Clin Child Fam Psychol Rev (2016) 19:392–402 DOI 10.1007/s10567-016-0211-4 Improving Treatment Response for Paediatric Anxiety Disorders: An Information-Processing Perspective Sarah Ege1 • Marie Louise Reinholdt-Dunne2 Published online: 1 September 2016 Ó Springer Science+Business Media New York 2016 Abstract Cognitive behavioural therapy (CBT) is considered the treatment of choice for paediatric anxiety disorders, yet there remains substantial room for improvement in treatment outcomes. This paper examines whether theory and research into the role of information-processing in the underlying psychopathology of paediatric anxiety disorders indicate possibilities for improving treatment response. Using a critical review of recent theoretical, empirical and academic literature, the paper examines the role of information-processing biases in paediatric anxiety disorders, the extent to which CBT targets informationprocessing biases, and possibilities for improving treatment response. The literature reviewed indicates a role for attentional and interpretational biases in anxious psychopathology. While there is theoretical grounding and limited empirical evidence to indicate that CBT ameliorates interpretational biases, evidence regarding the effects of CBT on attentional biases is mixed. Novel treatment methods including attention bias modification training, attention feedback awareness and control training, and mindfulness-based therapy may hold potential in targeting attentional biases, and thereby in improving treatment response. The integration of novel interventions into an The original version of this article was revised: The order of author was incorrect and the first author affiliation was missing. This has been corrected in this version. & Marie Louise Reinholdt-Dunne marie.reinholdt@psy.ku.dk Sarah Ege sarah.ege@sshf.no 1 Sørlandet Sykehus HF, Kristiansand, Norway 2 University of Copenhagen, Copenhagen, Denmark 123 existing evidence-based protocol is a complex issue and faces important challenges with regard to determining the optimal treatment package. Novel interventions targeting information-processing biases may hold potential in improving response to CBT for paediatric anxiety disorders. Many important questions remain to be answered. Keywords Information processing
Anxiety
Child
Attention Introduction Anxiety Disorders Paediatric (i.e. child and adolescent) anxiety disorders are a collection of syndromes characterised by excessive fear and anxiety, and related behavioural disturbances. Dysfunctional avoidance behaviours represent a core symptomatic feature (Salum et al. 2013). Anxiety disorders are the most common class of paediatric mental disorder (Beesdo-Baum and Knappe 2012). A lifetime prevalence study reported that 31.9 % of the US adolescents met diagnostic criteria for an anxiety disorder, with rates for individual subtypes ranging from 2.2 % for generalised anxiety disorder, to 19.3 % for specific phobia (Merikangas et al. 2010). Onset is often early, occurring by age 6 in 50 % of affected adolescents (Merikangas et al. 2010). Paediatric anxiety disorders can be considered gateway conditions for adult psychopathology (Britton et al. 2011), increasing the risk for being affected by either the same anxiety disorder, a different anxiety disorder or other forms of psychopathology later in life (Mohr and Schneider 2013). As such, paediatric anxiety disorders present an important target for intervention. Clin Child Fam Psychol Rev (2016) 19:392–402 Cognitive Behavioural Therapy (CBT) for Children and Adolescents The cognitive behavioural (CB) tradition holds that aberrant cognitive functioning maintains psychopathology. Traditionally, CBT aims to target and modify this cognitive dysfunction using a structured, present-focused, goal-directed psychotherapy based on cognitive and behavioural techniques (Clark and Beck 2010). CBT is considered an empirically supported treatment for a range of adult disorders (Hofmann et al. 2013) and is regarded as the treatment of choice for both adult and paediatric anxiety disorders (Cowart and Ollendick 2011; Mohr and Schneider 2013). Aims In a large, long-term follow-up study of CBT for paediatric anxiety disorders, Kendall, Safford, Flannery-Schroeder and Webb (2004) highlighted that a third of children fail to benefit from treatment, and called for future research into the needs of treatment non-responders. The aim of the present paper is to do this by examining potential shortcomings of CBT, and how limitations with existing treatment methods may be addressed. Treatment response is a complex issue and is likely to depend on a wide variety of factors. For example, factors independent of the core components of CBT such as therapist fidelity to the treatment method and external factors such as patient life events and changes in circumstances may have a role in some cases of poor treatment response. Other issues concern factors such as age, gender, parental psychopathology, comorbidity, primary diagnosis, symptom severity (e.g. Hudson et al. 2015) and the question of parental involvement in treatment (for review, see Breinholst et al. 2012 or Manassis et al. 2014). While the contribution of multiple factors in poor treatment response is recognised by the authors, it is not within the scope of this paper to examine these issues. Rather, the focus of the present paper is to examine whether there is room for improvement in the extent to which CBT targets the underlying psychopathology of paediatric anxiety disorders, and how this may be achieved. In recent years, the adoption of a developmental psychopathology perspective in paediatric anxiety disorder research has uncovered important insights into aetiological, risk and protective factors, which may help to elucidate what supports and impedes positive treatment outcomes. Treatment research has also seen the development of novel interventions that may help to ameliorate anxious psychopathology and may thereby hold potential in efforts to improve treatment response. Multiple variables have been investigated for their potential role in paediatric anxiety 393 disorders and as targets for treatment, including genetics, temperament, learning processes, parental influences, cognitive processing and emotion regulation. This paper limits its scope to examining the research question in relation to information-processing. To summarise, the aim of the present paper is to address the need for research into poor treatment response in anxious children by examining potential shortcomings of CBT from an information-processing perspective, and how these limitations may be dealt with. Materials and Methods The paper presents an exploratory review of theoretical, empirical and academic literature on paediatric anxiety disorders, CBT treatment of anxiety disorders and novel treatment interventions, with a focus on information-processing biases. The review presents a selective rather than exhaustive analysis, with emphasis on recent findings. The review focuses on literature published over the last decade, from 2004 to 2014. Adult-based literature is used sparingly. Literature was obtained from searches of Google Scholar and Psycinfo, recent books and accessing relevant citations from these sources. The review may be defined as exploratory rather than strictly systematic, utilising a broad and flexible literature search that was not restricted to a few listable search terms. This methodology was selected to allow for the integration of a broad range of information and flexibility in examining the research question (see discussion section for further information on the advantages and disadvantages of this approach). Paediatric Anxiety Disorders Fear and Anxiety in Children and Adolescents Fear and anxiety are adaptive emotional responses that are essential for survival (Blackford and Pine 2012). Fear occurs in relation to a threat stimulus that has the potential to cause immediate harm, whereas anxiety occurs in anticipation of a threat that is not immediately present. These emotional responses involve cognitive representations, physiological changes and behavioural responses that prepare an individual to deal with the threat. Fear and anxiety are regarded dysfunctional when the emotional responses are not in proportion to the threat and cause impairment and distress (Salum et al. 2013). As a universal part of development, children experience normative, transient fears and anxieties. However, for some children, these fears persist and new fears develop (Blackford and Pine 2012). 123 394 Anxiety Disorders For the purpose of this paper, anxiety disorders are primarily defined as the categories listed in the latest version of the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association 2013), which includes: separation anxiety disorder (SAD), selective mutism (SM), specific phobia (SP), social anxiety disorder/social phobia (SoP), panic disorder (PD), agoraphobia (AG) and generalised anxiety disorder (GAD). Some of the literature included in the review also broaches obsessive–compulsive disorder (OCD) and post-traumatic stress disorder (PTSD), which are also characterised by anxiety and avoidance behaviour, but no longer appear in the anxiety disorders category of the DSM-5. Information-Processing Biases in the Aetiology and Treatment of Paediatric Anxiety According to a contemporary cognitive-neurobiological model of information-processing recently articulated by Hofmann et al. (2012), the processing of threat-information involves a sequence of stages, each of which is associated with different neurobiological correlates, and anxious cognitions may be defined as specific features of this process. The model is based upon the hypervigilance–avoidance hypothesis, which postulates that cognitive processing in anxiety is characterised by an initial hypervigilance towards threat, followed by later avoidance processes. A recent review article suggested that this process involves facilitated attention (faster detection of threat stimuli than non-threat stimuli), difficulty in disengaging from threat and attentional avoidance (Cisler and Koster 2010). According to Hofmann et al. (2012), hypervigilance towards threat is an automatic and subconscious process. In contrast, avoidance processes occur at a later, slower-acting, and more consciously controllable stage of threat processing, and represent the application of maladaptive emotion regulation strategies. Some evidence indicates that biases in the processing of threat-information can have a causal role in the development of anxiety (see Field and Lester 2010, for a review). For example, training attentional biases towards negative emotional stimuli (MacLeod et al. 2002) and training threatening interpretation biases (Mackintosh et al. 2006; Wilson et al. 2006) in adults have been found to be associated with greater anxious and depressive emotional reactions and state anxiety, respectively, in response to a stress task. Attentional Biases Attentional biases are most commonly assessed in the recent literature using the dot-probe task (Manassis 2013). 123 Clin Child Fam Psychol Rev (2016) 19:392–402 This paradigm involves the brief, simultaneous presentation of a threat cue, such as a picture of an angry face, to one half of the visual field, and a neutral cue, such as a neutral face, to the other half of the visual field. Following disappearance of the cues, a dot is presented in the location previously occupied by one of the stimuli, and the child is requested to press a button as quickly as possible to indicate which side the dot appears on. A tendency to react more quickly to dots appearing in threat locations relative to neutral locations is inferred as indicating a bias towards threat. Conversely, a slower reaction time signals a bias away from threat, which is indicative of threat avoidance (Shechner et al. 2012). A recent review article by Shechner et al. (2012) concluded that the weight of the evidence suggests that anxious children, unlike their healthy counterparts, exhibit an attentional bias towards threat. The researchers noted that not all studies find attentional bias towards threat in anxious children and that attentional bias away from threat has been documented in some scenarios. Shechner et al. argue that the finding of threat avoidance is less frequent and has typically occurred in relation to unique, high-stress contexts. The duration of cue exposures also appears to be important. According to Shechner et al. bias towards threat has been more robustly observed when stimuli are presented for 500 ms or less, with findings more inconsistent for longer presentations. These findings appear to be consistent with the hypervigilance–avoidance hypothesis, in that research using longer presentation times may capture both of these processes, and lead to mixed results. CBT’s Effects on Attentional Bias Mixed findings have been reported with regard to CBT’s effects on attentional bias. Attentional allocation is considered to be an automatic, fast-acting process that is minimally amenable to conscious control (Hofmann et al. 2012; Manassis 2013). Consequently, it has been proposed that vigilance to threat is likely to be relatively resistant to the conscious efforts to change cognition that are practised in CBT (Manassis 2013). There is some empirical support for this view, with two studies documenting that threatrelated attentional biases in anxious children did not significantly reduce following CBT (Manassis et al. 2013; Waters et al. 2008). Some other studies report a more complicated picture. One study reported that the effects of stepped-care CBT for paediatric anxiety differed depending on the direction of pre-treatment bias (Legerstee et al. 2010). The first phase of therapy in this study was child-focused, and included 10 child sessions and 2 parental sessions. The second stage consisted of a 10-session, child–parent focused CBT intervention. Children who responded to treatment with the Clin Child Fam Psychol Rev (2016) 19:392–402 loss of any anxiety disorder were found to demonstrate significant reductions in biased processing of threat, providing evidence of change in attentional bias with CBT. Pre-treatment threat avoidance was associated with positive response in the first phase of treatment, pre-treatment attention towards threat was associated with later/delayed treatment response, and non-response was associated with the absence of pre-treatment threat bias. These findings indicate that attentional biases towards threat require more treatment sessions than cases characterised by threat avoidance and/or are more responsive to child–parent focused CBT than child-focused CBT. Another study indicated that there may be a nonlinear relationship between change in threat bias and change in anxiety symptoms (Waters et al. 2012). Similar to the findings reported for the first stage of Legerstee et al.’s (2010) study, Waters et al. (2012) found that attentional bias away from threat was significantly modified following 10 sessions of CBT plus a booster session and that bias towards threat was only nonsignificantly reduced. However, symptom reduction was greater in the children that demonstrated a pre-treatment attentional bias towards threat, indicating a nonlinear relationship between change in threat bias and change in anxiety. In summary, the reviewed studies report inconsistent findings regarding CBT’s effects on attentional biases. The most consistent finding reported in the studies is that attentional bias towards threat is not significantly changed with CBT (Manassis et al. 2013; Waters et al. 2012; Waters et al. 2008). Methodological limitations may account for some of the variability in findings, such as the type of stimulus used in the dot-probe task and variations in CBT protocol. Interpretation Bias According to an information-processing perspective of anxiety disorders (see Daleiden and Vasey 1997, for a review and elaboration), an important subsequent stage in the processing of threat-information is the more consciously controllable interpretation of stimuli as threatening or non-threatening. At this stage, in informationprocessing, children with anxiety disorders have been found to demonstrate interpretation bias: the tendency to appraise ambiguous situations and stimuli as threatening (Field et al. 2011). This bias has typically been examined by asking children to interpret ambiguous words (homophones, e.g. die versus dye), stories or pictures and evaluating the child’s perception of associated threat (Manassis 2013). For example, one of the more recent studies of interpretation bias in childhood anxiety disorders found that clinically anxious children interpreted ambiguous stories as more threatening than non-anxious children and 395 non-anxious ‘at risk’ children with anxious parents. The clinically anxious children demonstrated greater degrees of negative emotion and a reduced perception of influencing ability in relation to ambiguous stories (Waters et al. 2008). CBT’s Effects on Interpretation Bias The targeting of interpretation biases using cognitive reappraisal strategies and behavioural experiments represents a central element of CBT. Interpretation biases are considered to occur at a later stage of information-processing which is more amenable to conscious control, and as such, may be expected to be receptive to CBT (Manassis 2013). In line with this view, a study found that although attentional biases remained unchanged in anxious children following CBT, threat interpretation bias (as assessed on an ambiguous story task) significantly reduced (Waters et al. 2008). The Relationship Between Attentional and Interpretation Bias The relationship between attentional and interpretation bias has been examined in a randomised controlled trial (RCT) in a sample of female young adults (White et al. 2011). Using a modified version of the dot-probe task, the experimental group was trained to develop an attentional bias towards threat, whereas the control group viewed targets but was not trained to develop a bias. The group that underwent training to attend to threat was found to be more likely to make subsequent threatening interpretations of ambiguous information relative to controls. These findings suggest that threat-related biases in the earlier stages of processing affect subsequent processing at the interpretive stage. Onset and Acquisition of Attentional and Interpretation Bias How attentional and interpretation biases come to develop is not currently well understood (Vasey et al. 2014). Based on a developmentally focused review of the empirical literature, Field and Lester (2010) suggest that normative attentional biases to threat appear to be present early in development and that persistence of this bias depends on the moderating effect of developmental influences. In contrast, interpretation bias appears to develop in concordance with a child’s cognitive and social development and to causally influence the development of anxiety. Some evidence indicates that there may be a reciprocal relationship between child and parent factors in the development of interpretation biases, whereby parental expectancies of child threat-cognitions progressively 123 396 develop in response to the child’s anxious cognitions, and in turn escalate the child’s bias towards making anxious interpretations (Creswell et al. 2011). Clin Child Fam Psychol Rev (2016) 19:392–402 Improving Treatment Response in Paediatric Anxiety An Examination of Novel Interventions A Pivotal Role for Attentional Control in Paediatric Anxiety Disorders Several theoretical accounts underscore the key role attentional control plays in regard to threat-related biases (e.g. Eysenck et al. 2007). Attention gates the engagement of other cognitive processes (Shechner et al. 2012), and as such, attentional control is regarded by some researchers to be essential for all other forms of executive functioning (De Luca and Leventer 2008) and learning (Shechner et al. 2012). Elaborating on ideas articulated by Shechner et al. (2012), from this perspective, the impact of CBT on executive functions such as appraisal ability (the ability to accurately interpret situations and stimuli), emotion regulation and problem-solving, and the acquisition of new learning, may be limited due to attentional dysfunction. For example, attentional bias towards threat and difficulties disengaging from threat may limit the cognitive resources available for making rational appraisals, selecting adaptive emotion regulation strategies in the face of threat, and extinction learning. Attentional bias towards threat and difficulties disengaging from threat may also undermine extinction learning in exposure by maintaining a heightened level of fear that interferes with habituation. The idea that threat-related attentional dysfunction interferes with other aspects of cognition has some empirical support. Reinholdt-Dunne, Mogg and Bradley (2009) utilised an emotional Stroop face task to investigate the modulating effects of executive attentional control and trait anxiety on cognitive processing. Participants were asked to name the colours of angry, happy and neutral faces in order to measure interference effects of attention to emotional stimuli on the cognitive processing required for task performance. The findings revealed that a combination of poor attentional control and high trait anxiety were associated with significant interference in the processing of angry faces relative to neutral faces. These findings suggest that poor attentional control and high trait anxiety are associated with difficulty ignoring task-irrelevant emotional information, thereby limiting allocation of cognitive resources to the task. This study was based on a sample of mostly female, undergraduate adults, which limits generalisability of the findings and their meaning in relation to anxious children. In summary, based on the ideas presented here, insufficient amelioration of attentional bias may weaken the effects of CBT on the broader underlying psychopathology that characterises paediatric anxiety disorders. 123 Based on the ideas presented above, change in attentional dysfunction has the potential to initiate change in downstream cognitive processes that are mediated by attention, such as appraisal ability, learning acquisition and emotion regulation: core mechanisms through which CBT appears to operate. Treatment methods that target attentional dysfunction may therefore hold potential for improving the effectiveness of CBT. This section examines whether developments and insights from recent treatment literature can elucidate effective means of targeting attentional dysfunction in the treatment of paediatric anxiety disorders. Three novel interventions are discussed: attention bias modification training (ABMT), mindfulness-based therapy (MBT) and attention feedback awareness and control training (A-FACT). Attention Bias Modification Training Evidence implicating a key role for attentional biases in anxiety disorders has inspired the development of ABMT, a new treatment paradigm that uses the dot-probe task to modify attentional bias. In the treatment of attentional bias towards threat, the probe is repeatedly presented in the location of the neutral stimulus rather than the threat stimulus. After systematic repetition, expectance of this contingency is theorised to induce an implicitly learned (i.e. automatic) attentional bias away from the threat (BarHaim 2010). In recent years, there has been a surge of research examining the use of ABMT in the treatment of paediatric anxiety disorders. Reductions in anxiety with ABMT have been documented in a case series (Cowart and Ollendick 2011) and in larger RCTs utilising placebo ABMT protocols not designed to modulate distribution of attention (Bar-Haim et al. 2011; Eldar et al. 2012). Mixed evidence has been reported with regard to the placebo protocols. Some evidence indicates that placebo AMBT reduces anxiety (Bar-Haim et al. 2011), whereas other evidence suggests that it does not (Eldar et al. 2012). Some research has looked at the augmenting effects of ABMT. ABMT has been found to effectively augment CBT (Shechner et al. 2014) and CBT and pharmacotherapy (Riemann et al. 2013) in the treatment of paediatric anxiety disorders. The utility of ABMT in treatment non-responders has also been examined. A recent case series found that anxious children who had not previously responded to CBT Clin Child Fam Psychol Rev (2016) 19:392–402 exhibited significant reductions in child-report measures of anxiety with the training (Bechor et al. 2014). Additional findings documented in the ABMT literature include ameliorating effects on depression in anxious children (Bechor et al. 2014), generalisability of changes in attentional bias to new sets of stimuli (Shechner et al. 2014), and evidence indicating that exposure and desensitisation to threat stimuli is not a significant mechanism through which ABMT reduces anxiety (Eldar et al. 2012). Mixed findings have been reported regarding changes in threat bias with AMBT. ABMT has been found to significantly facilitate disengagement of attention from threat (Bar-Haim et al. 2011) and to promote a shift in attentional bias from towards threat to away from threat (Shechner et al. 2014). A couple of studies have reported pre-treatment bias away from threat in their samples (Bechor et al. 2014; Cowart and Ollendick 2011). These latter findings are inconsistent with the notion that symptom change is achieved via the reduction in attentional bias towards threat, and suggest that ABMT may operate through other mechanisms. Although ABMT emerged from studies demonstrating attentional bias towards threat in anxiety disorders and was designed to modify this bias, another viable explanation for the ameliorating effects of ABMT on anxiety is that ABMT promotes a more general improvement in flexible attentional control, irrespective of its valence-related directionality. This account offers an explanation for why ABMT has been found to reduce anxiety in the absence of pre-treatment bias towards threat. Understanding the underlying mechanisms of action in ABMT has important implications for who should receive ABMT. For example, not all children demonstrate an attentional bias towards threat, and it is not known whether these children would benefit from ABMT (Bar-Haim 2010). From the perspective that ABMT exerts its effects by eliminating an attentional bias towards threat, the use of ABMT would appear to be redundant in cases where there is no preexisting attentional bias towards threat. The concern has also been raised that if anxiety is associated with threat avoidance in some children, further training of attention away from threat may exacerbate symptoms (Shechner et al. 2012). However, evidence showing that ABMT significantly reduced anxiety in small samples including children with pre-treatment bias away from threat (Bechor et al. 2014; Cowart and Ollendick 2011) provides preliminary evidence that disputes this concern. If, on the other hand, ABMT leads to a general increase in flexible attentional control, children may benefit from treatment regardless of whether they exhibit a preexisting attentional bias (Bar-Haim 2010). Research findings that have shown placebo ABMT training to reduce anxiety (i.e.: Bar-Haim et al. 2011; Shechner et al. 2014) may be interpreted as evidence that 397 ABMT exerts its effects though improving attentional control. Placebo training diverts attention with equal probability both towards and away from threat. As such, reductions in anxiety may arise from practicing more flexible allocation of attention. Evidence that ABMT significantly reduced anxiety in small samples that included children with pre-treatment bias away from threat (Bechor et al. 2014; Cowart and Ollendick 2011) also suggests that the effects of ABMT on anxiety are unlikely to be due to the elimination of an attentional bias towards threat and fits in with the perspective that ABMT reduces anxiety through other mechanisms. None of the studies reviewed here were based on samples of young children, with the minimum age of the samples being 8 years for the majority of the studies. These findings therefore do not inform on the effectiveness of ABMT for the younger age range. At this stage, there is also a lack of follow-up studies examining the long-term impact of ABMT, which needs to be addressed in future research. Based on the limited evidence presented here, it appears that ABMT is associated with favourable shortterm treatment outcomes for paediatric anxiety. However, variations in study design such as differences in stimulus presentation time, number of sessions, and number of trials per session complicate between-study comparisons, and mixed and inconsistent findings preclude conclusions about the mechanisms through which ABMT exerts its effects. Lastly, it is also worth noting that a recent meta-analysis, looking at the effects of ABMT on anxiety and depression in adults, only observed small effects, which highlights the need for more research in this area before any firm conclusions can be drawn about its effects on emotional disorders (Hallion and Ruscio 2011). Mindfulness-Based Therapy According to Bishop et al. (2004), mindfulness promotes sustained attention, the ability to flexibly switch attentional focus and inhibition of elaborative processing and rumination and may be in part conceptualised as the self-regulation of attention. Based on this conceptualisation, MBT may offer another means of addressing attentional dysfunction in the treatment of paediatric anxiety disorders. Mindfulness is a meditative therapy that focuses attention towards current experience and involves observation of moment-to-moment changes in thoughts, feelings and physical sensations. Attention is anchored in present experience though sustained attention on the breath. When the mind wanders to attend to a thought, feeling or sensation, the experience is acknowledged, and attention is flexibly directed back to the anchor of the breath (Bishop et al. 2004). This practice is theorised to facilitate a decentred perspective, whereby thoughts and feelings are 123 398 experienced as transient events rather than objective reflections of reality, which interrupts elaborative processing and rumination (Segal et al. 2002). MBT is most commonly regarded as referring to the manualised programs of mindfulness-based stress reduction (MBSR; Kabat-Zinn 2005) and mindfulness-based cognitive therapy (MBCT; Segal et al. 2002), which consist of group mindfulness meditative practice and meditative homework over a course of 8 weeks. MBCT is closely modelled on MBSR, but is tied to a cognitive theory of recurrent depression and was developed for the treatment of this disorder. There has been increasing interest in MBT in recent years and its utility in alleviating a range of problems in adults, and a meta-analysis suggests that the intervention is effective in reducing anxiety, depressive, and stress symptoms (Khoury et al. 2013). More recently, an interest in the effect of MBCT in treating children and adolescents with psychological disorders has also arisen (Burke 2010). However, to date, only a small body of literature has examined the use of mindfulness-based approaches in paediatric samples. This section therefore draws upon evidence based on both paediatric and adults samples, to discuss whether MBT, as an adjunct to CBT, could offer potential in improving treatment response in the treatment of paediatric anxiety disorders. Another meta-analysis found MBT to be moderately effective in the treatment of anxiety symptoms in adults (Hofmann et al. 2010). Effect sizes were reported to be smaller but still significant in controlled studies relative to uncontrolled studies, and significantly greater than the effect sizes found for placebo treatment in a separate metaanalysis (Smits and Hofmann 2009). However, the majority of the studies reviewed examined the effects of MBT on anxiety symptoms in a range of psychiatric and medical disorders, with few studies specifically examining anxiety disorders. As such, these findings offer only tentative, preliminary evidence to suggest that MBT may be an effective treatment for adult anxiety disorders. More research examining the efficacy of MBT in the treatment of clinical levels of anxiety is needed, and research specifically targeted at paediatric samples is necessary to examine whether the intervention is effective with children and adolescents. MBCT has been adapted for use with children (Lee et al. 2008). An RCT of mindfulness-based cognitive therapy for children (MBCT-C) found the program to be associated with significant reductions in anxiety symptoms in a subsample of six children who reported clinically elevated symptoms at pre-treatment assessment (Semple et al. 2010). These findings tentatively suggest that MBT may be an effective treatment for paediatric as well as adult anxiety disorders and indicate some initial promise for the integration of mindfulness methods to improve treatment 123 Clin Child Fam Psychol Rev (2016) 19:392–402 response. However, outcome studies do not give any indications as to the mechanisms through which MBT exerts its effects. This is important, as the theoretical rationale presented here for the utility of MBT in enhancing CBT response rests on the assumption that MBT might target and ameliorate attentional dysfunction. Offering insight into this issue, a group of researchers tested Bishop et al.‘s (2004) theory that MBT strengthens attentional control via the promotion of sustained attention, switching of attention and inhibition of elaborate processing, in study of healthy adults (Anderson et al. 2007). Although a 12-week mindfulness program was associated with improvements in emotional well-being, no improvements in attentional control were found relative to a control group. These findings do not support the position that MBT improves attentional control and suggest that mindfulness exerts positive effects on well-being through other mechanisms. It has been suggested in this paper that the use of treatment methods that target attentional processing of threat may lead to downstream effects on other attentionmediated processes such as cognitive appraisal ability, emotion regulation and learning. Relevant to these ideas, a recent study compared cognitive reappraisal ability in adults with a history of MBCT, adults with a history of CBT, and adults without a history with any type of therapy (Troy et al. 2013). The study found a history of MBCT to be associated with higher cognitive reappraisal ability than a history of CBT or no-therapy. Post hoc analysis also revealed that a history of both MBCT and CBT was associated with significantly higher cognitive reappraisal ability scores than a history of CBT-only or no-therapy, and a nonsignificantly higher cognitive reappraisal ability than a history of only MBCT. This study suggests that the combined effects of MBT and CBT are associated with better functioning of attention-mediated downstream functions than CBT alone. Troy et al. (2013) suggest that future research should examine whether a hybridised therapy may lead to greater improvements in cognitive reappraisal ability than either therapy alone. In consideration of the theoretical discussion that has been presented in this paper, there would seem to be reasonable justification for further research examining this question in relation to paediatric anxiety disorders. Attention Feedback Awareness and Control Training Very recently, a new intervention paradigm has been proposed and trialed for its use in ameliorating attentional biases and the cascade of maladaptive effects mediated by these biases (Bernstein and Zvielli 2014). Drawing upon the importance of feedback systems in learning and selfregulation of behaviour, A-FACT provides real-time Clin Child Fam Psychol Rev (2016) 19:392–402 feedback about attentional allocation to facilitate attentional awareness and self-regulatory control of attention. Like ABMT, the treatment involves an emotional dotprobe paradigm. However, in contrast to ABMT, probe location is random. In addition, feedback about attentional allocation is provided at intervals throughout the task, and participants are encouraged to learn the feedback in order to reduce their bias and to attend equally to both threatening and neutral pictures. The effectiveness of A-FACT has been trialed in an RCT of highly anxious young-adult participants recruited from a university (Bernstein and Zvielli 2014). An active placebo condition received the same instructions and completed the same dot-probe task as a condition that received the A-FACT intervention, but did not receive the feedback. Baseline performance on the dot-probe task was also measured. The group that received the A-FACT intervention was found to exhibit a statistically and clinically significant reduction in attentional bias to threat relative to the active placebo control group: 68.2 % of the A-FACT group showed no attentional bias towards threat at post-intervention, compared to 33.3 % of participants in the control group. Participants in the A-FACT group also showed faster emotional recovery following exposure to anxiogenic stressors relative to the control group. The A-FACT condition also demonstrated a reduction in behavioural avoidance, although these effects were not statistically significant. In summary, these initial findings suggest that A-FACT may be a useful adjunctive treatment for anxiety disorders and may have downstream effects on variables that maintain anxiety and hinder its effective treatment. However, this study only provides preliminary data on the effectiveness of the intervention, and further research will be needed if these assumptions are to be substantiated. Whether the effects of the intervention extend to clinically anxious children also requires further research. An advantage of A-FACT is that it trains patients to direct their attention in a balanced and controlled way rather than training attentional biases in a particular direction. This methodology appears appropriate for treatment of anxiety regardless of whether the ameliorative effects of attention training are attributable to reductions in attentional bias towards threat or to improvements in attentional control, and regardless of whether a child demonstrates attentional bias towards or away from threat. Novel Interventions: Summary Treatment methods that target attentional dysfunction in paediatric anxiety disorders are in their early stages of development. On the theoretical side, these interventions appear to have the potential to improve treatment response by strengthening attentional control, ameliorating 399 attentional bias and freeing up cognitive capacity for more effective engagement in CBT. AMBT has received the most attention in the research literature as a treatment for anxious children, with the available evidence indicating that the intervention is associated with favourable shortterm treatment outcomes for paediatric anxiety. MBT was found to have the weakest empirical support, with the available evidence indicating that this intervention does not bring about improvements attentional control (Anderson et al. 2007). A-FACT is a new intervention with similarities to ABMT, which appears to deal with the issues raised regarding the underlying mechanisms of change (removing bias towards threat versus promoting more flexible attentional control), and cases in which children demonstrate attentional bias away from threat. A-FACT has only recently been trialed. Preliminary data indicates that the intervention warrants further investigation. Discussion Informed by the literature and research on informationprocessing, the present paper has reviewed potential limitations of CBT for paediatric anxiety disorders and how these limitations may be addressed. In recent years, therapies that include attentional training strategies have been developed, which emphasise the theoretical view that anxiety disorders are associated with selective attention towards threat-related information. More research is needed to examine the efficacy of these methods in the treatment of paediatric anxiety disorders, and whether they can improve response to CBT. Clinical Implications The dissemination of new therapies into clinical practice extends beyond efficacy, requiring consideration of practical issues such as treatment acceptability and accessibility, the optimal treatment package and cost. A preliminary discussion of this issue may help to elucidate potential challenges in the dissemination of these interventions, and how these challenges may be dealt with. Treatment Acceptability and Accessibility Treatment acceptability and accessibility may influence treatment-related variables such as motivation, compliance and attrition and ultimately treatment response. It is therefore important to consider new interventions in this context. The MBCT-C program involves 15 min of home practice, 6 days per week, and continued practice is deemed important. Although this seems reasonable in terms of time 123 400 demands, this guidance appears to be arbitrary, and much longer durations of home practice are recommended for adult versions of the therapy (Kabat-Zinn 2005; Segal et al. 2002). There may be developmental variation in the optimal duration of practice sessions. For example, children may benefit from increasingly longer practice durations as their executive functions mature. Depending on the required dosage for positive effects, there is potential for quite substantial time demands with MBT. This may present a challenge for children, particularly as it is not typically the child that requests treatment in the first place (Beidel and Alfano 2011). An additional challenge associated with MBT is the practitioner requirement of extensive personal experience with the practice (Kabat-Zinn 2005; Segal et al. 2002). Whether effective dissemination of a hybridised form of CBT and MBT necessitates clinician training in both interventions and extensive personal experience with MBT has important implications for the availability and accessibility of treatment. There is perhaps greater potential for wider dissemination of electronical-based treatments such as ABMT and A-FACT, as these treatments may be able to be delivered by professionals after limited training and may potentially also be independently utilised by the child or adolescent. Another advantage of these computer-based interventions is their potential for integration with computerised CBT. The increasing accessibility of electronic technology such as computers and smartphones means that these interventions may be made available at home, or at any other convenient time. Smartphone delivery of ABMT has been successfully trialled in adults for reducing social anxiety (Enock et al. 2014). ABMT was administered in three brief, daily sessions over a 4-week period, and pre-post analyses revealed significant reductions in social anxiety for the smartphone intervention compared to waitlist control. The accessibility of computer or smartphone delivered interventions also permits dosage to be manipulated according to the child’s needs and has the potential to reduce therapist contact time and treatment costs. Developmental considerations are also likely to be important for treatment acceptability and compliance. For example, less cognitively mature children may require more guidance and benefit more from therapies delivered by clinicians, whereas more cognitively mature or older children may find computer-based treatments more acceptable. The Optimal Treatment Package A central issue for dissemination of novel interventions in the treatment of paediatric anxiety disorders is ascertaining the optimal treatment package for optimal treatment response. For example, which components are necessary, 123 Clin Child Fam Psychol Rev (2016) 19:392–402 and at what dosage should they be integrated with CBT? Should CBT be kept in its current standardised format with techniques added to the treatment package, or could certain elements of CBT be removed or provided at a lower dosage to make room for new methods without adding to treatment length? Which of the interventions should be incorporated and to what degree? Should different CBT packages be developed for different problem profiles? Should the novel interventions be used as a primer to CBT, applied alongside CBT, or both? The optimal treatment package may also vary as a function of multiple variables such as developmental level, type of attentional dysfunction (e.g. direction of bias), personal preferences, comorbid disorders, gender and cultural influences. There are some risks associated with the development of new treatment methods and hybridised therapies. As the number of therapies grow, the risk of producing therapeutic generalists who do not have sufficient skill and training across each of the therapies increases. Clinicians also have to become familiarised with new techniques and receive continuous supervision, which takes time and is costly. Furthermore, endless combinations may make it close to impossible to research treatment effect given that all of the therapies are likely helpful and large samples will be needed to test all of the possibilities. Having said this, the introduction of new therapies is never without potential difficulties, and we must continue to advance the field. It should be noted that improving treatment outcomes for anxious children does not rest solely on developing the ultimate treatment package. The development of novel interventions provides a broader range of treatment options for anxious children and may help clinicians to match a treatment to a particular patient. Limitations The review should be considered in light of its limitations. Firstly, the review and its conclusions are based on limited literature from an emerging field. The validity of conclusions is in part dependent on the quality and limitations of the studies reviewed, with research limitations including the use of small sample sizes, self-report measures and issues with generalisability of findings across different stages of development, and the use of cross-sectional studies. Secondly, the review may be defined as exploratory rather than strictly systematic, in that the literature search was broad and flexible and was not restricted to a few listable search terms. This methodology opens up the possibility for researcher bias in the selection of relevant literature and limits possibilities for systematic replication of findings. However, the advantage of this methodology is that it allows for the integration of a broader range of Clin Child Fam Psychol Rev (2016) 19:392–402 information and greater flexibility in examining the research question. The systematic structure of the paper, the critical nature of the review and a reflexive researcher awareness over the potential for bias may be considered to minimise the limitations associated with this exploratory approach. Conclusion Consolidation of recent insights from developmental psychopathology and treatment literature indicates promise for improving response to CBT for paediatric anxiety disorders via the use of novel treatment methods that target attentional dysfunction. Further research is needed to substantiate these preliminary indications, and great challenges lie ahead with regard to ascertaining the optimal treatment package for optimal treatment response. References American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author. Anderson, N. D., Lau, M. A., Segal, Z. V., & Bishop, S. R. (2007). Mindfulness-based stress reduction and attentional control. Clinical Psychology and Psychotherapy, 14(6), 449–463. doi:10.1002/cpp.544. Bar-Haim, Y. (2010). Research review: Attention bias modification (ABM): A novel treatment for anxiety disorders. Journal of Child Psychology and Psychiatry, 51(8), 859–870. doi:10.1111/j. 1469-7610.2010.02251.x. Bar-Haim, Y., Morag, I., & Glickman, S. (2011). Training anxious children to disengage attention from threat: A randomized controlled trial. Journal of Child Psychology and Psychiatry, 52(8), 861–869. doi:10.1111/j.1469-7610.2011.02368.x. Bechor, M., Pettit, J. W., Silverman, W. K., Bar-Haim, Y., Abend, R., Pine, D. S., et al. (2014). Attention bias modification treatment for children with anxiety disorders who do not respond to cognitive behavioral therapy: A case series. Journal of Anxiety Disorders, 28(2), 154–159. doi:10.1016/j.janxdis.2013.09.001. Beesdo-Baum, K., & Knappe, S. (2012). Developmental epidemiology of anxiety disorders. Child and Adolescent Psychiatric Clinics of North America, 21, 457–478. doi:10.1016/j.chc.2012. 05.001. Beidel, D. C., & Alfano, C. A. (2011). Child anxiety disorders (2nd ed.). New York, NY: Routledge. Bernstein, A., & Zvielli, A. (2014). Attention feedback awareness and control training (A-FACT): Experimental test of a novel intervention paradigm targeting attentional bias. Behaviour Research and Therapy, 55, 18–26. doi:10.1016/j.brat.2014.01.003. Bishop, S. R., Lau, M., Shapiro, S., Carlson, L., Anderson, N. D., Carmody, J., et al. (2004). Mindfulness: A proposed operational definition. Clinical Psychology: Science and Practice, 11(3), 230–241. doi:10.1093/clipsy.bph077. Blackford, J. U., & Pine, D. S. (2012). Neural substrates of childhood anxiety disorders: A review of neuroimaging findings. Child and Adolescent Psychiatric Clinics of North America, 21(3), 501–525. doi:10.1016/j.chc.2012.05.002. 401 Breinholst, S., Esbjørn, B. H., Reinholdt-Dunne, M. L., & Stallard, P. (2012). CBT for the treatment of child anxiety disorders: A review of why parental involvement has not enhanced outcomes. Journal of Anxiety Disorders, 26(3), 416–424. doi:10.1016/j. janxdis.2011.12.014. Britton, J. C., Lissek, S., Grillon, C., Norcross, M. A., & Pine, D. S. (2011). Development of anxiety: The role of threat appraisal and fear learning. Depression and Anxiety, 28(1), 5–17. doi:10.1002/ da.20733. Burke, C. (2010). Mindfulness-based approaches with children and adolescents: A preliminary review of current research in an emergent field. Journal of Child and Family Studies, 19(2), 133–144. doi:10.1007/s10826-009-9282-x. Cisler, J. M., & Koster, E. H. W. (2010). Mechanisms of attentional biases towards threat in anxiety disorders: An integrative review. Clinical Psychology Review, 30, 203–216. doi:10.1016/j.cpr. 2009.11.003. Clark, D. A., & Beck, A. T. (2010). Cognitive theory and therapy of anxiety and depression: Convergence with neurobiological findings. Trends in Cognitive Sciences, 14(9), 418–424. doi:10. 1016/j.tics.2010.06.007. Cowart, M. J. W., & Ollendick, T. H. (2011). Attention training in socially anxious children: A multiple baseline design analysis. Journal of Anxiety Disorders, 25(7), 972–977. doi:10.1016/j. janxdis.2011.06.005. Creswell, C., Shildrick, S., & Field, A. (2011). Interpretation of ambiguity in children: A prospective study of associations with anxiety and parental interpretations. Journal of Child and Family Studies, 20(2), 240–250. doi:10.1007/s10826-010-9390-7. Daleiden, E. L., & Vasey, M. V. (1997). An information-processing perspective on childhood anxiety. Clinical Psychology Review, 17(4), 407–429. doi:10.1016/S0272-7358(97)00010-X. De Luca, C. R., & Leventer, R. J. (2008). Developmental trajectories of executive functions across the lifespan. In V. Anderson, R. Jacobs, & P.-J. Anderson (Eds.), Executive functions and the frontal lobes: A lifetime perspective (pp. 23–47). Hove, East Sussex: Psychology Press. Eldar, S., Apter, A., Lotan, D., Edgar, K. P., Naim, R., Fox, N. A., et al. (2012). Attention bias modification treatment for pediatric anxiety disorders: A randomized controlled trial. American Journal of Psychiatry, 169(2), 213–220. doi:10.1176/appi.ajp. 2011.11060886. Enock, P. M., Hofmann, S. G., & McNally, R. J. (2014). Attention bias modification training via smartphone to reduce social anxiety: A randomized, controlled multi-session experiment. Cognitive Therapy and Research, 38(2), 200–216. doi:10.1007/ s10608-014-9606-z. Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7(2), 336. Field, A. P., Hadwin, J. A., & Lester, K. J. (2011). Information processing biases in child and adolescent anxiety: A developmental perspective. In W. K. Silverman & A. P. Field (Eds.), Anxiety disorders in children and adolescents (2nd ed., pp. 103–128). Cambridge: Cambridge University Press. Field, A. P., & Lester, K. J. (2010). Is there room for ‘‘development’’ in developmental models of information processing biases to threat in children and adolescents? Clinical Child and Family Psychology Review, 13(4), 315–332. doi:10.1007/s10567-0100078-8. Hallion, L. S., & Ruscio, A. M. (2011). A meta-analysis of the effect of cognitive bias modification on anxiety and depression. Psychological Bulletin, 137(6), 940–958. doi:10.1037/a0024355. Hofmann, S. G., Asmundson, G. J. G., & Beck, A. T. (2013). The science of cognitive therapy. Behavior Therapy, 44(2), 199–212. doi:10.1016/j.beth.2009.01.007. 123 402 Hofmann, S. G., Ellard, K. K., & Siegle, G. J. (2012). Neurobiological correlates of cognitions in fear and anxiety: A cognitive– neurobiological information-processing model. Cognition and Emotion, 26(2), 282–299. doi:10.1080/02699931.2011.579414. Hofmann, S. G., Sawyer, A. T., Witt, A. A., & Oh, D. (2010). The effect of mindfulness-based therapy on anxiety and depression: A meta-analytic review. Journal of Consulting and Clinical Psychology, 78(2), 169–183. doi:10.1037/a0018555. Hudson, J. L., Keers, R., Roberts, S., Coleman, J. R., Breen, G., Arendt, K., et al. (2015). Clinical predictors of response to cognitive-behavioral therapy in pediatric anxiety disorders: The genes for treatment (GxT) study. Journal of the American Academy of Child and Adolescent Psychiatry, 54(6), 454–463. doi:10.1016/j.jaac.2015.03.018. Kabat-Zinn, J. (2005). Full catastrophe living: How to cope with stress, pain and illness using mindfulness meditation. London: Piatkus Books Ltd. Kendall, P. C., Safford, S., Flannery-Schroeder, E., & Webb, A. (2004). Child anxiety treatment: Outcomes in adolescence and impact on substance use and depression at 7.4-year follow-up. Journal of Consulting and Clinical Psychology, 72(2), 276–287. doi:10.1037/0022-006X.72.2.276. Khoury, B., Lecomte, T., Fortin, G., Masse, M., Therien, P., Bouchard, V., et al. (2013). Mindfulness-based therapy: A comprehensive meta-analysis. Clinical Psychology Review, 33(6), 763–771. Lee, J., Semple, R. J., Rosa, D., & Miller, L. (2008). Mindfulnessbased cognitive therapy for children: Results of a pilot study. Journal of Cognitive Psychotherapy, 22(1), 15–28. doi:10.1891/ 0889.8391.22.1.15. Legerstee, J. S., Tulen, J. H. M., Dierckx, B., Treffers, P. D. A., Verhulst, F. C., & Utens, E. M. W. J. (2010). CBT for childhood anxiety disorders: Differential changes in selective attention between treatment responders and non-responders. Journal of Child Psychology and Psychiatry, 51(2), 162–172. doi:10.1111/j. 1469-7610.2009.02143.x. Mackintosh, B., Mathews, A., Yiend, J., Ridgeway, V., & Cook, E. (2006). Induced biases in emotional interpretation influence stress vulnerability and endure despite changes in context. Behavior Therapy, 37(3), 209–222. doi:10.1016/j.beth.2006.03.001. MacLeod, C., Rutherford, E., Campbell, L., Ebsworthy, G., & Holker, L. (2002). Selective attention and emotional vulnerability: Assessing the causal basis of their association through the experimental manipulation of attentional bias. Journal of Abnormal Psychology, 111(1), 107–123. doi:10.1037/0021-843X.111.1.107. Manassis, K. (2013). Cognitive findings in childhood anxiety: Translations for clinical practice. Translational Neuroscience, 4(1), 88–95. doi:10.2478/s13380-013-0110-9. Manassis, K., Hum, K. M., Lee, T. C., Zhang, G., & Lewis, M. D. (2013). Threat perception predicts cognitive behavioral therapy outcomes in anxious children. Open Journal of Psychiatry, 3(1), 141–148. doi:10.4236/ojpsych.2013.31A009. Manassis, K., Lee, T. C., Bennett, K., Zhao, X. Y., Mendlowitz, S., Duda, S., et al. (2014). Types of parental involvement in CBT with anxious youth: A preliminary meta-analysis. Journal of Consulting and Clinical Psychology, 82(6), 1163–1172. doi:10. 1037/a0036969. Merikangas, K. R., He, J., Burstein, M., Swanson, S. A., Avenevoli, S., Cui, L., et al. (2010). Lifetime prevalence of mental disorders in US adolescents: Results from the national comorbidity studyadolescent supplement (NCS-A). Journal of the American Academy of Child and Adolescent Psychiatry, 49(10), 980–989. doi:10.1016/j.jaac.2010.05.017. Mohr, C., & Schneider, S. (2013). Anxiety disorders. European Child and Adolescent Psychiatry, 22(1), 17–22. doi:10.1007/s00787012-0356-8. 123 Clin Child Fam Psychol Rev (2016) 19:392–402 Reinholdt-Dunne, M. L., Mogg, K., & Bradley, B. P. (2009). Effects of anxiety and attention control on processing pictorial and linguistic emotional information. Behaviour Research and Therapy, 47(5), 410–417. doi:10.1016/j.brat.2009.01.012. Riemann, B. C., Kuckertz, J. M., Rozenman, M., Weersing, V. R., & Amir, N. (2013). Augmentation of youth cognitive behavioral and pharmacological interventions with attention modification: A preliminary investigation. Depression and Anxiety, 30(9), 822–828. doi:10.1002/da.22127. Salum, G. A., DeSousa, D. A., Conceição do Rosário, M., Pine, D. S., & Manfro, G. G. (2013). Pediatric anxiety disorders: From neuroscience to evidence-based clinical practice. Revista Brasileira de Psiquiatria, 35, 3–21. doi:10.1590/1516-4446-2013-S108. Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-based cognitive therapy for depression. New York, NY: Guilford Press. Semple, R. J., Lee, J., Rosa, D., & Miller, L. F. (2010). A randomized trial of mindfulness-based cognitive therapy for children: Promoting mindful attention to enhance social-emotional resiliency in children. Journal of Child and Family Studies, 19(2), 218–229. doi:10.1007/s10826-009-9301-y. Shechner, T., Britton, J. C., Perez-Edgar, K., Bar-Haim, Y., Ernst, M., Fox, N. A., et al. (2012). Attention biases, anxiety, and development: Toward or away from threats or rewards? Depression and Anxiety, 29(4), 282–294. doi:10.1002/da.20914. Shechner, T., Rimon-Chakir, A., Britton, J. C., Lotan, D., Apter, A., Bliese, P. D., et al. (2014). Attention bias modification treatment augmenting effects on cognitive behavioral therapy in children with anxiety: Randomized controlled trial. Journal of the American Academy of Child and Adolescent Psychiatry, 53(1), 61–71. doi:10.1016/j.jaac.2013.09.016. Smits, J. A. J., & Hofmann, S. G. (2009). A meta-analytic review of the effects of psychotherapy control conditions for anxiety disorders. Psychological Medicine, 39(02), 229–239. doi:10. 1017/S0033291708003498. Troy, A. S., Shallcross, A. J., Davis, T. S., & Mauss, I. B. (2013). History of mindfulness-based cognitive therapy is associated with increased cognitive reappraisal ability. Mindfulness, 4(3), 213–222. doi:10.1007/s12671-012-0114-5. Vasey, M. W., Bosmans, G., & Ollendick, T. H. (2014). The developmental psychopathology of anxiety. In M. Lewis, & K. D. Rudolph (Eds.), Handbook of developmental psychopathology (pp. 543–560). US: Springer. http://link.springer.com.ep.fjernad gang.kb.dk/chapter/10.1007/978-1-4614-9608-3_27. Waters, A. M., Craske, M. G., Bergman, R. L., & Treanor, M. (2008a). Threat interpretation bias as a vulnerability factor in childhood anxiety disorders. Behaviour Research and Therapy, 46(1), 39–47. doi:10.1016/j.brat.2007.10.002. Waters, A. M., Mogg, K., & Bradley, B. P. (2012). Direction of threat attention bias predicts treatment outcome in anxious children receiving cognitive-behavioural therapy. Behaviour Research and Therapy, 50(6), 428–434. doi:10.1016/j.brat.2012.03.006. Waters, A. M., Wharton, T. A., Zimmer-Gembeck, M. J., & Craske, M. G. (2008b). Threat-based cognitive biases in anxious children: Comparison with non-anxious children before and after cognitive behavioural treatment. Behaviour Research and Therapy, 46(3), 358–374. doi:10.1016/j.brat.2008.01.002. White, L. K., Suway, J. G., Pine, D. S., Bar-Haim, Y., & Fox, N. A. (2011). Cascading effects: The influence of attention bias to threat on the interpretation of ambiguous information. Behaviour Research and Therapy, 49(4), 244–251. doi:10.1016/j.brat.2011. 01.004. Wilson, E. J., MacLeod, C., Mathews, A., & Rutherford, E. M. (2006). The causal role of interpretive bias in anxiety reactivity. Journal of Abnormal Psychology, 115(1), 103–111. doi:10.1037/ 0021-843X.115.1.103. Deviant Behavior ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/udbh20 Racial Differences in the Applicability of Bronfenbrenner’s Ecological Model for Adolescent Bullying Involvement Jun Sung Hong, Simon C. Hunter, Jinwon Kim, Alex R. Piquero & Chelsey Narvey To cite this article: Jun Sung Hong, Simon C. Hunter, Jinwon Kim, Alex R. Piquero & Chelsey Narvey (2021) Racial Differences in the Applicability of Bronfenbrenner’s Ecological Model for Adolescent Bullying Involvement, Deviant Behavior, 42:3, 404-424, DOI: 10.1080/01639625.2019.1680086 To link to this article: https://doi.org/10.1080/01639625.2019.1680086 Published online: 23 Oct 2019. Submit your article to this journal Article views: 5956 View related articles View Crossmark data Citing articles: 7 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=udbh20 DEVIANT BEHAVIOR 2021, VOL. 42, NO. 3, 404–424 https://doi.org/10.1080/01639625.2019.1680086 Racial Differences in the Applicability of Bronfenbrenner’s Ecological Model for Adolescent Bullying Involvement Jun Sung Honga, Simon C. Hunterb, Jinwon Kimc, Alex R. Piquerod,e, and Chelsey Narveyd a Wayne State University, Detroit, MI, USA; bUniversity of Strathclyde, Glasgow, UK; cSungkyunkwan University, Seoul, South Korea; dUniversity of Texas at Dallas, Richardson, TX, USA; eMonash University, Melbourne, Australia ABSTRACT Objectives: Social scientists have devoted much theoretical and empirical attention to studying the correlates of bullying perpetration and victimization. Much less attention has been devoted to studying race differences in the correlates of bullying behaviors despite the importance of these when designing effective and focused prevention and intervention programs. Methods: Utilizing data from the 2009 to 2010 Health Behavior in SchoolAged Children (HBSC) study in the United States, this study applies Bronfenbrenner’s ecological model to bullying in order to examine how various interrelated systems are associated with bullying perpetration, victimization, and their concordance in a nationally representative sample of adolescents. Results: Findings shown important similarities, as well as some differences, across race in how key parental and peer relationships relate to aspects of involvement in bullying. Directions for future research are noted. ARTICLE HISTORY Received 18 April 2019 Accepted 7 October 2019 Bullying is a type of behavior that is repeatedly perpetrated by an individual or a group of individuals against a target (Gladden et al. 2014). Recent national data indicate that in 2017, about 20% of students (ages 12–18) reported being bullied during the school year; of those who reported being bullied, about 41% thought bullying would occur repeatedly (Musu-Gillette et al. 2019). The prevalence of bullying, coupled with high levels of maladjustment that it is associated with, has led to widespread anti-bullying efforts (Birkland and Lawrence 2009; Hall 2017). Anti-bullying programs have been widely developed and their effectiveness has been tested (Gaffney, Farrington, and Ttofi 2019; Merrell et al. 2008; Scherr and Larson 2010; Ttofi and Farrington 2011). According to the Bureau of Justice Statistics, 75.5% of public schools provide some form of training to teachers and aides in recognizing bullying (Zhang, Musu-Gillette, and Oudekerk 2016). Despite these efforts, findings have been inconsistent (Ferguson et al. 2007; Hall 2017). Several studies evaluating a widely used anti-bullying program in U.S. schools have reported positive results (Black and Jackson 2007; Limber et al. 2004). However, one study on the effectiveness of this program in 10 public middle schools reported that victimization decreased among Whites, but no similar effects were found for other racial groups (Bauer, Lozano, and Rivara 2007). This may be, in part at least, because there is little understanding of the different causes and processes underpinning the use and experience of bullying across different racial groups. Significant differences exist between Black and White youth with respect to a number of different risk factors. Black youth are more likely to reside in disadvantaged neighborhoods, have compromised familial situations, be exposed to violence, and have limited educational opportunities and attainment (Piquero 2015; Wilson 1987). As a result, CONTACT Jun Sung Hong fl4684@wayne.edu University of Texas at Dallas, Richardson, TX, USA & Monash University, Melbourne, Australia Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/udbh. © 2019 Taylor & Francis Group, LLC DEVIANT BEHAVIOR 405 bullying programs that do not pay attention to these differences and incorporate them into programmatic efforts will likely not have the same effect on Blacks as on Whites. Implementing best practices for bullying requires a comprehensive understanding and description of bullying and victimization risks across racially diverse youth. Scholars have proposed an ecological approach to assessing factors related to the risk that certain youth have to be involved in bullying (Shetgiri, Lin, and Flores 2013; You, Kim, and Kim 2014) as well as an ecologically based prevention strategy (Espelage 2004). The central tenet of the ecological perspective is that adolescent development is shaped by the ongoing qualities of various social settings in which the youth is embedded (Bronfenbrenner 1979). While a large percentage of school districts provide bully-recognition training to teachers, Bronfenbrenner’s (1979, 1994) perspective underscores the importance of recognizing the quality of teachers and conditions of schools that might differ across individuals, specifically those who are Black and are more likely to come from a lower socioeconomic status (SES) background. Moreover, Black and White youth differ in their accumulated exposure to multiple environmental risk factors (Piquero 2015). This accumulated exposure is crucial to understand the different needs that individuals might have. In particular, adolescents differ in their susceptibility toward environmental influences, both positive and negative (Belsky, Bakermans-Kranenburg, and van IJzendoornm 2002). These differences might be especially apparent for individuals of different racial/ethnic backgrounds, who have been exposed to significant differences in Bronfenbrenner’s nested structures. Moreover, the individuals and groups that comprise an adolescent’s microsystem might interact differently across the races. For instance, White adolescents might have parents that are more involved in their school. If this is the case, then a more thorough understanding is necessary in order to ensure programming is sensitive to the differences within the environment in which these schools are located and from which the adolescents are living. While an increasing number of bullying programs exist, some programs might be more effective than others because of these differences. Accordingly, the aim of this study is to apply the ecological model to explore whether factors related to bullying, victimization, and bullying/victimization are similar across Whites and Blacks in the U.S. Theoretical framework Bronfenbrenner (1977, 1979) proposed that individual development and behavior can be influenced by the ecological environment, which is regarded as a set of interrelated, nested structures. An individual is an inseparable part of multiple, interrelated systems that shape adolescent developmental processes, including the microsystem (relations of individuals with immediate settings), mesosystem (interrelations among the microsystems), exosystem (settings which do not directly influence the individual), and macrosystem (cultural or subcultural patterns) (Bronfenbrenner 1977, 1979). An important aspect of the ecological model is that developmental influences (e.g., peer relations) are shaped by the characteristics of the community in which the youth resides (Szapocznik and Coatsworth 1999). These influences contribute toward the racial identity development of adolescents and have consequences for their psychosocial wellbeing (Hughes et al. 2006). For years, research has been conducted on the risk and protective factors of bullying and victimization at the systems noted above. Microsystem-level factors include occurrences and relationships in the immediate environment, such as dynamics in the home, peer groups, and school. In the home setting, research reveals that parental monitoring, parent-adolescent communications, and parental supports reduced bullying and victimization risks (Conners-Burrow et al. 2009; Elsaesser et al. 2017). Theories, from attachment theory to social learning theory, have been applied to account for how relations with parents might influence adolescents’ bullying involvement (Hong et al. 2018). Attachment theorists might argue that youth with insecure attachment with their parents through lack of parental monitoring, communication, and support might be at an elevated risk of victimization because they may find it difficult connecting with their peers (Allen et al. 2007). 406 J. S. HONG ET AL. With respect to peer-level factors, bullying and victimization are positively linked to deviant peer affiliation (Espelage, Holt, and Henkel 2003) but are negatively linked with supportive friendships and time spent with peers (Bollmer et al. 2005; Kendrick, Jutengren, and Stattin 2012). Deviant peer affiliation can increase adolescents’ problem behaviors, which are often learned and reinforced in peer groups (Elliott and Menard 1996). Youth who regularly associate with deviant peers also have an increased risk of victimization, as they are perceived by their peers as potential targets due to low guardianships (Lauritsen, Sampson, and Laub 1991). Research also offers support for the potential protective functioning of supportive friendships and time spent with peers, such as providing a buffer against victimization (Bollmer et al. 2005). School-level factors have been researched extensively, and protective factors in school that are found to diminish bullying risks include teacher support, teachers’ involvement, and school bonding (Flaspohler et al. 2009; Wei et al. 2010). School environment is recognized as a salient influence in an adolescent’s adjustment (Aspy et al. 2012), and research shows that the more exposure adolescents have to environmental assets, the less likely they are involved in violent behaviors (Aspy et al. 2012). Moreover, a positive school environment can function to enhance the adoption of and commitment to prevention program as well as to increase help-seeking behavior, which can reduce bullying risk (Bradshaw et al. 2009; Eliot et al. 2010). In terms of mesosystem-level, although the home is the main context in which child development occurs – especially in the first five to 6 years of life before formal schooling begins, it is but one of numerous settings in which developmental process(es) can and do take place (Bronfenbrenner 1979). This system level is conceptualized as the interrelations among two or more microsystems (e.g., family and peer groups), each of which includes the individual (Bronfenbrenner 1994). Examples of mesosystems are interrelations between the adolescent’s peer group or school and the home environment. For instance, parental involvement and interactions with others (e.g., teachers) can influence adolescents’ behavior and interactions with peers in school (Lee and Song 2012). Involvement in violence can be reinforced through deviant peer association (Akers 1998), which may occur as a result of weak bonds, as indicated by, for example, a lack of communication and interactions in the home. Research on exosystem- and macrosystem-level factors related to bullying and victimization is limited. This is unfortunate as psychological development of adolescents is influenced not only by direct settings (e.g., home) but also by broader level occurrences which may affect the adolescent’s interactions in these settings, such as economic conditions (Bronfenbrenner 1979). Exosystem is defined as linkages and processes between two or more settings. However, only one directly affects the individual (Bronfenbrenner 1979). Macrosystem is defined as the cultural “blueprint” that may influence the social structures and activities occurring in the immediate system levels (Bronfenbrenner 1994). Examples of the macrosystem are “material resources, opportunity structures, alternatives throughout the life course, lifestyles and customs, and shared knowledge and cultural beliefs” (Eamon 2000:261). Some studies have explored macrosystem-level factors, including SES, income inequality, and poverty, and how they might elevate bullying risk in adolescents. Findings suggest that poverty and residence in communities with highincome inequality are associated with victimization (Carlson 2006; Chaux, Molano, and Podlesky 2009; Elgar et al. 2009). According to Carlson (2006), higher levels of poverty were associated with victimization. In a wider sense, adolescents in countries with high-income inequality report more bullying than those in countries with low-income inequality (Elgar et al. 2009). Poverty is related to power differentials between those with access to resources and those without access, which might lead to bullying perpetrated by those with more power over those with less power (Chaux, Molano, and Podlesky 2009). Race and bullying It has been reported that bullying involvement varies across race (Scherr and Larson 2010), although there is a more complex picture concerning involvement. Studies have documented that Black adolescents are involved in more perpetration, relative to adolescents of other racial groups DEVIANT BEHAVIOR 407 (Carlyle and Steinman 2007; Wang, Iannotti, and Nansel 2009), while other studies report no racial differences (e.g., Seals and Young 2003). In addition, Blacks experience higher rates of victimization than adolescents of other races (Koo, Peguero, and Shekarkhar 2012; Rhee, Lee, and Jung 2017). Also, according to the Department of Justice, more Black students (20%) reported being frequently teased, made fun of or called names, or socially excluded than White students (15%) (Zhang, MusuGillette, and Oudekerk 2016). In contrast, according to Juvonen, Graham, and Schuster (2003), Whites were significantly more likely to be classified as victims than their Black, Hispanic, and Asian peers. Sawyer, Bradshaw, and O’Brennan (2008) also found that Black youth tended to be less likely than their White peers to indicate being a bullying victim. Spriggs et al. (2007) found that parental communication, social isolation, and relations with classmates were negatively associated with bullying across racial/ethnic groups, but that living with two biological parents was a protective factor for Whites only. The study also found that two school-level factors, satisfaction, and performance, were negatively related to bullying for Whites yet were irrelevant for Blacks. In a more recent study, fathers’ parental monitoring was found to be negatively related to bullying for Whites, while not significant for Blacks (Hong, Ryou, and Piquero 2017). These results offer some (albeit limited) support for the contention that there may be distinctive ways in which ecological factors operate in the lives of adolescents of different racial or ethnic groups. The present study The present study builds on Hong, Ryou, and Piquero's (2017) study, which explored family-level factors related to bullying and victimization experiences of Blacks and Whites. More specifically, we investigate whether there are racial differences in ecological level factors associated with subtypes of bullying involvement (perpetration, victimization, bully/victim) at the microsystem, mesosystem, and macrosystem. This study contributes to the literature in several respects. First, studies have found inconsistent results with respect to differences in bullying perpetration and victimization, suggesting the need for more research on this topic. Moreover, research needs to look not only at the differences in rates but also in understanding the underlying factors. By examining factors at the microsystem, mesosystem, and macrosystem this study provides a more thorough and detailed background on the differences in variables associated with bullying involvement across race. Given the differences in exposure to risk factors that Black and White adolescents experience, there is a reason to believe that these factors may operate differently across race. Understanding these differences is critical for anti-bullying program implementation because awareness of potential differences between races in Bronfenbrenner’s ecological model can help to ensure that victims, perpetrators, and bully/victims are provided with the appropriate intervention for their specific needs. The research questions are as follows: (1) Are the microsystem, mesosystem, and macrosystem factors differentially associated with bullying for White and Black youth when controlling for sex and age? (2) Are the microsystem, mesosystem, and macrosystem factors differentially associated with victimization for both racial groups when controlling for sex and age? and (3) Are the microsystem, mesosystem, and macrosystem factors differentially associated with bullying/victimization for both racial groups when controlling for sex and age? Methods Sample and data Data were derived from the 2009 to 2010 Health Behavior in School-Aged Children (HBSC) study in the U.S. The HBSC is a standardized, international World Health Organization study consisting of repeated cross-sectional surveys in the 43 participating countries. Data were collected through school-based surveys utilizing random sampling to select a proportion of adolescents, aged 11, 13, and 15 years (Currie et al. 2012). The primary sampling units (districts 408 J. S. HONG ET AL. comprising one or more public schools) were stratified within each Census Division. The districts were classified as urban or rural, based on a comprehensive list of schools from the Quality Education Data. The primary sampling units had at least 10 schools, and those with large enrollments were considered as separate primary sampling units. A total of 1,302 primary school units were created, and a sample of 94 primary school units were selected. Also, a list of private and Catholic schools were obtained from the Quality Education Data and were assigned based on their locations to the 1,302 primary sampling units. All private and Catholic schools were eligible for inclusion into the 94 sampled primary school units. In the second stage, schools were selected from the sampled primary school units, and 314 schools participated in the study. In the final stage, classes were selected from the schools designated for sampling students from specific grades. Respondents consisted of public, Catholic, and private school students in grades 5–10 in 50 states and the District of Columbia. In the original sampling, 475 schools were considered to be eligible. Of these schools, 161 schools did not participate, and of the 314 schools, 31 did not complete the questionnaire. The school-based survey includes a self-reported questionnaire completed by students in the classroom and covers a range of health indicators and health-related behaviors, along with life circumstances (Roberts et al. 2009). Survey questions include information on socio-demographic factors, social background, social context, health outcomes, health behaviors, and risk behaviors (Roberts et al. 2009). The survey took approximately 45 minutes to complete and was administered in a classroom by teachers who read scripts that explained the procedure. Data for the study are from the cross-sectional 2009–2010 data set. Table 1 presents the descriptive statistics for the total sample, White sample, and Black sample. Table 2 shows a cross-tabulation of the four bullying subgroups (uninvolved, victims-only, bullies-only, bully/victims) across the two racial groups, the results of which indicate a significant association between the two variables (χ2 = 29.56, p < .001, φc = .082). As is clear from the standardized residuals reported in Table 2, the significant effect was driven by certain roles that Blacks and Whites take on. Blacks were overrepresented in the uninvolved role and underrepresented in the victim and the bully/victim roles. In contrast, Whites were overrepresented in the victim role. Table 1. Descriptive statistics of the study variables. Total (N =4,466) N(%) Age Sex Male Female Parental monitoring Mother’s parental monitoring Father’s parental monitoring Parent-child communication Elder brother/sister communication Parental support Parental treatment Number of friends Time spent with friends/peers Delinquent friend influences Positive peer relations in school Family socioeconomic status Bullying victimization Bullying perpetration Bully-victim Uninvolved Victim only Bullying only Bully/victim M 13.88 Whites (N =3,386) SD 1.26 2,223(49.8) 2,243(50.2) M 13.85 Blacks (N =1,080) SD 1.24 1,701(50.2) 1,685(49.8) 10.97 9.39 7.13 4.58 10.16 4.06 7.07 10.29 8.13 10.95 3.43 13.92 12.66 2,509(56.2) 632(14.2) 773(17.3) 552(12.4) N(%) 1.55 2.77 2.00 2.49 1.79 1.15 1.32 5.11 3.99 2.49 0.91 5.62 4.60 M 13.97 SD 1.34 10.73 8.13 6.74 5.37 9.95 3.97 7.10 10.95 8.24 11.02 3.38 14.07 13.17 1.59 3.14 2.16 2.75 1.86 1.24 1.35 5.31 4.26 2.62 .96 6.37 5.44 522(48.3) 558(51.7) 11.05 9.79 7.25 4.32 10.22 4.09 7.05 10.08 8.10 10.92 3.45 13.87 12.50 1,845(54.5) 524(15.5) 572(16.9) 445(13.1) N(%) 1.54 2.51 1.93 2.35 1.76 1.13 1.31 5.03 3.90 2.44 .89 5.35 4.28 664(61.5) 108(10.0) 201(18.6) 107(9.9) DEVIANT BEHAVIOR 409 Table 2. Cross-tabulation for race by bully subgroups, showing n, row-percentages, and standardized residuals. Bully Subgroups Race Black adolescents White adolescents Total n % z n % z N % Uninvolved 664 60.8% 2.1 1,845 54.5% −1.2 2,509 56.2% Victims-only 108 10.2% −3.5 524 15.5% 1.9 632 14.2% Bullies-only 201 19.0% 1.2 572 16.9% −0.7 773 17.3% Bully/victims 107 10.1% −2.1 445 13.1% 1.2 552 12.4% Total 1,080 100.0% χ2 29.56a φc .082a 3,386 100.0% 4,466 100.0% p < .001. a Measures Perpetration was measured with the question, “How often have you bullied another student(s) at school in the past couple of months in the way listed below” with eleven subcategories including: (a) “I called another student(s) mean names, and made fun of, or teased him or her in a hurtful way; (b) “I kept another student(s) out of things on purpose, excluded him or her from my group of friends, or completely ignored him or her”; (c) “I hit, kicked, pushed, shoved around, or locked another student(s) indoors”; (d) “I spread false rumors about another student(s) and tried to make others dislike him or her”; (e) “I bullied another student(s) with mean names and comments about his or her race or color”; (f) “I bullied another student(s) with mean names and comments about his or her religion”; (g) “I made sexual jokes, comments, or gestures to another student(s)”; (h) “I bullied another student(s) using a computer or e-mail messages or pictures”; (i) “I bullied another student(s) using a cell phone”; (j) “I bullied others outside of school using a computer or email messages or pictures”; and (k) “I bullied others outside of school using a cell phone”. Response options are 0 = I have not bullied another student in this way in the past couple of months, 1 = it has only happened once or twice, 2 = 2 or 3 times a month, 3 = about once a week, and 4 = several times a week. The final perpetration measure is the sum of the eleven items (α = .92). Victimization was measured with the following question, “How often got bullied” with eleven subcategories that are identical to the perpetration items noted above but were re-worded to reflect victimization (e.g., “I was called names, was made fun of, or teased in a hurtful way.”) (α = .88). Response options are also identical to perpetration but reworded to reflect victimization. Bully/victim was measured using two items, “How often got bullied” and “How often have you bullied another student(s) at school in the past couple of months”. Response options are 0 = I haven’t been bullied/ haven’t bullied another student at school the past couple of months, 1 = only once or twice, 2 = 2 or 3 times a month, 3 = about once a week, and 4 = several times a week. All responses were dichotomized as 0 = I haven’t been bullied/haven’t bullied and 1 = I have been bullied/bullied more than once, and then combined. These dichotomized responses were classified into four clusters: 1 = uninvolved, 2 = victimonly, 3 = bully-only, and 4 = bully/victim. Microsystem variables included family-level factors. Parental monitoring was measured with the questions, “How much does your mother (or female guardian) really know about … ?” and “How much does your father (or male guardian) really know about … ?” with the following subcategories, “Who your friends are”, “Where you are after school”, and “Where you go at night”. Response options initially were: 1 = s/he knows a lot, 2 = s/he knows a little, 3 = s/he doesn’t know anything, and 4 = don’t have/see mother/father/ guardian and were reverse coded. They were summed for each item. Parental monitoring was divided into “by mother” (α = .75) and “by father” (α = .91), and the variables were summed, respectively, to either mother or father subscales. Parent–child communication was measured with the same question asked twice (once for “mother” and once for “father”): “How easy is it for you to talk to the following persons about things that really bother you?” The response options initially were: 1 = very easy, 2 = easy, 3 = difficult, 4 = very difficult, and 5 = don’t have or see this person; they were reverse coded. The two items were summed. Elder brother/sister communication was also measured with the same question asked twice: “How is it for you to talk to the following persons about things that really bother you?” This question was asked for “Elder 410 J. S. HONG ET AL. brother(s)” and “Elder sister(s)”. Response options initially were 1 = very easy, 2 = easy, 3 = difficult, 4 = very difficult, and 5 = don’t have or see this person. They were reverse coded and were summed for the two items. Parental support was measured with the statement, “My parents/guardian” with the following subcategories, “helps me as much as I need”, “understands my problems and worries”, and “makes me feel better when I am upset” (α = .80). Response options initially were 1 = almost always to 4 = don’t have or don’t see parents/guardians; they were reverse coded. Parental treatment consists of one question, “Have your parent(s) treated you fairly?” with response options, 1 = never to 5 = always. Also included are peer-level factors. Number of friends was measured with the question, “At present, how many close male and female friends do you have?” with the response option for males and females, 1 = none to 4 = three or more. Time spent with friends/peers was measured with three questions, “How many days per week do you usually spend time with friends right after school?”, “How many evenings per week do you usually spend out with your friends”, and “How often do you talk to your friend(s) on the phone or send them text messages or have contacts through the internet?” (α = .64). Response options for the first question range from 0 days to 6 days, from 0 evenings to 7 evenings for the second question, and 1 = rarely or never to 5 = every day for the last question. Since the three questions have different response options, linear transformation was applied for the response options of the first and last questions in order to convert them to a common metric. The range of the response for the three items therefore was adjusted from 0 to 7. Delinquent friend influences were measured with the question, “How many of your friends would you estimate … ” with the following subcategories: (a) smoke cigarettes, (b) drink alcohol, (c) get drunk at least once a week, (d) smoke/use marijuana, and (e) carry a weapon (α = .88). Response options range from 1 = none to 5 = all. Positive peer relations in school was measured with the following three statements (α = .74): “The student in my class(es) enjoy being together”, “Most of the students in my class(es) are kind and helpful”, and “Other students accept me as I am”. Response options were 1 = strongly agree to 5 = strongly disagree but were reverse coded so that higher scores reflect more positive peer relations. Mesosystem variables included delinquent friend influences × parent–child communication and delinquent friend influence × elder brother/sister communication, which were generated using meancentred versions of the relevant variables (Aiken, West, and Reno 1991). The macrosystem variable, family SES, was measured with the question, “How well off do you think your family is?” Response options were 1 = very well off to 5 = not at all well off but were reverse coded so that a higher score reflects higher family SES. Covariates as originally measured in the study include age (“How old are you?”; 1 = 10 or younger, 2 = 11, 3 = 12, 4 = 13, 5 = 14, 6 = 15, 7 = 17, and 8 = 17 or older) and sex (“Are you a boy or a girl?”; 0 = boy and 1 = girl). Analyses Analyses included bivariate correlations, hierarchical multivariate regressions, and multinomial regressions separately for the White (N = 3,386) and Black (N = 1,080) samples. Multivariate regressions for victimization and perpetration were estimated using Ordinary Least Squares regression. To compare racial differences, coefficient comparisons were conducted using Paternoster et al.’s (1998) formula. To ease the interpretation of the results regarding bully-victims, results of multinomial logistic regression were converted into Relative Risk Ratios (RRR). All interaction terms were based on Aiken, West, and Reno (1991) analysis and interpretation methods of interaction effects in multiple regression. Simple slope analysis was used to interpret the interaction effect. Multinomial regressions were used to examine racial differences in adolescents’ status as a bully, victim, or bully/victim (compared to uninvolved status). Analyses were conducted using SPSS 18.0 and STATA 12 software.1 None of the correlations exceeded r = 0.51, which limits potential problems associated with collinearity in the model space. 1 DEVIANT BEHAVIOR 411 Results Tables 3 and 4 display the results of hierarchical multivariate regression for Whites and Blacks for victimization and perpetration, respectively. Hierarchical multivariate regression results In terms of victimization for Whites (see Table 3), we found that mother’s parental monitoring (B = −.13, p < .05), parental support (B = −.16, p < .05), parental treatment (B = −.46, p < .001), and positive peer relations in school (B = −.56, p < .001) were negatively related to victimization. On the other hand, parent–child communication (B = .13, p < .05), elder brother/sister communication (B = .09, p < .05), and delinquent friend influences (B = .13, p < .001) were positively associated with victimization. The interaction terms were not significant, nor did they alter the significance of the coefficient estimates reported above. With respect to victimization for the Black adolescent sample (see Table 3, Model B1), mother’s parental monitoring (B = −.41, p < .001), parental treatment (B = −.70, p < .001), number of friends (B = −.36, p < .05), and positive peer relations in school (B = −.40, p < .001) were negatively and significantly related to victimization in anticipated ways. Regarding the interaction terms, although the main effects of delinquent friend influences and parent–child communication on victimization were not significant, the interaction between delinquent friend influences × parent–child communication (B = −.06, p < .01) was negatively associated with victimization.2 Figure 1 displays the results of simple slope analysis for this particular interaction term for Blacks. As can be seen, the effect of high delinquent friends on victimization is diminished when parent– child communication is high. Conversely, when parent–child communication is low and delinquent friend influences are at their highest point, victimization is at its highest point. Regarding perpetration for Whites (see Table 4), mother’s parental monitoring (B = −.27, p < .001) and positive peer relations in school (B = −.16, p < .001) were negatively and significantly associated with perpetration. Time spent with friends/peers (B = .08, p < .001) and delinquent friend influences (B = .26, p < .001) exerted positive effects on perpetration. The main effects of parent–child communication and elder brother/sister communication were not significantly associated with perpetration, but interaction terms were found to be positive and significantly related to perpetration: delinquent friend influences × parent–child communication (B = .02, p < .05) and delinquent friend influences × elder brother/sister communication (B = .03, p < .001). As shown in Figure 2, when high delinquent peer influences are coupled with higher parent–child communication (easier in communication), perpetration risk is higher than when the corresponding variables are their low points. The same is observed for elder brother/sister communication and delinquent friend influences. For perpetration for Blacks, we found that parent–child communication (B = .20, p < .05), time spent with friends/peers (B = .10, p < .01), and delinquent friend influences (B = .18, p < .001) were positively associated with perpetration (see Table 4). Parental treatment (B = −.48, p < .01), number of friends (B = −.27, p < .05), and positive peer relations in school (B = −.26, p < .001) were negatively related to perpetration. Regarding the interaction terms (Model B2), delinquent friend influences × parent–child communication (B = −.06, p < .001) was negatively associated with perpetration.3 This is contrary to the 2 For the coefficient comparison tests across race, the corresponding Z statistics (Z-test) were calculated revealing mother’s parental monitoring (Z = 1.99, p < .05) and delinquent friend influences × parent–child communication (Z = 3.37, p < .001) were significant. This indicates that the effects of mother’s parental monitoring and delinquent friend influences × parent–child communication were significantly different between Whites and Blacks. Apart from the significant variables, the results of the corresponding Z-test indicate few differences between the two samples with respect to how the covariates relate to victimization. 3 Regarding the coefficient comparison tests on perpetration for Whites and Blacks, parental treatment (z = 2.12, p < .05), number of friends (z = 2.52, p < .05), delinquent friend influences × parent–child communication (z = 4.00, p < .001), and delinquent friend influences × elder brother/sister communication (z = 2.17, p < .05) were found to be significant, indicating that the coefficient estimates for these variables are significantly different from one another across race. *p < .05; **p < .01; ***p < .001. Age Sex Parental monitoring Mother’s parental monitoring Father’s parental monitoring Parent–child communication Elder brother/sister communication Parental support Parental treatment Number of friends Time spent with friends/peers Delinquent friend influences Positive peer relations in school Delinquent friend influences × Parent–child communication Delinquent friend influences × Elder brother/sister communication Family socioeconomic status Constant .06 .04 .06 .04 .07 .09 .07 .02 .03 .04 –.13* –.06 .13* .09* –.16* –.46*** –.13 –.01 .13*** –.56*** –.04 –.03 .05 .04 –.05 –.10 –.03 –.01 .09 –.26 β –.14 .03 .06 .04 .06 .04 .07 .09 .07 .02 .03 .04 .01 .01 SE .08 .18 Model A2 –.04 –.03 .05 .04 –.06 –.10 –.03 –.01 .10 –.26 .03 .03 β –.14 .03 –.11 .10 –.02 35.07*** 1.45 ΔR2 = .002, ΔF = 3.031* –.14* –.05 .13* .09* –.17* –.46*** –.13 –.01 .14*** –.56*** .02 .02 B –.59*** .29 Whites –.10 .10 –.02 34.87*** 1.45 R2 = .137, F = 41.151*** SE .08 .18 B –.59*** .27 Model A1 Table 3. Multivariate regression results for bullying victimization by race. .13 .07 .11 .07 .13 .18 .14 .04 .05 .08 SE .15 .39 –.10 .00 –.01 .06 .00 –.14 –.08 .03 .05 –.17 Β –.04 .03 B –.16 .43 .13 .07 .11 .07 .13 .18 .14 .04 .05 .08 .02 .02 SE .15 .39 Model B2 –.11 .00 –.01 .05 .01 –.13 –.08 .02 .03 –.17 –.10 –.04 β –.03 .03 –.02 .20 .00 29.94*** 3.01 ΔR2 = .011, ΔF = 6.535** –.43*** –.01 –.01 .12 .02 –.69*** –.35* .03 .04 –.41*** –.06** –.02 Blacks –.05 .20 –.01 30.53*** 3.01 R2 = .096, F = 8.680*** –.41** .01 –.03 .13 .01 –.70*** –.36* .03 .08 –.40*** B –.20 .40 Model B1 .95 –1.83 3.37*** 1.23 –.48 –1.20 1.21 1.47 1.99* Z –2.37* 412 J. S. HONG ET AL. *p < .05; **p < .01; ***p < .001. Age Sex Parental monitoring Mother’s parental monitoring Father’s parental monitoring Parent–child communication Elder brother/sister communication Parental support Parental treatment Number of friends Time spent with friends/peers Delinquent friend influences Positive peer relations in school Delinquent friend influences × Parent–child communication Delinquent friend influences × Elder brother/sister communication Family socioeconomic status Constant .05 .03 .05 .03 .05 .08 .06 .02 .02 .03 –.27*** –.04 .09 .06* –.05 –.12 .08 .08*** .26*** –.16*** –.10 –.02 .04 .03 –.02 –.03 .02 .09 .24 ...
Purchase answer to see full attachment