using a checklist with items pertaining to key features of the intervention (e.g. Campbell et al., 2008;Carter & Horner, 2009). The list was completed by one of the authors. The purpose of these checklists was either to observe teachers' practices during the implementation of the intervention (Trussell et al., 2016) or to assess students' progress in key areas of the intervention (Campbell & Anderson, 2008;Mong et al., 2011). In the other studies, treatment integrity was guaranteed using detailed protocols, in which teachers were asked to indicate whether each component was fully or partially implemented (De Martini-Scully et al., 2000;Wilkinson, 2005). The inclusion of a measure of fidelity check among individuals subjected to intervention or among teachers responsible for intervention assumes higher importance in the case of single-case designs.
The social validity of the intervention was measured in 15 studies (55.6%) through consumer satisfaction and treatment acceptability questionnaires. Although the social validity measures in some studies included the evaluation of intervention goals (e.g. Wu et al., 2011), the majority did not specifically measure the social significance of techniques used to decrease disruptive behaviours (e.g. Ling et al., 2011;Moore et al., 2005). Furthermore, treatment acceptability was assessed from both the teacher's and student's perspectives in 5 of the 16 studies with social validity measures (Cook et al., 2014;De Martini-Scully et al., 2000;Janney et al., 2013;Ling et al., 2011;Umbreit et al., 2004). One study also examined parents' perspectives (Wilkinson, 2005).

Limitations reported by the reviewed studies
All of the studies reported their limitations except one (Moore et al., 2001). A small sample size that could limit the generalizability of the results regarding the effects of the intervention was reported as a limitation by many of the studies (n=15; 55.6%). In addition, limited samples with regard to participants' characteristics (e.g. all participants were male, all participants were in elementary school, and all participants were from one class) were also identified as a limitation (e.g. Campbell & Anderson, 2008;Miller et al., 2015;Mong et al., 2011).
Limitations regarding methodological features were frequently reported. Sixteen studies (69.3%) reported the absence of a maintenance phase aimed at the measurement of the target behaviours and the intervention's long-term effectiveness (e.g. Cook et al., 2014;Regan & Howe, 2017). Extending the treatment and follow-up phases would have increased the robustness of the study design and strengthened confidence in the outcomes of the interventions. The failure to control the isolated effects of intervention components was reported as a limitation in five studies (18.5%).
Six studies reported the author's role as trainer or as primary data collector (22.2%) as a possible limitation because of the lack of control over reactivity effects (e.g. Vance et al., 2012;Wu et al., 2010).
Practical constraints that limited the number of observations were described in four studies (14.8%). These constraints made it difficult for authors to assess the various problem behaviours in different phases of the research design (Kilgus et al., 2016;Lo et al., 2011;Miller et al., 2015;Wilkinson, 2005).

Discussion
This systematic review aimed to synthesise the common elements of behaviour interventions for disruptive behaviours of children with SEBDs. Similar to the literature that describes the use of different terminology for SEBDs among students (Cooper & Cefai, 2013), this study also showed that a variety of terms are used to describe SEBDs. In most of the studies, students had not been officially diagnosed with SEBD; however, they exhibited significant behavioural problems and were considered at risk for emotional and behavioural disorders (Restori et al., 2007). Problem behaviours exhibited by students are particularly alarming, given the number of students who receive special education services. The impact of problem behaviours on long-term outcomes, such as continued poor academic performance, is an issue of concern (Mong et al., 2011).
Most of the reviewed studies used a combination of strategies to select participants, combining teacher nominations and/or team member observations and/or specific screening tools to identify problem behaviours (e.g. Brunch-Crump & Lo, 2017;Cook et al., 2014). Some studies included a pre-baseline phase in the design. This phase was described as crucial, either for the operational definition of the target problem behaviours or for the opportunity for participants to become familiar with the researcher. In this way, the authors aimed to reduce the reactivity and expectancy effects. In fact, reactivity or expectancy effects were mentioned as potential biases in six studies, in which the first author was the primary data collector.
The primary dependent variable of the reviewed studies was the problem behaviours of students. The results demonstrated that some studies considered a single, generic behaviour category, defined either as problem/disruptive behaviour or off-task behaviour, to be the target of the intervention. In other studies, the problem behaviour of students was treated as different discrete behaviours such as looking around, noncompliance, inappropriate talking, not staying in one's seat, and inappropriate use of materials. This differentiation seems to be motivated by the extent to which the intervention was preceded by the functional analysis of each behaviour. Within this context, the literature suggests that interventions designed on the basis of the function of behaviour effectively decrease problem behaviours in general classrooms (e.g. Briere & Simonsen, 2011;Umbreit & Ferro, 2015). In fact, about half of the studies analysed in this review included function-based interventions with favourable results (Trussell et al., 2016). However, there are also several studies with non-function-based interventions (Vance et al., 2012), with positive results. In fewer numbers, studies with both types of interventions were not conclusive in distinguishing the effects of function-based and nonfunction-based interventions due to the characteristics of the research design, which did not control the possibility of isolated effects (Carter & Horner, 2009). The number of effective non-function-based interventions supports using preventive and general interventions first and then proceeding to a more individual intervention, based on the function of the problem behaviour of the student, which is described as being more intensive and time-consuming (Sugai & Horner, 2002).
Most behaviour interventions combined antecedent-based and consequent-based strategies. Effective antecedentbased interventions included antecedent adjustments, such as the active teaching of classroom rules, the explicit teaching of appropriate behaviours, self-monitoring, and the modification of task assignments, such as reducing the overall length of assignments and/or breaking them into smaller sub-units. Consequent-based strategies useful for individual intervention included positive reinforcement and feedback and, less frequently, corrective statements.
The academic engagement of students was the main secondary dependent variable in the reviewed studies. The analyses demonstrated that both the dependent variables of problem behaviours and academic targets were effectively modified by interventions. Similar results were found in a meta-analysis of single-subject designs, which did not find a significant effect of target type on intervention response (Vannest et al., 2010).
With the exception of one study, all the studies in the present review measured problem behaviours using observation. Some studies reported the use of technology to support teachers and students in tracking and assessing progress in the targeted behaviours. Although technology can facilitate description and understanding of patterns of problem behaviours and can be used to monitor progress of behavioural interventions (Cohen & Rozenblat, 2015), few studies have examined the use of technology in interventions to manage disruptive behaviours of students with SEBDs (Merlo et al., 2018).
The extent to which the intervention was implemented with integrity was measured in many studies. Indeed, gathering data about the efficacy of interventions is important in determining its success. Nevertheless, a measure of fidelity check is essential to evaluate the extent to which interventions are comprehensively and consistently delivered among students or implemented in accordance with the intention of the treatment purposes (Little et al., 2002). The need for a treatment integrity measure is reinforced by the specificities of the research design of individual behavioural interventions. The technique most often used in the studies examined in the present review was a procedural checklist. Fuchs and Fuchs (1989) recommend the use of a component analysis checklist, on which researchers checked off whether a list of the components of an intervention strategy was implemented during the period of intervention observed.
Most of the interventions for disruptive behaviours analysed in this study appear to be equally effective. This finding supports previous research (De Martini-Scully et al., 2000;Stage & Quiroz, 1997). Therefore, factors such as ease of implementation and immediacy of results can also be considered by teachers in their decision to use an intervention. Within this context, social validity measures collect information about the social significance, appropriateness, and effectiveness of an intervention (Wolf, 1978) and provide a cue about the likelihood that interventions will be implemented in the future. The extent to which teachers approved an intervention and perceived it to be effective in decreasing children's behaviour problems was widely verified in studies that collected social validity data. However, not all studies specifically measured the social significance of the interventions. However, social significance can be observed when teachers select the students to be included in the sample, demonstrating the need to decrease disruptive behaviours on the part of some students. Similarly, the involvement of teachers in long-term interventions (e.g. Campbell & Anderson, 2011) also demonstrated the perceived value of the intervention goals.
The final methodological limitation was the common failure to not use the maintenance phase to check the durability of the intervention results. Maintenance of an effect is often needed to uphold the findings of single-case experimental designs and make it possible to attribute changes in the dependent variable to the independent variable (Smith, 2012).

Conclusion
This review of disruptive behaviour interventions identified several aspects in need of further development and exploration. More rigorous research is needed to verify that the interventions are effective for decreasing the rates of problem behaviours of students. Potential challenges include (a) the inclusion of a maintenance phase in the methodological design, (b) the control of isolated effects of each intervention component, (c) the development of a protocol to guide the consistent delivery of interventions, (d) the evaluation of social validity with particular attention to the implementer's opinion about the feasibility of the procedures. In addition to these methodological aspects, a more complete description of participants' characteristics is needed, one in which functional analysis is used in addition to descriptive analysis to determine the function of behaviour. For example, students whose behaviour functions are to escape or avoid something would react differently to an intervention compared with students whose behaviour is for receiving attention. Finally, very few studies have described the classroom setting, which has a strong influence on students' behaviour. Contextual factors, such as the conditions under which the intervention was implemented (e.g. the student's position in the class in the group; intervention delivered during instruction/smallgroup work/independent work), should also be examined. Information about these factors is essential for a more indepth understanding of problem behaviour intervention results.

Limitations
The findings of this review should be considered considering a few limitations. First, although a significant effort was made to include all studies of interventions to improve behaviours of elementary school children with problem behaviours, there remains the possibility that relevant studies meeting the eligibility criteria were not identified. Second, the use of such a comprehensive study selection criterion as 'behavioural problems' may represent an imbalance in the target population under analysis. Third, the decision to limit the present review to interventions conducted at the individual level led to the exclusion of studies using a set of relevant behavioural interventions, which used the class as the unit of analysis. Fourth, this review focused on classrooms. It remains unclear whether the interventions could be implemented and would have similar effects in other school settings (e.g. the cafeteria or playground). Finally, future studies should use meta-analysis to increase the robustness of inferences for synthetizing evidenced-based practices with students with SEBD.