Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Aug 27;33(5):e70330. doi: 10.1002/cpp.70330

Imagery Interventions and Cognitive‐Behavioural Therapy for Psychological Complaints Related to Bullying: A Systematic Review and Meta‐Analysis

Meng Liang 1,✉, Gerly Tamm 1,2,3, Olivia De Laere 1, Yannick Vander Zwalmen 1, Ernst H W Koster 1
PMCID: PMC13520821  PMID: 42658011

ABSTRACT

Bullying is associated with internalizing symptoms and impaired everyday functioning. Cognitive‐behavioural therapy (CBT) and imagery interventions have been studied as treatments to improve mental health after such experiences.

This meta‐analysis aimed to evaluate the effects of individual CBT and imagery interventions in targeting internalizing symptoms (negative affect, depression, anxiety and post‐traumatic stress symptoms) and improving self‐perception and coping in bullying victims.

Systematic searches of PubMed, PsycInfo, Embase and Web of Science identified 23 studies applying CBT (n = 15) and imagery interventions (n = 8) in bullying victims. Risk of bias and random‐effects meta‐analyses were conducted for pre ‐ post and pre ‐ follow‐up comparisons. Sensitivity analyses and meta‐regressions explored potential moderators.

Unadjusted analyses showed reductions in internalizing symptoms pre‐ to post ‐ intervention (k = 21, g = 0.72). These symptoms were maintained at follow‐up assessments (k = 9, g = 0.54). Similar initial improvements emerged for self‐perception and coping outcomes (k = 15, g = 0.75). However, substantial heterogeneity existed (I 2 = 68%–98%). Crucially, after adjusting for publication bias, the pre–post internalizing effect dropped to g = 0.26 and was no longer statistically significant. Conversely, the adjusted effect for self‐perception and coping increased to g = 0.98, whereas the follow‐up internalizing effect decreased (g = 0.37) but remained significant.

Both CBT and imagery interventions may offer potential benefits for bullying victims in reducing internalizing symptoms and in improving self‐perception and coping abilities. However, large heterogeneity, methodological limitations and publication bias warrant cautious interpretation. Further high‐quality controlled trials are needed to establish the robustness and clinical efficacy of these interventions.

Keywords: bullying, cognitive‐behavioural therapy, coping strategies, imagery interventions, internalizing symptoms, meta‐analysis

Summary

  • CBT and imagery interventions were associated with improvements in internalizing symptoms and self‐perception/coping outcomes among bullying victims.

  • Improvements in internalizing symptoms appeared to be maintained at follow‐up assessment.

  • Publication bias adjustment substantially reduced the observed effects for internalizing symptoms.

  • Substantial heterogeneity and methodological limitations warrant cautious interpretation of the findings.

Key Practitioner Points

  • CBT and imagery interventions show potential for supporting the mental health of bullying victims.

  • Intervention benefits may be particularly evident for self‐perception and coping outcomes.

  • Evidence for reductions in internalizing symptoms is less certain after accounting for potential publication bias.

  • Follow‐up findings suggest that some intervention gains may be maintained over time.

1. Introduction

Bullying is conceptualized as a form of interpersonal aversive experience embedded within a peer social context (Idsoe et al. 2021). It is defined as intentional and repetitive harmful behaviour between peers with significant power imbalance between the perpetrator and victim, such as differences in size, strength, age or social status (Olweus 1993). The victimization can take physical, verbal or relational forms (Kennedy 2020). Bullying victims are often shy, socially marginalized and have poorer academic performance compared to their peers (Zequinão et al. 2017). Many also avoid school to escape victimization (Rothon et al. 2010). The experience of bullying can have detrimental effects on self‐esteem, sense of belonging and acceptance (Cross et al. 2011).

Childhood peer victimization is a primary risk factor for enduring mental health challenges, as shown in multiple longitudinal studies (Reijntjes et al. 2010). For instance, bullying in primary school has been shown to predict higher rates of internalizing problems 2 years after the initial victimization (Arseneault et al. 2008). Recent research confirms that these risks for depression, anxiety and self‐harm persist into mid‐20s (Armitage et al. 2022). Large‐scale, long‐term follow‐up studies have documented a significantly increased incidence of depression and broad psychiatric disorders in early adulthood and up to 32 years later, after controlling for early childhood adjustment and family‐level factors (Farrington et al. 2011; Sourander et al. 2007). Notably, a 50‐year national prospective cohort study showed that people who were frequently bullied in childhood had a small but significant increased risk of suicide in adulthood (Geoffroy et al. 2023). These findings underscore that the psychological impact of bullying often persists long after the victimization has ceased. Moreover, meta‐analytic evidence indicates that victims are approximately two times more likely to develop depressive disorders than their nonvictimized peers (Ttofi et al. 2011), whereas 57% report symptoms of PTSD at clinical levels (Nielsen et al. 2015). It is important to note that bullying does not qualify as traumatic exposure in the diagnostic sense unless it involves ‘Criterion A’ events, such as actual or threatened death or serious injury (Friedman 2006). Yet, its psychological impact often follows a stress‐trauma continuum (Tehrani 2004). Although isolated peer conflicts may constitute situational stress, chronic and inescapable victimization can result in trauma‐related symptoms (Idsoe et al. 2012). Although prevention projects by governments, schools and families have been put forward to reduce bullying behaviour, it remains challenging to fully prevent bullying and its associated internalizing problems (Hutchings et al. 2025; Sun et al. 2025).

Cognitive‐Behavioural Therapy (CBT) is widely recommended as a first‐choice psychological treatment for children and adolescents with psychological complaints, characterized by persistent low mood, excessive worry, social withdrawal and post‐traumatic intrusions (Beck 2020; Gesteira et al. 2026). Bullying victims who experienced a highly threatening environment for many years have a heightened risk to develop maladaptive cognitive schemas regarding imminent danger, the trustworthiness of others and personal powerlessness and self‐worth (Friedman 2006). Initial research on the efficacy of CBT that addresses these schemas demonstrates that victims report significant improvements in internalizing complaints (Berry and Hunt 2009; Camargo et al. 2023; Fung 2018; Lydecker et al. 2024). Camargo et al. (2022) found that improving emotion regulation skills such as reappraisal of the bullying events resulted in lower depressive and anxious symptoms. These therapeutic effects are often sustained for months or even a year after intervention (DeRosier 2004; Fung 2018). Additionally, CBT appears to be effective in alleviating distress and emotional difficulties among bullying victims. For instance, Fox and Boulton (2003) and DeRosier (2004) integrated social skill training into a group intervention format, aiming to improve self‐esteem and reduced anxiety by addressing social deficits. Similarly, Austin et al. (2018) utilized an affirmative CBT framework specifically designed for transgender bullying victims, focusing on identity intersectionality. Furthermore, other training interventions have been tailored not only to directly target bullying victims but also to their social environments, including classmates, schools and parents, to improve social climate and alleviate the emotional distress, specifically by reducing depressive affect and bolstering victims' sense of self‐worth (Vliek et al. 2019).

In recent years, psychological interventions focused on mental imagery have gained increased attention in the treatment of traumatic experiences. Imagery interventions are a therapeutic technique focusing on perceptual and affective reprocessing of memories (Holmes et al. 2006). Imagery interventions aim to directly modify traumatic mental representations through positive imagery, often incorporating multisensory elements such as sound and touch to enhance emotional processing (Holmes et al. 2006). International clinical guidelines recommend imagery interventions as first‐line treatments for trauma such as Trauma‐Focused Cognitive‐Behavioural Therapy (TF‐CBT), including imaginal exposure and imagery rescripting (ImRs), and eye movement desensitization and reprocessing (Ganslmeier et al. 2025). In such interventions, aversive memories are rescripted in a more safe or positive way (Hagenaars and Arntz 2012; Romano et al. 2020). Although these imagery techniques are often an important component of CBT protocols (Hackmann et al. 2011), they have typically been examined as independent, stand‐alone interventions within the bullying literature. For instance, Watson et al. (2016) found that ImRs could effectively alleviate the immediate stress associated with victimization. Several studies have reported positive outcomes from ImRs for bullying victims, including reductions in intrusions and significant relief from negative affect regardless of the imaginal content (Watson et al. 2016, 2021; Yamada et al. 2018). Norton et al. (2021) further demonstrated that even two sessions of ImRs could significantly change victims' evaluations, maladaptive appraisals and negative core beliefs.

Although the number of primary studies examining CBT and imagery interventions for bullying victims has increased in recent years, findings remain inconsistent across studies. Many individual studies relied on relatively small samples, limiting statistical power and the precision of effect estimates. To our knowledge, no one has systematically reviewed the differences between the effects of CBT and imagery interventions in bullying victims. Synthesizing the available evidence allows for a more reliable estimate of overall treatment effects. Furthermore, where CBT primarily targets maladaptive cognitions, imagery interventions focus on modifying distressing memories. To date, their relative efficacy has not been systematically compared.

To provide a wide‐ranging synthesis, this study aimed to capture the effects of CBT and imagery interventions on two primary outcome domains: one related to mental health and the second to resilience. Beyond estimating overall effects, this study aimed to explore potential sources of variability that may influence treatment outcomes. Against this background, two key questions were posed: (1) Are imagery interventions as effective as CBT in alleviating negative emotional impact on bullying victims? (2) Which moderators (e.g., number of sessions and intervention duration) contribute to the effectiveness of these interventions? We sought to address these questions through a systematic review and meta‐analysis of the emerging literature.

2. Methods

The study was preregistered in the International Prospective Register of Systemic Reviews (PROSPERO; registration number: CRD42025610437) and conducted in accordance with the Preferred Reporting Items for Systemic Reviews and Meta‐Analysis statement (PRISMA; Page et al. 2021).

2.1. Search Strategy

Searches were carried out in the following electronic databases: PubMed, PsycInfo, Embase and Web of Science on 30 April 2025, and a second search to update the findings was conducted on 30 August 2025. The search was not limited to peer‐reviewed publications so that dissertations and other indexed grey literature could be included. The following search string was used to search on title and abstract level: (‘imag* exposure’ OR ‘imag* rescripting’ OR ‘imag* morphing’ OR ‘emotive imagery’ OR ‘cognitive behavior therap*’ OR ‘cognitive behaviour therap*’ OR ‘CBT’) AND (‘bully*’ or ‘bulli*’ OR ‘victim*’ OR ‘verbal abuse’ OR ‘physical abuse’ OR ‘interpersonal violence’ OR ‘aggression’ OR ‘ostraci*’ OR ‘exclusion’ OR ‘rejection’ OR ‘harassment’ OR ‘mobbing’). To reduce publication bias, we employed a snowballing strategy (backward and forward citation tracking) via Google Scholar for all studies meeting the inclusion criteria.

2.2. Inclusion and Exclusion Criteria

The following inclusion criteria were used: (1) Study Objective: The goal of the publication should be the evaluation of the effectiveness of an intervention for bullying victims; (2) Intervention Type: Interventions should include CBT or CBT‐based interventions (e.g., Cognitive Processing Therapy (CPT) or social skill training) or imagery interventions (e.g., imagery rescripting); (3) Target Population: Studies included participants with experiences of being bullied if they utilized standardized measures to assess bullying victimization or if participants self‐identified as victims according to the criteria defined by the original study authors; (4) Outcome Measures: Outcomes included internalizing symptoms or self‐perception and coping strategy as a primary outcome, measured using validated, standardized questionnaires; (5) Data Availability: The study reports effect sizes or relevant data that allows for the computation of effect sizes; (6) Study Design: The design includes pre, post or follow‐up assessments; (7) Language: The paper is written in English. Purely qualitative studies and case studies were excluded.

2.3. Study Selection Process and Data Extraction

There were initially 3484 studies obtained through database searching, and those results were imported into reference manager software Endnote for review. After removing 1029 duplicate articles, there were 2455 studies remaining. Two authors (ML and ODL) conducted the title/abstract screening phase and full‐text screening independently. In cases of disagreement between the two reviewers, the third author (GT) joined to resolve the conflicts. In addition, nine more articles were found through snowballing (citation tracking), and one further study was incorporated following a reviewer's recommendation. Although grey literature was sought through these methods, no unpublished studies met the eligibility requirements. The search and selection process is visualized in the PRISMA flow diagram (see Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

PRISMA flow diagram.

In the data extraction phase, the information extracted included (1) article information: title, publication year and country; (2) participant information: N (per group), age, male/female ratio and ongoing bullying or not; (3) intervention information (such as type of interventions, number of sessions, duration, follow‐up time, individual or group intervention); and (4) outcome information: effect size or data to allow calculation of Hedges' g effect sizes.

2.4. Risk of Bias (RoB) Assessment

The RoB assessment acts as a quality check for each study. An adapted version of the Downs and Black checklist was used to evaluate potential bias in the included studies (Downs and Black 1998). Five domains of bias were assessed: (1) bias arising from the randomization process, (2) bias due to deviations from intended interventions, (3) bias related to missing outcome data, (4) bias in the measurement of outcomes and (5) bias in the selection of reported results. Each study was evaluated across these five domains and rated as ‘low’, ‘some concerns’ or ‘high risk’. A ‘high risk’ rating suggests that the study's findings might be influenced by flaws in its design or execution, and ‘some concerns’ reflects a lack of sufficient information or minor methodological issues that could impact the reliability of the findings. Two authors (ML and ODL) independently carried out the RoB assessments, and disagreements were resolved through discussion.

2.5. Data Analysis

R (version 4.5.1) was used for analyses. A list of the R packages and their version numbers can be found in Table S1. Main outcomes of this study were grouped into two domains: (1) Internalizing symptoms specifically incorporated measures of depressive, anxiety and trauma‐related symptoms (Kovacs and Devlin 1998); (2) measures of self‐perception and regulatory coping mechanisms (e.g., self‐compassion, self‐esteem, self‐efficacy and cognitive emotion regulation). When a study reported multiple outcomes within the same outcome domain (e.g., depression, anxiety or stress), these outcomes were combined into a single study‐level effect size to maintain statistical independence and avoid double‐counting participants. Composite effect sizes and their variances were calculated following the procedures described by Cuijpers (2016). Because most primary studies did not report correlations among outcomes, a correlation of r = 0.50 was assumed between different variables. For between‐group comparisons, effect sizes were calculated based on postintervention or follow‐up differences between groups. If Hedges' g was reported that effect size was used. When other effect sizes were reported (e.g., Cohen's d, Eta Squared), the effect sizes were converted into Hedges' g. When no effect size was reported, postmeans Hedges' g was calculated using the reported means, SDs and sample sizes (Borenstein 2013). For within‐study designs, the pre–post or prefollow‐up correlations were used to account for their dependence. When no correlations were reported, authors were contacted for this information. If no response was received after reaching out three times, a correlation of 0.5 was assumed (Cuijpers 2016). Sensitivity analyses using alternative plausible correlation values (r = 0.30 and r = 0.70) were conducted for studies with missing pre–post correlations and for the calculation of composite effect sizes requiring assumed correlations among outcomes, in order to evaluate the robustness of the pooled estimates. The effect sizes were interpreted according to Cohen's recommendations (0.2 as ‘small’, 0.5 as ‘medium’ and 0.8 as ‘large’; Cohen et al. 2016).

Heterogeneity among studies was investigated using the Q test and quantified by I 2 , which represents the percentage of total variation across studies due to heterogeneity. I 2 is reported with its 95% confidence interval (Ioannidis et al. 2007). An I 2 < 30% is considered low heterogeneity, 31%–60% is considered moderate heterogeneity, 61%–75% is considered high heterogeneity, and 76%–100% is considered very high heterogeneity (Higgins et al. 2021). Heterogeneity for the primary outcome was visualized using Baujat plots (Baujat et al. 2002) to visually examine which studies contributed most to the heterogeneity. In addition, sensitivity analyses were conducted to examine the influence of individual studies on I 2 and to identify whether any study affected the overall heterogeneity.

To examine heterogeneity related to intervention type, subgroup analyses were conducted for CBT and imagery interventions. Intervention types were classified independently by two reviewers (ML and ODL), with a third reviewer (GT) serving as an arbitrator. Meta‐regression analyses were also performed to explore the effects of number of sessions, intervention modality (individual vs. group), participant age and gender. Based on study sample size, multivariate meta‐regression was conducted for subgroups with more than 10 studies, whereas univariate meta‐regression was performed for subgroups with fewer than 10 studies (Higgins et al. 2021). To summarize the meta‐analysis, forest plots were generated for the two outcome domains at pre – post phase and pre ‐ follow‐up phase, displaying individual study effect sizes with 95% confidence intervals, as well as the overall pooled effect size and its 95% confidence interval. Finally, the pooled effect size for the primary outcome was converted to a Number Needed to Treat (NNT) to clarify clinical relevance, calculated using the R package dmetar (Harrer et al. 2022), following the Kraemer and Kupfer method (Kraemer and Kupfer 2006). Given the inclusion of different designs in our meta‐analysis, the NNT values are defined here as the number of individuals treated to observe one case of clinically meaningful improvement. These values should be interpreted as an estimate of clinical utility within the observed samples. Publication bias was examined visually by funnel plots and quantified by Egger's regression and Duval and Tweedie's trim‐and‐fill procedure (Duval and Tweedie 2000). An alpha error of < 0.05 was considered a criterion for statistical significance.

Grading of Recommendations Assessment, Development and Evaluation (GRADE) was used to determine the certainty of our overall evidence. Because most included studies used uncontrolled pre–post designs, the baseline certainty of evidence was rated as low. Risk of bias was evaluated using an adapted Downs and Black checklist to assess whether limitations in study design were present that affect the quality of evidence. Inconsistency was assessed based on statistical heterogeneity thresholds, with substantial heterogeneity resulting in downgrading of evidence certainty where appropriate. Indirectness was defined as the mismatch between the PICO elements (e.g., population, intervention, comparison and outcome) and existing actual evidence. Imprecision was assessed by optimal information size and confidence intervals, considering factors such as the consistency of results across studies, directness of the evidence and the impact of publication bias (Puhan et al. 2014). Within the GRADE framework, if nonrandomized studies meet certain criteria (e.g., Large Magnitude of Effect, Dose–Response Gradient and Plausible Confounding), the certainty of the evidence can be upgraded by one or two levels. However, any upgrading is strictly precluded if the evidence suffers from serious limitations in other domains, such as serious risk of bias or inconsistency (Guyatt et al. 2011).

3. Results

3.1. Study Characteristics

For the title/abstract screening, the overall agreement between the two reviewers was high (interrater reliability Cohen's κ = 0.70). In the full‐text screening, 13 studies were included from searching the online databases. The interrater reliability at the second stage was high (κ = 0.75). There were 23 studies included in this meta‐analysis, with a total of 2460 participants. The interrater reliability of two reviewers for self‐perception and coping indices was acceptable (κ = 0.68), which was similar for internalized symptom indices (κ = 1.00). Six articles initially met inclusion criteria but were excluded because no response was received when we requested the necessary data from the authors. Across all included studies, 1199 participants were male (48.7%). The mean age of participants was 15.7 (SD = 5.63) years old. The sample size ranged from 8 (Austin et al. 2018) to 415 (DeRosier 2004). A total of 15 studies were classified as CBT interventions. These were categorized into two subgroups based on protocol adherence. Standard CBT (n = 3) strictly followed traditional CBT protocols focusing on cognitive restructuring and behavioural activation. CBT‐based interventions integrated CBT elements with other therapeutic components (e.g., social skill training, teacher and family involvement) or were specifically tailored to the unique needs of bullying victims (e.g., gender identity). Regarding imagery interventions, eight studies were categorized as stand‐alone imagery rescripting and imagery‐focused interventions. Stand‐alone imagery rescripting (n = 5) utilized imagery rescripting as primary therapeutic technique following established protocols (Arntz and Weertman 1999). Imagery‐focused interventions (n = 3) utilized the imaginal processing of trauma memories as therapeutic mechanism (e.g., via trauma narrative or imaginal exposure) (Ganslmeier et al. 2025). Notably, one TF‐CBT study was classified in this group because the intervention was primarily focused on the trauma narrative that was distinct from the verbal cognitive restructuring of CBT (Lydecker et al. 2024).

A total of 23 articles were included in the final analysis. The availability of data was as follows: 22 articles provided pre‐ and post ‐ intervention data, and 12 articles provided follow‐up data, with durations ranging widely from 1 week (Dugué et al. 2019) to 1 year (Fung 2018). Ten studies provided group comparisons, and 13 studies were single‐arm designs. Table 1 lists the characteristics of the included studies and the extracted data. A total of 44 effect sizes were extracted: 21 for internalizing symptoms between pre‐ and post ‐ comparison, nine for internalizing symptoms between pre‐ and follow‐up comparison and 14 for self‐perception and coping ability between pre‐ and post ‐ comparison. There were no follow‐up effect sizes for self‐perception and coping ability.

TABLE 1.

Characteristics of included studies.

Study Country Intervention name Intervention category Sample size (N total) Age Session Male% Single session duration (min) Follow‐up time Outcomes measures (scales)
Austin et al. (2018) USA AFFIRM CBT 8 14–18 8 Diversea NR 3 months BDI‐II, PCI‐A
Berry and Hunt (2009) Australia Confident kids program VS wait list CBT 44 13.04 8 100% 60 min 3 months

SCARED, CES‐DC, SPPC

SPPA
Chillemi (2015) Australia IRCB program CBT 54 14.7 1 86% 60 min 3 months GHSQ, ATSPPHS
DeRosier (2004) USA Social skills group intervention VS no‐treatment control group CBT 415 8.6 8 49.20% 50 to 60 min NR SASC‐R, SPPC
DeRosier and Marcus (2005) USA Social skills group intervention VS no‐treatment control group CBT 381 8.6 8 49.20% 50 to 60 min 1 year SASC‐R, SPPC
Dugué et al. (2019) Germany Imagery rescripting VS cognitive restructuring Imagery 36 42.56 2 25% 41 min 1 weeks BDI‐V, STAI‐T
Fite et al. (2019) USA ACTION VS No‐treatment control group CBT 24 9.01 24 62.50% 45–60 min NR SRCS, SMFQ, PROMIS
Fox and Boulton (2003) UK Social skill training programme CBT 28 9.64 8 25% NR 3 months RCMAI, CDI, GSW
Fung (2012) China Cognitive‐behavioural group therapy CBT 68 13.4 10 71% NR 6 months, 1 years STAXI
Fung (2018) China Cognitive‐behavioural group Therapy CBT 68 13.2 10 65% 90 min 1 years CBCL‐YSR
Healy and Sanders (2014) Australia Resilience triple P CBT 111 8.72 4 61% NR 9 month PFC
Hunt et al. (2022) Australia Cool kids CBT 261 10.9 10 57.10% 120 min NR SCAS
Lydecker et al. (2024) USA TF‐CBT for weight‐related bullying Imagery 28 13.71 12 32% 60 min NR RSES, SCS, PHQ‐9
McCarthy et al. (2022) Australia Imagery rescripting Imagery 33 30.34 12 53.10% NR NR BFNE‐S
Rajabi et al. (2017) Iran Cognitive‐behavioural group therapy VS no‐treatment control group CBT 30 11–18 12 100% 90 min NR YSR, Billings and Mouse's Coping Strategies Scale
Stewart et al. (2023) Australia ‘Bullying: The Power to Cope’ program VS no‐treatment control group CBT 115 11.45 5 50% 55 min 2 weeks STAI‐CH
Thorisdottir et al. (2022) Canada Internet‐delivered CPT VS therapist‐guided stress management VS waitlist group CBT 52 43.37 12 29% 30–50 min 1 months

PTCI, DASS‐21, ASI‐3,

DTS, BRS, CERQ‐short,
Twardawski et al. (2024) Germany Imagery intervention VS Imagery rehearsal Imagery 356 24.7 1 33% 6 min NR PANAS, Empowerment scale
Vliek et al. (2019) The Netherlands Topper training VS waiting list CBT 132 9.38 10 50% 90 min 6 months CDI, SPPC
Waechter et al. (2017) Canada Cognitive‐behavioural therapy with imagery exposure Imagery 91 31.7 12 37% 120 min NR SPIN
Watson et al. (2016) Australia Imagery rescripting Imagery 135 18.39 1 19% 11.25 min NR

DASS 21, PANAS

Evaluation response items
Watson et al. (2021) Australia Imagery rescripting Imagery 43 12.81 2 100% NR NR

PANAS‐C, DASS 21

Evaluation response items
Yamada et al. (2018) Japan Imagery Rescripting Imagery 19 38.4 2 42% 50 min NR BDI‐II
Open in a new tab

Note: NR: not reported; a Austin et al. (2018): Sample consists of gender‐diverse youth, including non‐binary (32%), queer (26%), transgender (11%) and other minoritized gender identities; only 5% identified as cisgender male. Internalizing Symptoms: Includes measures of depression (BDI‐II, CDI, CES‐DC, PHQ‐9, SMFQ and PFC), anxiety (SCARED, STAI‐T, STAI‐CH, SCAS, RCMAI, SPIN, SASC‐R, ASI‐3 and BFNE‐S) and general psychological distress or trauma‐related symptoms (DASS‐21, YSR, CBCL‐YSR, PTCI, PANAS, PANAS‐C, STAXI and DTS). Self‐perception and Coping: Includes measures of self‐concept/worth (SPPC, SPPA, RSES, GSW and SCS), coping strategies (PCI‐A, SRCS and CERQ‐short) and resilience/functional health (BRS, PROMIS, GHSQ and ATSPPHS). Scales: ASI‐3: Anxiety Sensitivity Index–3; ATSPPHS: Attitudes Towards Seeking Professional Psychological Help Scale—Short Form; BDI‐II: Beck Depression Inventory‐II; BFNE‐S: Brief Fear of Negative Evaluation‐Straightforwardly Worded; BRS: Brief Resilience Scale; CBCL‐YSR: Child Behaviour Checklist‐Youth Self‐Report; CDI: Children's Depression Inventory; CERQ‐short: Cognitive Emotion Regulation Questionnaire‐short; CES‐DC: Centre for Epidemiologic Studies Depression Scale for Children; DASS‐21: Depression Anxiety Stress Scales‐21; DTS: Distress Tolerance Scale; GHSQ: General Help‐Seeking Questionnaire; GSW: Global Self‐Worth; PANAS/‐C: Positive and Negative Affect Schedule (for Children); PCI‐A: Adolescent Proactive Coping Inventory; PFC: Preschool Feelings Checklist; PHQ‐9: Patient Health Questionnaire‐9; PROMIS: Patient‐Reported Outcomes Measurement Information System; PTCI: Post‐traumatic Cognitions Inventory; RCMAI: Revised Children's Manifest Anxiety Inventory; RSES: Rosenberg Self‐Esteem Scale; SASC‐R: Social Anxiety Scale for Children–Revised; SCAS: Spence Children's Anxiety Scale; SCARED: Screen for Child Anxiety Related Emotional Disorders; SCS: Self‐Compassion Scale; SMFQ: Mood and Feelings Questionnaire; SPIN: Social Phobia Inventory; SPPC/‐A: Self‐Perception Profile for Children/Adolescents; SRCS: Self‐Report Coping Scale; STAI‐T/‐CH: State–Trait Anxiety Inventory (Trait version/for Children); STAXI: State–Trait Anger Expression Inventory; YSR: Youth Self‐Report.

3.2. Risk of Bias

Overall, most studies showed low RoB across all aspects, particularly in reporting, internal validity bias and power. However, some concerns were more common in the domains of external validity and internal validity (confounding factors). Specifically, about 30.43% studies were rated as high risk for external validity, and 73.91% were rated as having some concerns regarding internal validity due to potential confounding factors. Regarding statistical power, three studies were rated as high risk for low statistical power, whereas most were rated as low risk. Figure 2 shows the summary plot and Figure S1 visualizes individual study RoB using a traffic light plot.

FIGURE 2.

FIGURE 2

Open in a new tab

Risk of bias summary plot.

3.3. Internalizing Symptoms Changes Between Pre‐ and Post‐CBT and Imagery Interventions

Across all included studies (k = 21), random‐effect models revealed a significant reduction in internalizing symptoms from pre‐ to post ‐ intervention assessments, see Figure 3 (g = 0.72, 95% CI [0.38, 1.06], p < 0.001). The studies exhibited substantial heterogeneity (τ 2  = 0.50, I 2 = 93.6%, 95%CI [91.4%, 95.2%], Q (20)  = 310.38, p < 0.001). Inspection of the funnel plot (see Figure S2) suggested some asymmetry, and Egger's test provided evidence of potential publication bias (t(19) = 2.70, p = 0.01). Trim‐and‐fill analysis imputed eight missing studies on the lower end of the funnel, which reduced the overall effect estimate from g = 0.72 to g = 0.26 (95% CI [−0.15, 0.66], p = 0.210). The Baujat plot highlighted Rajabi et al. (2017) as the study contributing most to the overall heterogeneity (see Figure S3).

FIGURE 3.

FIGURE 3

Open in a new tab

Forest plot on internalizing symptoms changes between pre‐ and pos ‐ tinterventions.

Calculation of NNT on the original effect sizes suggested that approximately two participants would need to receive imagery interventions (NNT = 2.11) and around three participants should receive CBT (NNT = 2.95) to achieve one participant showing clinically meaningful improvement. Taking into account publication bias, the NNTs increase to 3.39 for imagery interventions and 14.20 for CBT respectively.

When comparing intervention types using subgroup analyses, no significant differences were found between CBT and imagery interventions (p = 0.380). Pooled effect sizes indicated g = 0.89 (95% CI [0.40, 1.39]) for imagery interventions and g = 0.62 (95% CI [0.12, 1.12]) for CBT, with high heterogeneity in both groups (imagery: τ 2  = 0.29, I 2 = 92.70%; CBT: τ 2  = 0.63, I 2 = 92.70%). Subgroup analyses by experimental design revealed no significant differences between controlled (k = 8, g = 0.74) and uncontrolled (k = 5, g = 0.48) studies with CBT subgroup (p = 0.511). Similarly, no significant differences were observed between controlled (k = 2, g = 0.83) and uncontrolled (k = 6, g = 0.92) designs for imagery interventions (p = 0.876). The meta‐regression results showed that number of sessions, age and gender did not moderate the effect for either imagery interventions or for CBT (all p's > 0.05). Figure S4 shows the bubble plots for these analyses.

Leave‐one‐out sensitivity analysis indicated that no single study substantially altered the overall heterogeneity or the pooled effect size. The I 2 remained high (91%–94%) and the pooled effect remained highly significant (p < 0.001). Rajabi et al. (2017) had the largest impact on τ 2, reducing it from 0.49 to 0.24, suggesting it contributes somewhat more to between‐study heterogeneity than other studies. Overall, the results of the meta‐analysis were robust to the exclusion of any individual study.

3.4. Internalizing Symptoms Change From Pre to Follow‐Up CBT

The random‐effects model indicated a significant reduction in internalizing symptoms from pre‐ to follow‐up assessments in CBT studies, see Figure 4 (k = 9, g = 0.54, 95% CI [0.29, 0.79], p = 0.001). The studies showed moderate to substantial heterogeneity (τ 2  = 0.07, I 2 = 65.6%, 95% CI [30.1%, 83.1%], Q (8) = 23.27, p = 0.003). Visual inspection of the funnel plot (see Figure S5) did not reveal asymmetry. Egger's test did not find a significant effect (t(7) = 1.73, p = 0.13). Trim‐and‐fill analysis imputed three potentially missing studies, resulting in a slightly reduced overall effect estimate, changing g = 0.54 to g = 0.37 (95% CI [0.08, 0.66], p = 0.02). The Baujat plot (Figure S6) identified one study contributing to heterogeneity (Healy and Sanders 2014).

FIGURE 4.

FIGURE 4

Open in a new tab

Forest plot on internalizing symptoms changes between pre‐ and follow‐up CBT.

Calculation of NNT suggested that approximately three participants would need to receive the CBT intervention for one to achieve a meaningful reduction in internalizing symptoms (NNT = 3.39). After considering publication bias, the NNT increased to 4.85. Subgroup analyses by experimental design revealed no significant differences between controlled (k = 5, g = 0.66) and uncontrolled (k = 4, g = 0.39) studies with CBT subgroup (p = 0.160). Comparison of potential moderators using meta‐regression did not reveal significant effects for number of sessions, age and gender (all p's > 0.05). Figure S7 shows the bubble plots for these analyses. A leave‐one‐out sensitivity analysis indicated that no single study has a strong impact on the overall effect size. Heterogeneity I 2 ranged between 25% and 70%, indicating moderate to substantial heterogeneity. Removing any individual study did not substantially change the significance of the overall effect (all p's < 0.01).

3.5. Self‐Perception and Coping Results Between Pre‐ and Post‐CBT and Imagery Interventions

Across all interventions (k = 15), the random‐effects model indicated a significant improvement in self‐perception and coping ability from pre‐ to post ‐ interventions in Figure 5 (g = 0.75, 95% CI [0.22, 1.27], p = 0.009). There was high heterogeneity (τ 2 = 0.87, I 2 = 99.0%, 95% CI [98.8%, 99.2%], Q (14) = 1438.61, p < 0.001). Visual inspection of the funnel plot (Figure S8) and Egger's test (t(13) = −0.50, p = 0.624) did not indicate significant publication bias. The trim‐and‐fill procedure suggested the presence of three potentially missing studies, which changed the overall effect from g = 0.75 to g = 0.98 (95% CI [0.47, 1.48], p < 0.001). The Baujat plot (Figure S9) identified Rajabi et al. (2017) as the study contributing most to overall heterogeneity.

FIGURE 5.

FIGURE 5

Open in a new tab

Forest plot on self‐perception and coping changes between and post ‐ interventions.

NNT was calculated based on the overall effect sizes for each subgroup. For CBT, roughly three participants would need to receive the intervention for one to experience a positive outcome (NNT = 2.63). For imagery interventions, the number of participants needed to be treated was around two (NNT = 2.09). After considering publication bias, the NNT increased to 6.25 for CBT and the NNT for imagery interventions decreased to 1.46.

The comparison between CBT and imagery interventions revealed no significant difference between intervention types on self‐perception and coping (Q (1)  = 0.26, p = 0.611). The pooled effect size for CBT was g = 0.64 (95% CI [−0.10; 1.38]), whereas imagery interventions had an effect size of g = 0.90 (95% CI [−0.15, 1.96]). We found high heterogeneity in both groups (CBT: τ 2  = 0.88, I 2  = 94.7%; imagery: τ 2  = 0.97, I 2  = 99.5%). Subgroup analyses by experimental design revealed no significant differences between controlled (k = 6, g = 0.75) and uncontrolled (k = 3, g = 0.43) studies with CBT subgroup (p = 0.533). Similarly, no significant differences were observed between controlled (k = 2, g = 1.03) and uncontrolled (k = 4, g = 0.83) designs for imagery interventions (p = 0.876).

The meta‐regression results showed that potential variables including number of sessions, age and gender did not moderate the effect (all p's > 0.05). Figure S10 shows the bubble plots. Leave‐one‐out sensitivity analysis showed that removing individual studies did not substantially affect the overall heterogeneity or the pooled effect size.

3.5.1. Sensitivity Analyses

To investigate the robustness of results, sensitivity analyses were conducted to evaluate the impact of the assumed pre–post correlation used in effect size calculations. For within‐group designs where the pre–post or pre ‐ follow‐up correlation was missing (pre – post internalizing symptoms: n = 1; pre – post self‐perception and coping: n = 2; pre ‐ follow‐up internalizing symptoms: n = 4), an assumed r = 0.50 was used. Varying the assumed pre – post correlation (r = 0.30 and r = 0.70) did not meaningfully alter the pooled effect sizes or their statistical significance (see Table S2).

In addition, sensitivity analyses were performed to examine the influence of the assumed correlation (r = 0.50) among multiple outcomes when calculating study‐level composite effect sizes. For within‐group designs where the pre – post or prefollow‐up correlation was missing (pre–post internalizing symptoms: n = 4; pre–post self‐perception and coping: n = 12; prefollow‐up). Using alternative correlations of r = 0.30 and r = 0.70, instead of the primary assumption, yielded virtually identical pooled effect sizes, confidence intervals and levels of statistical significance across all primary outcomes (see Table S3).

3.5.2. Certainty of Evidence Assessment

Finally, the overall evidence was examined using GRADE, which is summarized in Table 2. As most of included studies were uncontrolled pre – post designs, the initial starting level of certainty for all three outcomes was anchored at Low. Upgrading was not applicable for any outcomes due to the presence of methodological limitations. For pre – post Internalizing symptoms, the evidence was downgraded to Very Low, due to the observed risk of bias, inconsistency of results (I 2  = 93.6%) and indication of publication bias (Egger's test p = 0.01). Similarly for pre ‐ follow‐up internalizing symptoms, the final rating was also Very Low. The evidence was downgraded for risk of bias and inconsistency (I 2  = 65.6%). Publication bias was not detected (p = 0.13). Finally, for pre ‐ post self‐perception and coping ability, the evidence was downgraded to Very Low, receiving a decrease due to risk of bias and inconsistency of results (I 2  = 99.0%). No publication bias was detected.

TABLE 2.

GRADE certainty of evidence table.

Outcomes N of studies Risk of bias Inconsistency Indirectness Imprecision Other considerations Certainty
Pre–post internalizing symptomsa 21 Serious Very serious Not serious Not serious

Publication bias strongly suspected

All plausible residual confounding would reduce the demonstrated effect

⨁◯◯◯

Very low
Prefollow‐up Internalizing symptomsa 9 Serious Serious Not serious Not serious All plausible residual confounding would reduce the demonstrated effect

⨁◯◯◯

Very low
Pre–post self‐perception and coping abilityb 15 Serious Very serious Not serious Not serious All plausible residual confounding would reduce the demonstrated effect

⨁◯◯◯

Very low
Open in a new tab

Note: a Internalizing symptoms: Aggregated from scales assessing depression (e.g., BDI‐II, CDI, CES‐DC, PHQ‐9, SMFQ and PFC), anxiety (e.g., SCARED, STAI, SCAS, RCMAI, SPIN, SASC‐R, ASI‐3 and BFNE‐S) and general distress or trauma‐related symptoms (e.g., DASS‐21, YSR, CBCL‐YSR, PTCI, PANAS, PANAS‐C, STAXI and DTS); b Self‐perception and coping ability: Aggregated from scales assessing self‐concept/worth (e.g., SPPC, SPPA, RSES, GSW and SCS), coping strategies (e.g., PCI‐A, SRCS and CERQ‐short) and resilience/functional health (e.g., BRS, PROMIS, GHSQ and ATSPPHS).

4. Discussion

This systematic review and meta‐analysis set out to estimate the overall effects of CBT and imagery interventions on internalizing symptoms and related outcomes in bullying victims. The results suggested that both types of interventions were associated with significant improvements in two types of outcomes: internalizing symptoms and self‐perception and coping abilities. However, these initial findings should be interpreted with caution given the substantial publication bias and large, unexplained heterogeneity observed across studies. We will discuss each of these issues below.

CBT and imagery interventions may provide potential effective treatments for bullying victims. Improvements were observed across multiple studies, suggesting that both approaches may target psychological processes associated with recovery from bullying victimization. Conceptually, their common mechanism lies in their ability to promote cognitive and emotional changes. CBT encourages participants to adopt a more rational and balanced perspective by helping them identify and question the validity and utility of their interpretations (Hales et al. 2015). Imagery interventions (especially ImRs) take a different approach, changing the meaning of events (Moritz et al. 2018). Imagery rescripting allows clients to reexperience disturbing past events from a new perspective, inconsistent with the original threatening valence, generating a new representation that carries a less threatening valence and less emotional distress (Saulsman et al. 2019).

Both approaches ultimately aim to promote new learning and therapeutic change through different therapeutic processes. Some studies suggested that improvements associated with CBT interventions may be maintained following the end of treatment (Berry and Hunt 2009; DeRosier and Marcus 2005; Dugué et al. 2019; Fung 2018; Healy and Sanders 2017; Thorisdottir and Asmundson 2022; Vliek et al. 2019). Currently, no imagery intervention studies examined long‐term effects. The pooled effect sizes in this meta‐analysis suggested potential beneficial effects. However, publication bias strongly influenced both the overall effect sizes and their NNTs. After trim‐and‐fill, the estimates of NNT and their directions changed in different intervention types. Thus, the underlying efficacy of interventions may be more modest than the uncorrected results initially suggested. Given the inclusion of uncontrolled pre – post designs, these NNT values should be interpreted as an estimate of the number of individuals treated to observe one case of clinically meaningful improvement rather than a definitive causal effect. Especially, the decrease in imagery intervention NNT estimate should be interpreted particularly cautiously. Because it might reflect instability due to limited included studies rather than a stronger intervention effect. For instance, Rajabi et al. (2017) reported a large effect size, which could have led to an overestimation of the pooled effect size. The sensitivity analyses indicated that removing this study from the internalizing symptoms model decreased interstudy heterogeneity and removing this study from the self‐perception and coping strategy model resulted in a decrease in τ 2. However, in both cases, the overall pooled effect remained statistically significant.

This indicates that the high heterogeneity observed in this review limits the interpretability of a single pooled effect size. Sensitivity analyses explored potential sources of heterogeneity but could not explain all variability, suggesting that the efficacy of CBT and imagery interventions may be sensitive to specific contextual or procedural factors that are not yet consistently reported in primary studies.

In summary, CBT and imagery interventions appear to represent complementary therapeutic approaches for bullying victims. The imagery approach may elicit stronger and more immediate emotional engagement, whereas CBT provides broader cognitive restructuring skills that sustain long‐term gains. However, despite these encouraging findings, several methodological constraints warrant caution in interpreting the results. The magnitude of observed effects must be viewed in light of the substantial publication bias and unexplained heterogeneity revealed in our analyses.

5. Limitations and Future Directions

While this meta‐analysis provides valuable insights, several limitations must be acknowledged. First, the initial pooled analysis was influenced by publication bias, and the main effect on internalizing symptoms was reduced after adjustment, suggesting the true effect may be overestimated. To overcome this, future clinical trials must prioritize study preregistration to ensure all outcomes are available for synthesis.

Second, the analysis is affected by high unexplained statistical heterogeneity; the included studies differed substantially in methodology and outcomes across the included studies. Although we tried to identify the sources of this difference through subgroup analysis or meta‐regression analysis, these were mostly unsuccessful. Consequently, a consensus on standardized intervention protocols and the adoption of core outcome measures are needed to facilitate more meaningful cross‐study comparisons and provide more precise effect size estimates.

Third, we did not compare the difference between passive and provocative victims of bullying. We acknowledge the clinical importance of these distinct profiles, but the number of primary studies reporting disaggregated data for provocative victims was insufficient to run subgroup analysis. Consequently, victims were pooled to ensure adequate statistical power for the overall synthesis. Future research should rigorously measure and report implementation fidelity and provide detailed data on victim characteristics to determine whether the effectiveness of interventions varies across different types of peer bullying.

Fourth, this meta‐analysis treated CBT and imagery interventions as separate categories to organize the included studies. However, in clinical practice, these techniques are often integrated. To move the field forward, more direct‐comparison trials are required, particularly comparing standalone imagery interventions against standard CBT, and CBT that integrates imagery techniques. Such studies should also include long‐term follow‐up assessments to determine if the observed effects are maintained over time, addressing the current lack of longitudinal efficacy data.

Finally, a major limitation of this study is the heavy reliance on uncontrolled pre – post studies. Without control conditions, the observed symptom improvement may be influenced by a lot of factors unrelated to the intervention. Therefore, both the pooled effect sizes and the derived NNT estimates should be interpreted cautiously. Future research utilizing randomized controlled designs with consistent control groups will be essential to refine these clinical utility metrics.

6. Conclusion

In summary, this meta‐analysis suggests that both CBT and imagery may offer potential in addressing internalizing symptoms caused by bullying victimization. Both approaches help individuals re‐evaluate adverse experiences and develop a more adaptive sense of self. These findings highlight the potential value of integrative treatment strategies that combine cognitive restructuring with emotion‐focused techniques to potentially enhance the immediacy and persistence of therapeutic effects. However, the current evidence is limited by significant publication bias, high heterogeneity and a lack of standardized intervention protocols, all of which necessitate careful interpretation of the findings. Future research should prioritize preregistered trials, consistent outcome measures and rigorous implementation assessments to clarify the true efficacy and sustainability of these interventions. By addressing these methodological shortcomings, future research can refine our understanding of how cognitive and imagery approaches may collectively contribute to the resilience and recovery of individuals who have experienced bullying.

Ethics Statement

This meta‐analysis was based on data extracted from previously published studies and did not involve direct interaction with human participants or the collection of personally identifiable information. Therefore, ethical approval is not needed.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Summary of software packages and version numbers.

Table S2: Sensitivity analyses of pooled effect sizes under different assumed pre–post/follow‐up correlations.

Table S3: Sensitivity analyses of composite effect sizes using different assumed correlations among outcomes.

Figure S1: Risk of bias traffic light plot.

Figure S2: S2A Funnel plot of internalized symptoms comparing pre‐ and post ‐ intervention. S2B Funnel plot of internalized symptoms after applying the trim‐and‐fill.

Figure S3: The Baujat plot of internalized symptoms between pre‐ and post ‐ intervention.

Figure S4: Meta‐regression bubble plots of the pooled effect size for internalized symptoms (pre‐ vs. post ‐ intervention): S4A: Association with gender ratio in CBT. S4B: Association with participants' age in CBT. S4C: Association with the number of sessions in CBT. S4D: Association with gender ratio in imagery intervention. S4E: Association with participants' age in imagery intervention. S4F: Association with the number of sessions in imagery intervention.

Figure S5:. S5A Funnel plot of internalized symptoms between pre‐ and follow‐up intervention. S5B Funnel plot of internalized symptoms between pre‐ and follow‐up intervention after trim‐and‐fill.

Figure S6: The Baujat plot of internalized symptoms between pre‐ and follow‐up intervention.

Figure S7: Meta‐regression bubble plots of the pooled effect size for internalized symptoms (pre‐ vs. follow‐up): S7A: Association with gender ratio in CBT. S7B: Association with participants' age in CBT. S7C: Association with the number of sessions in CBT.

Figure S8: S8A The Funnel plot of self‐perception and coping ability between pre‐ and postintervention. S8B The Funnel plot of self‐perception and coping ability between pre‐ and postintervention after trim‐and‐fill.

Figure S9: The Baujat plot self‐perception and coping ability between pre‐ and post ‐ intervention.

Figure S10: Meta‐regression bubble plots of the pooled effect size for self‐perception and coping ability (pre‐ vs. post ‐ intervention): S10A: Association with gender ratio in CBT. S10B: Association with age in CBT. S10C: Association with number of sessions in CBT. S10D: Association with gender ratio in imagery interventions. S10E: Association with age in imagery interventions. S10F: Association with number of sessions in imagery intervention.

CPP-33-e70330-s001.docx (553KB, docx)

Acknowledgements

M.L. was funded by the CSC scholarship (No. 202308340029). E.H.W.K. was supported by the Special Research Funds _Flanders (BOF.BAF.2024.0946.01). Y.V.Z. is funded by a concerted project (GOA) from the Special Research Fund (BOF) of Ghent University (reference number: BOF23‐GOA‐006). G.T. was supported by the European Commission action HORIZON‐TMA‐MSCA‐PF‐EF project MenNaRum (grant number: 101244202).

Liang, M. , Tamm G., De Laere O., Zwalmen Y. V., and Koster E. H. W.. 2026. “Imagery Interventions and Cognitive‐Behavioural Therapy for Psychological Complaints Related to Bullying: A Systematic Review and Meta‐Analysis.” Clinical Psychology & Psychotherapy 33, no. 5: e70330. 10.1002/cpp.70330.

Yannick Vander Zwalmen and Ernst H.W. Koster share last authorship.

Data Availability Statement

Data and analyses scripts for this meta‐analysis can be found on Open Science Framework (OSF): https://osf.io/wjzs6.

References

  1. Armitage, J. M. , Wang R. A. H., Davis O. S. P., and Haworth C. M. A.. 2022. “A Polygenic Approach to Understanding Resilience to Peer Victimisation.” Behavior Genetics 52, no. 1: 1–12. 10.1007/s10519-021-10085-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arntz, A. , and Weertman A.. 1999. “Treatment of Childhood Memories: Theory and Practice.” Behaviour Research and Therapy 37, no. 8: 715–740. 10.1016/S0005-7967(98)00173-9. [DOI] [PubMed] [Google Scholar]
  3. Arseneault, L. , Milne B. J., Taylor A., et al. 2008. “Being Bullied as an Environmentally Mediated Contributing Factor to Children's Internalizing Problems: A Study of Twins Discordant for Victimization.” Archives of Pediatrics & Adolescent Medicine 162, no. 2: 145–150. 10.1001/archpediatrics.2007.53. [DOI] [PubMed] [Google Scholar]
  4. Austin, A. , Craig S. L., and D'Souza S. A.. 2018. “An AFFIRMative Cognitive Behavioral Intervention for Transgender Youth: Preliminary Effectiveness.” Professional Psychology: Research and Practice 49, no. 1: 1–8. 10.1037/pro0000154. [DOI] [Google Scholar]
  5. Baujat, B. , Mahé C., Pignon J., and Hill C.. 2002. “A Graphical Method for Exploring Heterogeneity in Meta‐Analyses: Application to a Meta‐Analysis of 65 Trials.” Statistics in Medicine 21, no. 18: 2641–2652. 10.1002/sim.1221. [DOI] [PubMed] [Google Scholar]
  6. Beck, J. S. 2020. Cognitive Behavior Therapy: Basics and Beyond. Guilford Publications. [Google Scholar]
  7. Berry, K. , and Hunt C. J.. 2009. “Evaluation of an Intervention Program for Anxious Adolescent Boys Who Are Bullied at School.” Journal of Adolescent Health 45, no. 4: 376–382. 10.1016/j.jadohealth.2009.04.023. [DOI] [PubMed] [Google Scholar]
  8. Borenstein, M. , ed. 2013. Introduction to Meta‐Analysis (Nachdr.). Wiley. [Google Scholar]
  9. Camargo, L. F. D. , Rice K., and Thorsteinsson E.. 2022. “A Systematic Review and Empirical Investigation: Bullying Victimisation and Anxiety Subtypes Among Adolescents.” Australian Journal of Psychology 74, no. 1: 2145236. 10.1080/00049530.2022.2145236. [DOI] [Google Scholar]
  10. Camargo, L. F. D. , Rice K., and Thorsteinsson E. B.. 2023. “Bullying Victimization CBT: A Proposed Psychological Intervention for Adolescent Bullying Victims.” Frontiers in Psychology 14: 1122843. 10.3389/fpsyg.2023.1122843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chillemi, K. 2015. Does a Self‐Help Online Psycho‐Educational Program on Cyber Bullying Lead to Increased Confidence and Help‐Seeking Behaviours Amongst Adolescents? [Doctoral Dissertation]. Swinburne University of Technology. [Google Scholar]
  12. Cohen, N. , Margulies D. S., Ashkenazi S., et al. 2016. “Using Executive Control Training to Suppress Amygdala Reactivity to Aversive Information.” NeuroImage 125: 1022–1031. 10.1016/j.neuroimage.2015.10.069. [DOI] [PubMed] [Google Scholar]
  13. Cross, D. , Epstein M., Hearn L., Slee P., Shaw T., and Monks H.. 2011. “National Safe Schools Framework: Policy and Practice to Reduce Bullying in Australian Schools.” International Journal of Behavioral Development 35, no. 5: 398–404. 10.1177/0165025411407456. [DOI] [Google Scholar]
  14. Cuijpers, P. 2016. Meta‐Analyses in Mental Health Research. A Practical Guide. Vrije Universiteit. [Google Scholar]
  15. DeRosier, M. E. 2004. “Building Relationships and Combating Bullying: Effectiveness of a School‐Based Social Skills Group Intervention.” Journal of Clinical Child & Adolescent Psychology 33, no. 1: 196–201. 10.1207/S15374424JCCP3301_18. [DOI] [PubMed] [Google Scholar]
  16. DeRosier, M. E. , and Marcus S. R.. 2005. “Building Friendships and Combating Bullying: Effectiveness of S.S.GRIN at One‐Year Follow‐Up.” Journal of Clinical Child & Adolescent Psychology 34, no. 1: 140–150. 10.1207/s15374424jccp3401_13. [DOI] [PubMed] [Google Scholar]
  17. Downs, S. H. , and Black N.. 1998. “The Feasibility of Creating a Checklist for the Assessment of the Methodological Quality Both of Randomised and Non‐Randomised Studies of Health Care Interventions.” Journal of Epidemiology and Community Health 52, no. 6: 377–384. 10.1136/jech.52.6.377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dugué, R. , Renner F., Austermann M., Tuschen‐Caffier B., and Jacob G. A.. 2019. “Imagery Rescripting in Individuals With Binge‐Eating Behavior: An Experimental Proof‐of‐Concept Study.” International Journal of Eating Disorders 52, no. 2: 183–188. 10.1002/eat.22995. [DOI] [PubMed] [Google Scholar]
  19. Duval, S. , and Tweedie R.. 2000. “Trim and Fill: A Simple Funnel‐Plot–Based Method of Testing and Adjusting for Publication Bias in Meta‐Analysis.” Biometrics 56, no. 2: 455–463. 10.1111/j.0006-341X.2000.00455.x. [DOI] [PubMed] [Google Scholar]
  20. Farrington, D. P. , Loeber R., Stallings R., and Ttofi M. M.. 2011. “Bullying Perpetration and Victimization as Predictors of Delinquency and Depression in the Pittsburgh Youth Study.” Journal of Aggression, Conflict and Peace Research 3, no. 2: 74–81. 10.1108/17596591111132882. [DOI] [Google Scholar]
  21. Fite, P. J. , Cooley J. L., Poquiz J., and Williford A.. 2019. “Pilot Evaluation of a Targeted Intervention for Peer‐Victimized Youth.” Journal of Clinical Psychology 75, no. 1: 46–65. 10.1002/jclp.22697. [DOI] [PubMed] [Google Scholar]
  22. Fox, C. , and Boulton M.. 2003. “Evaluating the Effectiveness of a Social Skills Training (SST) Programme for Victims of Bullying.” Educational Research 45, no. 3: 231–247. 10.1080/0013188032000137238. [DOI] [Google Scholar]
  23. Friedman, M. J. 2006. “Posttraumatic Stress Disorder Among Military Returnees From Afghanistan and Iraq.” American Journal of Psychiatry 163, no. 4: 586–593. 10.1176/ajp.2006.163.4.586. [DOI] [PubMed] [Google Scholar]
  24. Fung, A. L. 2012. “Intervention for Aggressive Victims of School Bullying in Hong Kong: A Longitudinal Mixed‐Methods Study.” Scandinavian Journal of Psychology 53, no. 4: 360–367. 10.1111/j.1467-9450.2012.00953.x. [DOI] [PubMed] [Google Scholar]
  25. Fung, A. L. 2018. “Cognitive‐Behavioural Group Therapy for Pure Victims With Internalizing Problems: An Evidence‐Based One‐Year Longitudinal Study.” Applied Research in Quality of Life 13, no. 3: 691–708. 10.1007/s11482-017-9553-4. [DOI] [Google Scholar]
  26. Ganslmeier, M. , Reineck A., Aleksic M., et al. 2025. “Effects of Imagery‐Based Interventions on Voluntary Memory.” A Systematic Review. 10.1017/S0033291722000836. [DOI] [Google Scholar]
  27. Geoffroy, M. C. , Arseneault L., Girard A., Ouellet‐Morin I., and Power C.. 2023. “Association of Childhood Bullying Victimisation With Suicide Deaths: Findings From a 50‐Year Nationwide Cohort Study.” Psychological Medicine 53, no. 9: 4152–4159. 10.1017/s0033291722000836. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Gesteira, C. , Garcia‐Vera M. P., Sanz J., and Shultz J. M.. 2026. “Trauma‐Focused Cognitive‐Behavioral Therapy for Long‐Term Posttraumatic Stress Disorder, Major Depressive Disorder and Anxiety Disorders in Victims of Terrorism: A Randomized Clinical Trial.” Psychotherapy Research 36, no. 2: 308–323. 10.1080/10503307.2025.2467380. [DOI] [PubMed] [Google Scholar]
  29. Guyatt, G. H. , Oxman A. D., Schünemann H. J., Tugwell P., and Knottnerus A.. 2011. “GRADE Guidelines: A New Series of Articles in the Journal of Clinical Epidemiology.” Journal of Clinical Epidemiology 64, no. 4: 380–382. 10.1016/j.jclinepi.2010.09.011. [DOI] [PubMed] [Google Scholar]
  30. Hackmann, A. , Bennett‐Levy J., and Holmes E. A.. 2011. Oxford Guide to Imagery in Cognitive Therapy. OUP Oxford. [Google Scholar]
  31. Hagenaars, M. A. , and Arntz A.. 2012. “Reduced Intrusion Development After Post‐Trauma Imagery Rescripting; an Experimental Study.” Journal of Behavior Therapy and Experimental Psychiatry 43, no. 2: 808–814. 10.1016/j.jbtep.2011.09.005. [DOI] [PubMed] [Google Scholar]
  32. Hales, S. , Blackwell S. E., Di Simplicio M., Iyadurai L., Young K., and Holmes E. A.. 2015. “Imagery‐Based Cognitive‐Behavioral Assessment.” In Assessment in Cognitive Therapy, edited by Brown G. P. and Clarke D. A., 69–93. Guilford Press. [PubMed] [Google Scholar]
  33. Harrer, M. , Cuijpers P., Furukawa T. A., and Ebert D. D.. 2022. Doing Meta‐Analysis With R: A Hands‐On Guide. First ed. CRC Press, Taylor & Francis Group. 10.1201/9781003107347. [DOI] [Google Scholar]
  34. Healy, K. , and Sanders M.. 2017. “Antecedents of Treatment Resistant Depression in Children Victimized by Peers.” Child Psychiatry & Human Development 48, no. 1: 107–119. 10.1007/s10578-016-0658-z. [DOI] [PubMed] [Google Scholar]
  35. Healy, K. , and Sanders M. R.. 2014. “Randomized Controlled Trial of a Family Intervention for Children Bullied by Peers.” Behavior Therapy 45, no. 6: 760–777. 10.1016/j.beth.2014.06.001. [DOI] [PubMed] [Google Scholar]
  36. Higgins, T. , J. P, Thomas J., et al. 2021. Cochrane Handbook for Systematic Reviews of Interventions (Version 6.3). Cochrane. https://training.cochrane.org/handbook. [Google Scholar]
  37. Holmes, E. A. , Mathews A., Dalgleish T., and Mackintosh B.. 2006. “Positive Interpretation Training: Effects of Mental Imagery Versus Verbal Training on Positive Mood.” Behavior Therapy 37, no. 3: 237–247. 10.1016/j.beth.2006.02.002. [DOI] [PubMed] [Google Scholar]
  38. Hunt, C. , Bussey K., Peters L., Gaston J., Lo A., and Rapee R. M.. 2022. “School‐Based Victimization in Children and Adolescents Presenting for Cognitive Behavioural Treatment of Anxiety Disorders.” Behavioural and Cognitive Psychotherapy 50, no. 6: 590–603. 10.1017/S1352465822000303. [DOI] [PubMed] [Google Scholar]
  39. Hutchings, J. , Pearson R., Babu M., et al. 2025. “Participants' Roles in Bullying Among 7–11 Year Olds: Results From a UK‐Wide Randomized Control Trial of the KiVa School‐Based Program.” Behavioral Sciences 15, no. 2: 236. 10.3390/bs15020236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Idsoe, T. , Dyregrov A., and Idsoe E. C.. 2012. “Bullying and PTSD Symptoms.” Journal of Abnormal Child Psychology 40, no. 6: 901–911. 10.1007/s10802-012-9620-0. [DOI] [PubMed] [Google Scholar]
  41. Idsoe, T. , Vaillancourt T., Dyregrov A., Hagen K. A., Ogden T., and Nærde A.. 2021. “Bullying Victimization and Trauma.” Frontiers in Psychiatry 11: 480353. 10.3389/fpsyt.2020.480353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ioannidis, J. P. A. , Patsopoulos N. A., and Evangelou E.. 2007. “Uncertainty in Heterogeneity Estimates in Meta‐Analyses.” BMJ 335, no. 7626: 914–916. 10.1136/bmj.39343.408449.80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Kennedy, R. S. 2020. “A Meta‐Analysis of the Outcomes of Bullying Prevention Programs on Subtypes of Traditional Bullying Victimization: Verbal, Relational, and Physical.” Aggression and Violent Behavior 55: 101485. 10.1016/j.avb.2020.101485. [DOI] [Google Scholar]
  44. Kovacs, M. , and Devlin B.. 1998. “Internalizing Disorders in Childhood.” Journal of Child Psychology and Psychiatry 39, no. 1: 47–63. 10.1017/S0021963097001765. [DOI] [PubMed] [Google Scholar]
  45. Kraemer, H. C. , and Kupfer D. J.. 2006. “Size of Treatment Effects and Their Importance to Clinical Research and Practice.” Biological Psychiatry 59, no. 11: 990–996. 10.1016/j.biopsych.2005.09.014. [DOI] [PubMed] [Google Scholar]
  46. Lydecker, J. A. , Ozbardakci E. V., Lou R., and Grilo C. M.. 2024. “Trauma‐Focused Cognitive‐Behavioral Therapy for Adolescents Bullied Because of Weight: A Feasibility Study.” International Journal of Eating Disorders 57, no. 10: 2117–2127. 10.1002/eat.24257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. McCarthy, A. , Bank S. R., Campbell B. N. C., Summers M., Burgess M., and McEvoy P.. 2022. “An Investigation of Cognitive and Affective Changes During Group Imagery Rescripting for Social Anxiety Disorder.” Behavior Therapy 53, no. 5: 1050–1061. 10.1016/j.beth.2022.04.012. [DOI] [PubMed] [Google Scholar]
  48. Moritz, S. , Ahlf‐Schumacher J., Hottenrott B., et al. 2018. “We Cannot Change the past, but We Can Change Its Meaning. A Randomized Controlled Trial on the Effects of Self‐Help Imagery Rescripting on Depression.” Behaviour Research and Therapy 104: 74–83. 10.1016/j.brat.2018.02.007. [DOI] [PubMed] [Google Scholar]
  49. Nielsen, M. B. , Tangen T., Idsoe T., Matthiesen S. B., and Magerøy N.. 2015. “Post‐Traumatic Stress Disorder as a Consequence of Bullying at Work and at School. A Literature Review and Meta‐Analysis.” Aggression and Violent Behavior 21: 17–24. 10.1016/j.avb.2015.01.001. [DOI] [Google Scholar]
  50. Norton, A. R. , Abbott M. J., Dobinson K. A., Pepper K. L., and Guastella A. J.. 2021. “Rescripting Social Trauma: A Pilot Study Investigating Imagery Rescripting as an Adjunct to Cognitive Behaviour Therapy for Social Anxiety Disorder.” Cognitive Therapy and Research 45: 1180–1192. [Google Scholar]
  51. Olweus, D. 1993. “Bullies on the Playground: The Role of Victimization.” In Children on Playgrounds: Research Perspectives and Applications, 85–128. SUNY Series, Children's Play in Society. [Google Scholar]
  52. Page, M. J. , McKenzie J. E., Bossuyt P. M., et al. 2021. “The PRISMA 2020 Statement: An Updated Guideline for Reporting Systematic Reviews.” BMJ: n71. 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Puhan, M. A. , Schunemann H. J., Murad M. H., et al. 2014. “A GRADE Working Group Approach for Rating the Quality of Treatment Effect Estimates From Network Meta‐Analysis.” BMJ 349, no. sep24 5: g5630. 10.1136/bmj.g5630. [DOI] [PubMed] [Google Scholar]
  54. Rajabi, M. , Bakhshani N.‐M., Saravani M. R., Khanjani S., and Javad Bagian M.. 2017. “Effectiveness of Cognitive‐Behavioral Group Therapy on Coping Strategies and in Reducing Anxiety, Depression, and Physical Complaints in Student Victims of Bullying.” International Journal of High Risk Behaviors and Addiction, (Inpress. 10.5812/ijhrba.41463. [DOI] [Google Scholar]
  55. Reijntjes, A. , Kamphuis J. H., Prinzie P., and Telch M. J.. 2010. “Peer Victimization and Internalizing Problems in Children: A Meta‐Analysis of Longitudinal Studies.” Child Abuse & Neglect 34, no. 4: 244–252. 10.1016/j.chiabu.2009.07.009. [DOI] [PubMed] [Google Scholar]
  56. Romano, M. , Moscovitch D. A., Huppert J. D., Reimer S. G., and Moscovitch M.. 2020. “The Effects of Imagery Rescripting on Memory Outcomes in Social Anxiety Disorder.” Journal of Anxiety Disorders 69: 102169. 10.1016/j.janxdis.2019.102169. [DOI] [PubMed] [Google Scholar]
  57. Rothon, C. , Edwards P., Bhui K., Viner R. M., Taylor S., and Stansfeld S. A.. 2010. “Physical Activity and Depressive Symptoms in Adolescents: A Prospective Study.” BMC Medicine 8: 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Saulsman, L. M. , Ji J. L., and Mcevoy P. M.. 2019. “The Essential Role of Mental Imagery in Cognitive Behaviour Therapy: What Is Old Is New Again.” Australian Psychologist 54, no. 4: 237–244. 10.1111/ap.12406. [DOI] [Google Scholar]
  59. Sourander, A. , Jensen P., Rönning J. A., et al. 2007. “What Is the Early Adulthood Outcome of Boys Who Bully or Are Bullied in Childhood? The Finnish “From a Boy to a Man” Study.” Pediatrics 120, no. 2: 397–404. 10.1542/peds.2006-2704. [DOI] [PubMed] [Google Scholar]
  60. Stewart, J. , Bernard M. E., and California State University, Long Beach . 2023. “Empowering the Victims of Bullying: The ‘Bullying: The Power to Cope’ Program.” Journal of Evidence‐Based Psychotherapies 23, no. 2: 147–172. 10.24193/jebp.2023.2.15. [DOI] [Google Scholar]
  61. Sun, Y. , Li T., Liu F., Liu X., and Zhen R.. 2025. “The Effectiveness of Virtual Reality (VR)‐Based Anti‐School Bullying Interventions: A Systematic Review and Meta‐Analysis.” Interactive Learning Environments: 1–21. 10.1080/10494820.2025.2553116. [DOI] [Google Scholar]
  62. Tehrani, N. 2004. “Bullying: A Source of Chronic Post Traumatic Stress?” British Journal of Guidance & Counselling 32, no. 3: 357–366. 10.1080/03069880410001727567. [DOI] [Google Scholar]
  63. Thorisdottir, A. S. , and Asmundson G.. 2022. “Internet‐Delivered Cognitive Processing Therapy for Individuals With a History of Bullying Victimization: A Randomized Controlled Trial.” Cognitive Behaviour Therapy 51, no. 2: 143–169. 10.1080/16506073.2021.1938663. [DOI] [PubMed] [Google Scholar]
  64. Ttofi, M. M. , Farrington D. P., Lösel F., and Loeber R.. 2011. “Do the Victims of School Bullies Tend to Become Depressed Later in Life? A Systematic Review and Meta‐Analysis of Longitudinal Studies.” Journal of Aggression, Conflict and Peace Research 3, no. 2: 63–73. 10.1108/17596591111132873. [DOI] [Google Scholar]
  65. Twardawski, M. , Gollwitzer M., Altenmüller M. S., et al. 2024. “Victim Empowerment and Satisfaction: The Potential of Imagery Rescripting.” European Journal of Social Psychology 54, no. 6: 1182–1197. 10.1002/ejsp.3073. [DOI] [Google Scholar]
  66. Vliek, L. , Overbeek G., and Orobio De Castro B.. 2019. “Effects of Topper Training on Psychosocial Problems, Self‐Esteem, and Peer Victimisation in Dutch Children: A Randomised Trial.” PLoS ONE 14, no. 11: e0225504. 10.1371/journal.pone.0225504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Waechter, S. , Rowa K., Milosevic I., Shnaider P., Antony M. M., and McCabe R. E.. 2017. “Social Anxiety and the Accuracy of Memory for Childhood Teasing Frequency.” Journal of Cognitive Psychotherapy 31, no. 3: 151–157. 10.1891/0889-8391.31.3.151. [DOI] [PubMed] [Google Scholar]
  68. Watson, H. , Rapee R., and Todorov N.. 2016. “Imagery Rescripting of Revenge, Avoidance, and Forgiveness for Past Bullying Experiences in Young Adults.” Cognitive Behaviour Therapy 45, no. 1: 73–89. 10.1080/16506073.2015.1108360. [DOI] [PubMed] [Google Scholar]
  69. Watson, H. , Todorov N., and Rapee R. M.. 2021. “Healing From Bullying in Early Adolescent Boys: The Positive Impact of Both Forgiveness and Revenge Fantasies.” Journal of School Violence 20, no. 2: 139–152. 10.1080/15388220.2020.1858425. [DOI] [Google Scholar]
  70. Yamada, F. , Hiramatsu Y., Murata T., et al. 2018. “Exploratory Study of Imagery Rescripting Without Focusing on Early Traumatic Memories for Major Depressive Disorder.” Psychology and Psychotherapy: Theory, Research and Practice 91, no. 3: 345–362. 10.1111/papt.12164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Zequinão, M. A. , Cardoso A. A., Silva J. L. D., et al. 2017. “Academic Performance and Bullying in Socially Vulnerable Students.” Journal of Human Growth and Development 27, no. 1: 19. 10.7322/jhgd.127645. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1: Summary of software packages and version numbers.

Table S2: Sensitivity analyses of pooled effect sizes under different assumed pre–post/follow‐up correlations.

Table S3: Sensitivity analyses of composite effect sizes using different assumed correlations among outcomes.

Figure S1: Risk of bias traffic light plot.

Figure S2: S2A Funnel plot of internalized symptoms comparing pre‐ and post ‐ intervention. S2B Funnel plot of internalized symptoms after applying the trim‐and‐fill.

Figure S3: The Baujat plot of internalized symptoms between pre‐ and post ‐ intervention.

Figure S4: Meta‐regression bubble plots of the pooled effect size for internalized symptoms (pre‐ vs. post ‐ intervention): S4A: Association with gender ratio in CBT. S4B: Association with participants' age in CBT. S4C: Association with the number of sessions in CBT. S4D: Association with gender ratio in imagery intervention. S4E: Association with participants' age in imagery intervention. S4F: Association with the number of sessions in imagery intervention.

Figure S5:. S5A Funnel plot of internalized symptoms between pre‐ and follow‐up intervention. S5B Funnel plot of internalized symptoms between pre‐ and follow‐up intervention after trim‐and‐fill.

Figure S6: The Baujat plot of internalized symptoms between pre‐ and follow‐up intervention.

Figure S7: Meta‐regression bubble plots of the pooled effect size for internalized symptoms (pre‐ vs. follow‐up): S7A: Association with gender ratio in CBT. S7B: Association with participants' age in CBT. S7C: Association with the number of sessions in CBT.

Figure S8: S8A The Funnel plot of self‐perception and coping ability between pre‐ and postintervention. S8B The Funnel plot of self‐perception and coping ability between pre‐ and postintervention after trim‐and‐fill.

Figure S9: The Baujat plot self‐perception and coping ability between pre‐ and post ‐ intervention.

Figure S10: Meta‐regression bubble plots of the pooled effect size for self‐perception and coping ability (pre‐ vs. post ‐ intervention): S10A: Association with gender ratio in CBT. S10B: Association with age in CBT. S10C: Association with number of sessions in CBT. S10D: Association with gender ratio in imagery interventions. S10E: Association with age in imagery interventions. S10F: Association with number of sessions in imagery intervention.

CPP-33-e70330-s001.docx (553KB, docx)

Data Availability Statement

Data and analyses scripts for this meta‐analysis can be found on Open Science Framework (OSF): https://osf.io/wjzs6.

Articles from Clinical Psychology & Psychotherapy are provided here courtesy of Wiley

RESOURCES