The impact of online situational interventions on hostile attribution bias and emotion regulation difficulties: an empirical study from the perspective of crime prevention

Abstract

Introduction

Adolescent aggression leading to crime is a growing global issue, and understanding its antecedents is crucial. Previous studies have shown that hostile attribution bias and difficulties in emotion regulation are significant psychological factors influencing aggression. However, traditional interventions are limited by time and space constraints. Online contextual interventions offer flexibility but have remained underutilized in this field. This study examines the predictive role of hostile attribution bias and difficulties in emotion regulation in adolescent aggression and explores the impact of online interventions on reducing these biases, assessing their potential for preventing adolescent crime.

Objective

The study aims to first verify the significant predictive role of hostile attribution bias and emotion regulation difficulties in adolescent aggression. Secondly, it seeks to explore whether online contextual interventions effectively reduce hostile attribution bias and improve emotion regulation ability, ultimately aiming to reduce aggression and prevent crime.

Methods

This research consists of two studies. Study 1 involved a survey of 559 adolescents from Wenzhou, China, using the Adolescent Aggression Questionnaire, Hostile Attribution Bias Scale, and Difficulties in Emotion Regulation Scale. Pearson correlation, regression analysis, and mediation tests were employed to analyze the data. Study 2 employed a randomized controlled design, dividing participants into experimental and control groups. The experimental group received online interventions targeting hostile attribution and emotion regulation ability, while the control group viewed unrelated documentaries. Post-intervention, both groups were reassessed, and independent sample t-tests were used to evaluate the effects.

Results

Study 1 showed that hostile attribution bias (r =.577, p <.001) and emotion regulation difficulties (r =.630, p <.001) significantly predicted adolescent aggression, explaining 50.6% of the behavioral variance (R² = 0.506). Study 2 demonstrated a significant reduction in hostile attribution bias (M = 55.15, SD = 18.35, t = 3.05, p =.010) and emotion regulation difficulties (M = 33.46, SD = 14.02, t = 4.81, p <.001) in the experimental group post-intervention, while the control group showed no significant changes.

Conclusions

(1) Hostile attribution bias and emotion regulation difficulties significantly predict adolescent aggression; (2) Online contextual interventions effectively reduce hostile attribution bias and improve emotion regulation abilities in adolescents; (3) The study provides theoretical and empirical support for using online interventions in behavioral correction and crime prevention in adolescents; (4) Online interventions offer a convenient, efficient approach to preventing aggression and crime among adolescents.

Clinical trial number

This study was retrospectively registered in the Chinese Clinical Trial Registry (ChiCTR2500100303, registration date: April 7, 2025).

Introduction

The rise in juvenile delinquency, particularly aggressive behavior, poses significant challenges to individuals, families, and society at large. Aggressive behavior reflects not only an individual’s psychological state but also broader changes in social structures and cultural values, which are particularly relevant in China’s collectivist society, where individual behavior is often shaped by group norms [1, 2, 3]. This cultural context may amplify or mitigate aggression depending on social influences. Two critical factors influencing aggressive behavior are hostile attribution bias and emotion regulation abilities. Hostile attribution bias refers to the tendency of individuals to interpret ambiguous social situations as hostile [4], while emotion regulation refers to the ability to manage and modulate emotional responses [5]. Understanding the impact of these two factors on adolescent aggression is crucial for developing effective intervention strategies.

Given China’s unique socio-cultural context, including the influence of collectivist culture on individual behavior, exploring adolescent aggression and psychological interventions within this framework is particularly important. Online interventions, including computer-assisted cognitive-behavioral therapy (CBT) and cognitive bias modification, have been increasingly used in adolescent mental health to improve emotion regulation ability and reduce aggression [6, 5]. These methods enable adolescents to engage in safe, controlled environments where they can practice coping strategies and modify cognitive biases that contribute to aggressive behavior. Compared to traditional face-to-face interventions, online contextual interventions offer advantages of efficiency, convenience, and ease of dissemination [7, 8].

Thus, this study examines the impact of online contextual interventions within the aforementioned social and technological contexts, aiming to examine the impact of online contextual interventions on hostile attribution bias and emotion regulation in adolescents, with the ultimate goal of reducing aggressive behavior and potential criminal risk.

Hostile attribution bias, emotion regulation, and adolescent aggressive behavior

Aggression is a multidimensional construct that includes behavioral, cognitive, and emotional components. According to Buss and Perry, aggression includes three dimensions: aggressive behavior, aggressive cognition, and aggressive emotion [6].

There is a close relationship between adolescent aggressive behavior and both hostile attribution bias and emotion regulation abilities. Hostile attribution bias refers to the tendency of individuals to interpret ambiguous social situations as hostile. Nasby found that highly aggressive adolescents are more likely to perceive ambiguous social scenarios as hostile, a finding that has been supported by subsequent research [9, 4, 10]. Further research by Wilkowski et al. demonstrated that effectively recruiting cognitive control resources can mitigate the aggressive behaviors resulting from hostile attribution. This suggests that cognitive control plays a critical role in reducing the negative impact of hostile attribution on behavior [11].

Emotion regulation ability refers to the capacity to manage and control emotional responses, and deficits in this ability, known as emotion regulation difficulties, are significantly associated with aggressive behavior in adolescents. Insufficient emotion regulation abilities may lead to overreactions in response to provocation, thereby increasing the occurrence of aggressive behaviors [5]. Conversely, higher levels of emotion regulation can effectively reduce the occurrence of aggressive behavior [12, 13].

Moreover, there is a strong connection between hostile attribution bias and emotion regulation abilities. Hostile attribution bias directly influences adolescent aggressive behavior and indirectly exacerbates it by weakening emotion regulation abilities. This indicates that reducing hostile attribution bias and improving emotion regulation abilities can effectively intervene in adolescent aggressive behavior [4, 10]. Therefore, conducting a mediation analysis helps to understand the role of emotion regulation abilities in the relationship between hostile attribution bias and aggressive behavior, and provides a theoretical basis for interventions aimed at reducing aggressive behavior.

Cognitive-Emotion Regulation Theory emphasizes that emotional responses are not only triggered by external situations but also influenced by individuals’ cognitive evaluations of the situation and their difficulties in emotion regulation [5]. Adolescents with a high level of hostile attribution bias tend to interpret ambiguous social cues as hostile. This cognitive bias not only exacerbates emotional responses (such as anger) but also disrupts their emotion regulation process [10]. Research indicates that difficulties in emotion regulation amplify the negative effects of hostile attribution bias, thereby promoting aggressive behavior [14]. Therefore, from the perspective of cognitive-emotion regulation theory, difficulties in emotion regulation play a critical mediating role between hostile attribution bias and aggressive behavior. By focusing on the emotion regulation process, we can better understand how these cognitive and emotional interactions jointly promote aggressive behavior and provide a more refined theoretical framework for interventions.

Social-Cognitive Theory further emphasizes the role of self-regulation in managing emotional and behavioral responses [2]. Bandura argues that individuals with difficulties in emotion regulation, particularly those lacking cognitive control and impulse inhibition, are more likely to exhibit maladaptive behaviors, such as aggression [15]. In adolescents with a strong hostile attribution bias, difficulties in emotion regulation lead to impulsive aggressive reactions [16]. Therefore, the degree of emotion regulation difficulties becomes the key mechanism regulating the impact of hostile attribution bias on aggressive behavior [11]. By introducing emotion regulation difficulties as a mediating variable, this study reveals how hostile attribution bias exacerbates aggressive behavior through the mediation of emotion regulation difficulties. This theoretical framework not only helps to understand the role of emotion regulation difficulties between hostile attribution bias and aggressive behavior but also provides strong theoretical support for interventions aimed at reducing adolescent aggressive behavior.

Therefore, a deep understanding of the impact of hostile attribution bias and emotion regulation abilities on adolescent aggressive behavior is essential for developing effective intervention strategies. Future research should further explore how enhancing emotion regulation abilities and reducing hostile attribution bias can prevent aggressive behavior in adolescents, thereby lowering their risk of engaging in criminal activities. This line of research not only contributes to theoretical advancements but also provides a scientific basis for the development of practical intervention strategies [17].

The impact of online contextual intervention on hostile attribution Bias and emotion regulation

This study explores the roles of hostile attribution bias and emotion regulation in adolescent aggressive behavior, grounded in the General Aggression Model (GAM) and Social Information Processing Theory (SIP). The General Aggression Model integrates various theories to explain the complexity of aggressive behavior, emphasizing the interaction between internal cognitive, emotional, and physiological responses, and external situational factors [6, 10]. It posits that aggressive behavior results from these interactions, shaped by both long-term and short-term processes that influence the emergence of aggression based on individual experiences and environmental context [16, 15].

Conversely, Social Information Processing Theory focuses on how individuals interpret and respond to social information. Dodge suggests that individuals undergo a series of steps in processing social information: encoding information, interpreting it, clarifying goals, selecting responses, and implementing actions. This theory particularly emphasizes hostile attribution bias—the tendency to interpret ambiguous actions as hostile—which significantly increases the likelihood of aggressive responses [16, 19].

With advancements in information technology, online contextual interventions have become prominent in psychological research and applications. These interventions utilize virtual environments to simulate real-world social situations, providing interactive experiences and practice opportunities to help individuals enhance their cognitive and emotional response patterns [20, 8]. For instance, O’Toole & Dennis-Tiwary highlight the effectiveness of cognitive bias modification, a technique that trains individuals to reinterpret ambiguous social cues more positively, thereby reducing aggressive responses [20]. Integrating cognitive bias modification with computer-assisted CBT can enhance emotion regulation and reduce hostile attributions, offering a comprehensive framework for mitigating aggressive behavior in adolescents [20].

Online contextual interventions offer several advantages over traditional face-to-face methods, such as overcoming the limitations of time and space, enhancing accessibility, and allowing repeated practice in safe, controlled environments [8]. Research supports the effectiveness of these interventions in reducing hostile attribution bias and improving emotion regulation abilities. For example, situational storytelling training has been shown to positively affect cognitive recovery and emotion regulation, allowing adolescents to repeatedly practice coping strategies and refine their responses to ambiguous situations [20, 10].

The effectiveness of online contextual interventions has been validated across various cultural contexts. Studies indicate that such interventions can significantly reduce hostile attribution bias and improve emotion regulation, thus decreasing aggressive behavior [20, 11]. These findings underscore the broad potential of online contextual interventions in mental health and behavioral correction, providing robust theoretical and practical support for their application [21].

Building on previous research and theoretical models, this study designs experimental materials specifically tailored to the psychological and developmental characteristics of Chinese adolescents. By examining the effectiveness of this intervention technique, the study aims to confirm its utility in reducing hostile attribution bias and enhancing emotion regulation, thereby contributing new theoretical insights and practical strategies for juvenile crime prevention.

Research hypotheses

Based on the theoretical background and literature review discussed above, this study proposes the following hypotheses: (1) Hostile attribution bias and difficulties in emotion regulation will significantly and positively predict aggressive behavior in adolescents; (2) Online contextual interventions will significantly reduce hostile attribution bias and difficulties in emotion regulation in adolescents; (3) Online contextual interventions will significantly decrease aggressive behavior in adolescents. This study systematically explores how hostile attribution bias and emotion regulation influence adolescent aggression, as well as the mechanisms by which online contextual interventions operate, this study aims to provide scientific evidence and practical strategies for juvenile crime prevention.

Study 1: the relationship between hostile attribution bias, emotion regulation abilities, and aggressive behavior in adolescents

Research methodology

Participants

This study randomly selected one middle school, one high school, and one university in Wenzhou, Zhejiang Province, located in southern China, and distributed 200 questionnaires at each institution.Participants ranged in age from 12 to 22 years, and the inclusion criteria required no severe psychological or neurological disorders and no history of substance abuse. All participants voluntarily signed informed consent forms before completing the questionnaires. Data from 41 participants were excluded due to failure on validity checks, invalid responses, or failure to meet the questionnaire completion criteria. The final sample included data from 559 participants: 189 middle school students, 195 high school students, and 175 university students. The sample consisted of 190 males and 366 females, with an average age of 17.67 years (SD = 1.96 years). Demographic variables such as gender, age, place of household registration, family background, and education level were included in the analysis.

Measurement instruments

The Hostile Attribution Bias Scale used in this study is a localized adaptation of the Word Sentence Association Test for Obsessive-Compulsive Disorder (Riemann et al., 2013) [14]. This Chinese version of the scale consists of 32 items and utilizes a 6-point rating system (1 = completely unrelated, 6 = highly related), covering two dimensions: hostile attribution and benign attribution. The scale quickly assesses participants’ tendency to interpret ambiguous social situations by presenting vague sentences followed by words related to hostility or benevolence. For example, “Sentence: The door slams shut in front of you. Word: Insulting.” This study primarily focuses on the 16 items related to the hostile attribution dimension, with higher scores indicating more severe hostile attribution bias. The Chinese version of the scale has demonstrated good reliability and validity in both adolescent and adult samples, with a reliability coefficient of 0.82 in this study, indicating high consistency and reliability.

The Difficulties in Emotion Regulation Scale used in this study is based on the original scale developed by Gratz and Roemer (2004) [5] and was revised into Chinese by Wang Li et al. in 2007 [12]. The scale includes 36 items, rated on a 5-point Likert scale (1 = almost never, 5 = almost always), with some items requiring reverse scoring. Higher scores indicate greater difficulties in emotion regulation. The scale consists of six subscales: strategies for emotion regulation, goal-directed behavior under emotional arousal, emotional understanding, emotional awareness, impulse control when experiencing emotions, and acceptance of emotional responses. For example, items include “I understand my feelings” and “I pay attention to how I feel.” In this study, the overall test-retest reliability of the scale was 0.88, with an internal consistency reliability of 0.92. The reliability and validity of the subscales ranged from 0.79 to 0.88, indicating excellent psychometric properties for assessing difficulties in emotion regulation in adolescents.

The Adolescent Aggressive Behavior Scale used in this study is the Chinese version of the Buss-Perry Aggression Questionnaire, revised by Li Xianyun et al. in 2011 [6]. This scale consists of 30 items and measures five dimensions: physical aggression, verbal aggression, anger, hostility, and self-directed aggression. The scale uses a 5-point Likert rating system (1 = not applicable, 5 = completely applicable). Physical and verbal aggression are categorized as behavioral expressions of aggression, such as ‘In some situations, I can’t help but hit others.’ Anger is considered an emotional expression, while hostility represents a cognitive aspect. According to the scoring guidelines of this scale, higher total scores indicate higher levels of aggression. The internal consistency Cronbach’s alpha for the various dimensions ranges from 0.60 to 0.89, and the test-retest reliability intraclass correlation coefficient (ICC) ranges from 0.57 to 0.81, demonstrating good reliability and validity, indicating that the scale effectively measures levels of aggressive behavior in adolescents.

Research procedure

Before completing the questionnaires, all participants were informed of relevant instructions and precautions by the primary investigator. After ensuring that the participants understood the information, they voluntarily signed informed consent forms. Following this, the participants proceeded to complete the questionnaires.

Data analysis

Data were organized and analyzed using SPSS 17.0. First, descriptive statistical analyses were conducted on all variables. Next, Pearson correlation analysis was performed to explore the relationships between the variables. To examine the predictive roles of hostile attribution bias and emotion regulation difficulties on adolescent aggressive behavior, hierarchical regression analysis was conducted.

In hierarchical regression analysis, predictor variables were entered into the model in a stepwise manner to assess their incremental explanatory power. First, demographic variables (e.g., age, gender, and household registration) were entered as control variables. Second, the key psychological predictors, hostile attribution bias and emotion regulation difficulties, were added to determine whether they significantly predicted aggression beyond demographic controls.

Unlike simultaneous regression analysis, hierarchical regression allows for the evaluation of the unique contribution of each predictor variable by assessing the incremental changes in R². This approach ensures that the psychological factors’ predictive power is not confounded by demographic influences.

Finally, a mediation analysis was conducted to further investigate the indirect relationships among the variables.

Results

Correlation analysis of adolescent aggressive behavior, hostile attribution bias, and emotion regulation abilities

To explore the relationships between adolescent aggressive behavior, hostile attribution bias, and emotion regulation abilities, a Pearson correlation analysis was conducted on the data collected in Study 1. As shown in Table 1, there was a significant positive correlation between adolescent aggressive behavior and hostile attribution bias (r =.577, p <.001) [4, 10, 22, 23], indicating that as hostile attribution bias increases, the level of aggressive behavior in adolescents also increases. Additionally, there was a significant positive correlation between adolescent aggressive behavior and difficulties in emotion regulation(r =.630, p <.001), suggesting that greater difficulties in emotion regulation are associated with stronger aggressive behavior in adolescents. Furthermore, all sub-dimensions of emotion regulation abilities were significantly correlated with both aggressive behavior and hostile attribution bias, with specific correlation coefficients presented in Table 2. Finally, a significant positive correlation was also found between levels of hostile attribution bias and emotion regulation abilities (r =.507, p <.001), indicating that these psychological variables are interrelated and may collectively contribute to the manifestation of aggressive behavior in adolescents.

Multiple regression analysis of emotion regulation abilities, hostile attribution bias, and adolescent aggressive behavior

To examine the predictive roles of hostile attribution bias and emotion regulation difficulties on aggressive behavior, a hierarchical regression analysis was conducted. As shown in Table 2, predictor variables were entered into the model in a stepwise manner to assess their incremental explanatory power.

Step 1: Demographic variables (age, gender, and household registration) were entered into the model as control variables, yielding an R² of 0.333. This indicates that these demographic factors alone explain 33.3% of the variance in aggressive behavior.Step 2: Hostile attribution bias was added to the model, resulting in a significant increase in R² to 0.486 (ΔR² = 0.153, p <.001). This suggests that hostile attribution bias accounts for an additional 15.3% of the variance in aggression, beyond the influence of demographic variables.Step 3: Emotion regulation difficulties were then entered into the model, further increasing R² to 0.506 (ΔR² = 0.020, p =.003). This indicates that emotion regulation difficulties contribute an additional 2.0% of the variance in aggression, demonstrating an independent predictive effect.

Unlike simultaneous regression analysis, hierarchical regression allows for the assessment of the unique contribution of each predictor variable by evaluating incremental changes in explained variance (R²). The stepwise increase in R² suggests that both hostile attribution bias and emotion regulation difficulties significantly contribute to adolescent aggressive behavior beyond demographic influences.

The multiple regression analysis revealed that both emotion regulation abilities and hostile attribution bias are significant predictors of aggressive behavior. Specifically, the unstandardized coefficient for hostile attribution bias was B = 0.252, with a standard error of SE = 0.025, and the standardized coefficient was β = 0.350, t = 10.027, p <.001. For emotion regulation abilities, the unstandardized coefficient was B = 0.125, with a standard error of SE = 0.042, and the standardized coefficient was β = 0.106, t = 3.005, p =.003. Additionally, several sub-dimensions of hostile attribution bias and emotion regulation abilities significantly predicted aggressive behavior in adolescents. Among these, hostile attribution bias had the strongest predictive power for aggressive behavior, with a standardized coefficient of β = 0.350, indicating its most significant effect. Sub-dimensions of emotion regulation abilities, such as impulse control, regulation strategies, and goal-directed behavior, also significantly predicted aggressive behavior. Overall, these variables collectively accounted for a portion of the variance in aggressive behavior.

Mediation effect of emotion regulation abilities on the relationship between hostile attribution bias and adolescent aggressive behavior

This study utilized the SPSS PROCESS macro (Model 4) to examine the mediating role of emotion regulation abilities in the relationship between hostile attribution bias and adolescent aggressive behavior. As shown in Fig. 1; Tables 3 and 4, in the regression model for emotion regulation abilities, hostile attribution bias had a significant positive predictive effect on emotion regulation abilities, with a regression coefficient of 0.310 (p <.001). This model explained 25.73% of the variance (R² = 0.257). In the regression model for aggressive behavior, both hostile attribution bias and emotion regulation abilities had significant positive predictive effects on aggressive behavior, with regression coefficients of 0.249 (p <.001) and 0.532 (p <.001), respectively. This model explained 48.60% of the variance (R² = 0.486).

The mediation effect analysis showed that the indirect effect of hostile attribution bias on aggressive behavior through emotion regulation abilities was 0.165, with a Bootstrap confidence interval of [0.125, 0.207] (significant). The direct effect was 0.249, with a confidence interval of [0.200, 0.299], and this direct effect was statistically significant (p <.001).

The results of this study indicate that emotion regulation difficulties partially mediate the relationship between hostile attribution bias and adolescent aggressive behavior. Hostile attribution bias not only directly influences aggressive behavior, but also indirectly exacerbates it by enhancing emotion regulation difficulties [24]. This finding is consistent with previous research [16, 18, 11], further validating the mediating role of emotion regulation difficulties in the relationship between hostile attribution bias and aggressive behavior [10, 25]. Specifically, the study shows that enhancing emotion regulation abilities can effectively mitigate the negative impact of hostile attribution bias on aggressive behavior [18]. This finding provides theoretical support for interventions aimed at reducing adolescent aggressive behavior, highlighting the importance of enhancing emotion regulation abilities.

Table 1 Correlation analysis of variables

Table 2 Results of Multiple Regression Analysis

Table 3 Mediation Model of the Relationship Between Adolescent Aggressive Behavior, Hostile Attribution Bias, and difficulties in emotion regulation

Table 4 Mediation effects with bootstrapping

Fig. 1

Mediation Effect of Hostile Attribution Bias and Emotion Regulation Difficulties on Adolescent Aggressive Behavior

Discussion

This study conducted an in-depth analysis of the relationships between hostile attribution bias, emotion regulation abilities, and aggressive behavior in adolescents. The results indicated a significant positive correlation between hostile attribution bias and aggressive behavior, consistent with Nasby’s initial findings and subsequent studies [1, 4]. These studies confirm that when adolescents are inclined to interpret ambiguous social situations as hostile, their aggressive behavior tends to increase. The results of Study 1 further underscore the importance of hostile attribution bias in predicting aggressive behavior in adolescents.

The significant positive correlation between difficulties in emotion regulation and aggressive behavior supports relevant perspectives in emotion regulation theory [16, 9], suggesting that a lack of emotion regulation abilities may lead to overreactions to provocation, thereby increasing aggressive behavior. This finding is crucial for developing intervention measures for aggressive behavior, particularly in designing prevention programs aimed at enhancing emotion regulation abilities.

Moreover, the results of regression analysis and mediation effect testing revealed that emotion regulation abilities partially mediate the relationship between hostile attribution bias and aggressive behavior. Specifically, hostile attribution bias not only directly influences aggressive behavior but also indirectly increases it by impairing emotion regulation abilities. This finding provides theoretical support for interventions aimed at reducing adolescent aggressive behavior, highlighting the importance of improving emotion regulation abilities. The study by Wilkowski et al. further emphasizes the critical role of cognitive control in regulating aggressive impulses, particularly when faced with hostile triggers [26]. Recruiting cognitive control resources can significantly reduce the occurrence of aggressive behavior, a finding that aligns with the results of this study and supports the effectiveness of interventions that enhance emotion regulation abilities in reducing adolescent aggression.

Although this study used one-way analysis of variance (ANOVA) and independent samples t-tests to explore the influence of demographic variables such as age and place of household registration on hostile attribution bias and difficulties in emotion regulation, the results indicated that these variables were not significant predictors (F(3,554) = 2.304, p =.076; t = -0.628, p =.530). However, it is important to clarify that these variables were included as control variables in the regression analysis, rather than being used for correlation analysis. The decision to include age and place of household registration as control variables was based on previous research suggesting that these variables might be potentially related to the psychological factors under study. For example, Zimmermann and Iwanski (2014) noted that emotion regulation abilities improve with age, particularly during the transitional stages from adolescence to early adulthood [4]. Additionally, Crick and Dodge (1994) pointed out that hostile attribution bias is not only related to individual cognitive development but is also shaped by the broader social context [1].

Given that no significant results were found in our analysis, we recommend that future research consider larger sample sizes or different sociocultural settings to further explore the potential effects of these demographic variables on hostile attribution bias and emotion regulation abilities, in order to obtain more generalizable and robust findings. Moreover, when designing intervention strategies for adolescent aggressive behavior, consideration should be given to the developmental and cultural backgrounds of adolescents.

While this study achieved certain results, it also has some limitations. For example, the sample was limited to students in the Wenzhou area, which may affect the external validity of the findings. Future research should be conducted across a broader geographical and cultural context to verify the generalizability of the results. Additionally, this study relied on self-report questionnaires, which may introduce self-report bias. Future research could consider using multiple data collection methods, such as behavioral observations or assessments from teachers and parents, to enhance the reliability of the findings.

Based on the results of Study 1 and previous research, the research team introduced online contextual intervention technology in Study 2 to explore its impact on hostile attribution bias and emotion regulation abilities in adolescents. This will further investigate the effectiveness of this technology in intervening in adolescent aggressive behavior. By validating the effectiveness of online contextual interventions in Study 2, we hope to provide new theoretical support and practical references for juvenile crime prevention.

Study 2: the impact of online contextual interventions on hostile attribution bias and emotion regulation abilities in adolescents

Research methodology

Participants

This study recruited 180 university students (90 males and 90 females) from Wenzhou University who met the experimental criteria. The participants were aged between 18 and 23 years. All participants were required to meet the following inclusion criteria: right-handedness, no recent illness or medication use, no history of neurological or psychiatric disorders, and normal or corrected-to-normal vision. To ensure consistency in motor and cognitive responses, only right-handed participants were included in the study. Handedness has been shown to influence neural processing and response execution in cognitive and behavioral tasks [27]. Excluding left-handed participants helps control for potential variability in cognitive processing and ensures comparability of results. This criterion aligns with prior research that has adopted similar participant selection standards in behavioral and cognitive studies [28]. All participants voluntarily signed informed consent forms prior to the experiment and received appropriate compensation after the study. Each participant received a monetary compensation of 50 RMB for their time and effort. The amount of compensation was determined based on ethical considerations and standard research practices, ensuring that it was sufficient to acknowledge participants’ time without exerting undue influence [29, 30]. The 180 participants were then randomly assigned into two groups, Group A (45 males, 45 females) and Group B (45 males, 45 females), ensuring equal gender distribution.

The participants were further screened using the Difficulties in Emotion Regulation Scale and the Hostile Attribution Bias Scale to select those who met the specific criteria. Eight participants were excluded due to failing validity checks or incomplete responses. The remaining participants included 26 with high hostile attribution bias scores (top 30% on the Hostile Attribution Bias Scale) and 26 with high difficulties in emotion regulation scores (top 30% on the Difficulties in Emotion Regulation Scale), with an average age of 20.5 years (SD = 0.50 years). The 26 participants with high levels of hostile attribution bias were randomly assigned into two groups: Experimental Group 1 (13 participants) and Control Group 1 (13 participants). Similarly, the 26 participants with significant difficulties in emotion regulation were randomly assigned into two groups: Experimental Group 2 (13 participants) and Control Group 2 (13 participants).

Experimental design and materials

This experiment employed a single-factor between-groups design, primarily investigating the effects of online contextual interventions on hostile attribution bias and emotion regulation, as illustrated in Fig. 2. Participants were randomly assigned to different intervention themes, resulting in two experimental groups and two control groups: Experimental Group 1 (hostile attribution bias intervention) and Control Group 1, and Experimental Group 2 (emotion regulation intervention) and Control Group 2. Participants in Experimental Groups 1 and 2 underwent online contextual interventions targeting hostile attribution bias and difficulties in emotion regulation, respectively, under the guidance of the researchers. Meanwhile, participants in Control Groups 1 and 2 were organized to watch a documentary entirely unrelated to the content of the experiment. After the completion of the online contextual interventions, all participants were administered the Hostile Attribution Bias Scale and the Difficulties in Emotion Regulation Scale, and the relevant data were collected by the researchers.

The experimental materials include a life story situational intervention program developed by the researchers using the E-Prime software, as well as three scales previously used in Study 1. The researchers created six life story scenarios, consisting of three scenarios focused on hostile attribution bias intervention and three on difficulties in emotion regulation intervention. In developing these materials, the researchers referred to the story-based cognitive remediation training used by Khanna & Kendall [20] for psychiatric patients and the contextual settings of the Hostile Attribution Bias Scale. One scenario was randomly selected as the training material. These six scenarios were reviewed by three experts in applied psychology before being used in the intervention experiment.

An example of a specific scenario is as follows:“On a Saturday afternoon, Xiao Li was at home with her family. Her younger brother, Xiao Ming, was busy playing computer games, and due to weekend work commitments, their parents were unable to clean the house and expected Xiao Ming to tidy up his room. Despite multiple reminders from the parents, Xiao Ming locked himself in his room, insisting that he deserved more free time on weekends after a week of intense study. This led to disappointment and anger from the parents, who considered Xiao Ming to be selfish and disrespectful, creating tension in the household.” An example question based on this scenario is:

“Based on the scenario above, please complete the following task. 1. (Multiple choice) If you were Xiao Li, what actions would you take to ease the family tension? ( )

• A. Initiate a conversation with Xiao Ming to explain the reasons why the parents expect him to take responsibility at home.

• B. Communicate with the parents, expressing Xiao Ming’s reasonable need for relaxation and rest on the weekend.

• C. Ignore the family conflict and avoid getting involved.

• D. Use aggressive language to criticize both Xiao Ming and the parents, accusing them of ruining your mood.”

Participants were required to immerse themselves in the scenario and the roles of the characters, think critically, and make choices based on the situation.

In summary, the research framework for this study is established as shown in Fig. 2.

Experimental procedure

This experiment was divided into three phases: the preparation phase, the formal intervention phase, and the immediate post-test phase.

Preparation phase

Upon entering the experimental program, participants were instructed by the primary investigator to carefully read the operation manual and the relevant experimental requirements, and to fill in their personal information as instructed. After familiarizing themselves with the operational process, participants entered a practice mode as prompted by the program. In practice mode, the primary investigator presented a story scenario related to the theme. Participants were required to read and understand the story scenario thoroughly, then answer related single-choice and multiple-choice questions. If a question was left unanswered, participants could not proceed to the next one. Participants entered their answers into the answer box, and upon clicking “Submit,” the program automatically displayed the correct answers along with explanations. After completing the practice mode, participants could choose to proceed to the formal intervention experiment or continue practicing, based on their comfort level. There were no time or frequency limitations during the preparation phase.

Formal intervention phase

As shown in Fig. 3, during this phase, the program presented participants with two story scenarios related to the theme. Participants were required to read and comprehend each scenario carefully, with each scenario needing to be read for at least five minutes. If participants did not meet the minimum time requirement, the program would not display the questions and options. Participants could repeatedly read and review the scenarios and revise their answers to the questions by clicking on-screen buttons. Once participants had read both scenarios and answered all related questions, and had reviewed the correct answers and explanations for each question, they could exit the experimental program by clicking the designated button on the screen.

Immediate post-test phase

In this phase, participants from both groups were required by the primary investigator to complete the Hostile Attribution Bias Scale and the difficulties in emotion regulation Scale, and the data were collected immediately afterward.

Fig. 2

Flowchart of the between-group design in study 2

Fig. 3

Preview of the online contextual intervention experiment

Results

As shown in Table 5, the online contextual intervention had a significant impact on both hostile attribution bias and emotion regulation abilities in adolescents. For emotion regulation abilities, the mean scores for the experimental group before and after the intervention were M = 55.69 (SD = 17.57) and M = 33.46 (SD = 14.02), respectively, indicating that the intervention significantly reduced participants’ emotion regulation scores, t(12) = 4.81, p <.001. This suggests that the contextual intervention had a significant effect on improving the emotion regulation abilities of participants in the experimental group. In contrast, the control group’s mean emotion regulation scores before and after the intervention were M = 53.77 (SD = 11.23) and M = 50.85 (SD = 11.26), respectively, with no statistically significant change in scores, t(12) = 1.21, p =.250, indicating that emotion regulation abilities did not significantly change in the absence of intervention.

Regarding hostile attribution bias, the experimental group’s mean scores before and after the intervention were M = 68.46 (SD = 7.09) and M = 55.15 (SD = 18.35), respectively, with the intervention effect reaching statistical significance, t(12) = 3.05, p =.010, demonstrating that the intervention effectively reduced hostile attribution bias scores. The control group’s hostile attribution bias scores showed no significant change from M = 63.62 (SD = 4.56) to M = 62.15 (SD = 10.21), t(12) = 0.59, p =.565, further confirming that without intervention, hostile attribution bias scores did not significantly change.

The calculation of Cohen’s d further supported the effectiveness of the contextual intervention. For emotion regulation abilities, Cohen’s d was 1.33, indicating a large effect size. For hostile attribution bias, Cohen’s d was 0.85, also indicating a large effect.

Table 5 Comparison of pretest and posttest scores between the two groups

Discussion

Study 2 primarily aimed to evaluate the effects of online contextual interventions on hostile attribution bias and emotion regulation abilities in adolescents. These interventions were explored as psychological strategies aimed at mitigating aggression, a key risk factor for juvenile delinquency. While not all aggressive behaviors lead to criminal actions, reducing aggression can significantly decrease the likelihood of subsequent criminal behavior. Therefore, this study assessed the potential of these interventions in preventing aggressive behaviors and their associated antisocial outcomes. The results demonstrated that online contextual interventions significantly reduced hostile attribution bias and enhanced emotion regulation abilities in university students [26]. The selection of university students (18–23 years) as participants is based on several factors. First, this age group represents a transitional phase from adolescence to adulthood, a critical period for the development of emotion regulation and aggression-related behaviors [32]. This period is characterized by ongoing cognitive and emotional development, making university students an ideal group to study interventions targeting behavior outcomes like aggression and attribution biases. Second, university students typically experience significant emotional, social, and cognitive challenges during this time, which are comparable to challenges faced by adolescents but may involve more mature forms of emotional regulation. These students are still in a developmental stage where emotional and social difficulties are prominent, as noted by Moffitt (1993) [33], who found that even in young adulthood, individuals may face behavioral challenges similar to those seen in adolescence. However, given the study’s focus on a narrow age range, caution is needed in extending these findings to younger adolescent populations or individuals from different cultural contexts. This finding aligns with prior research highlighting the importance of cognitive control in mitigating aggression, particularly through interventions designed to enhance emotion regulation skills and modify cognitive biases [1, 4]. These interventions offer repeated practice and feedback within a virtual environment, enabling adolescents to manage their responses to ambiguous social cues more effectively, thereby reducing aggressive behaviors.

This study has significant theoretical and practical implications. Theoretically, it reinforces the critical role of difficulties in emotion regulation and cognitive restructuring in managing aggressive behavior, consistent with findings on the efficacy of computer-assisted cognitive-behavioral therapy (CBT) and cognitive bias modification [1, 4]. The findings suggest that contextual interventions can effectively reduce difficulties in emotion regulation and modify attributional styles, reducing aggressive behavior. However, since the study involved only university students aged 18–23, caution should be exercised in generalizing these results to younger adolescents, as developmental differences may influence the outcomes. This provides new empirical support for the application of emotion regulation and cognitive restructuring in managing adolescent aggression.

From a practical perspective, this study highlights the potential of online contextual interventions as accessible, cost-effective, and efficient methods for addressing aggression in adolescents. Techniques such as computer-assisted cognitive-behavioral therapy (CBT) and cognitive bias modification have demonstrated their ability to improve emotion regulation and correct maladaptive cognitive patterns by simulating real-life social situations. By providing immediate, actionable feedback, these interventions help adolescents practice adaptive coping strategies, enhancing their ability to manage ambiguous social cues and emotional responses [1, 4]. Schools and families can utilize these intervention methods to help adolescents enhance their emotion regulation and cognitive processing abilities, thereby reducing aggressive behaviors and potential criminal risks.

Despite the significant findings of this study, several limitations should be noted. First, the study lacks a longitudinal design, which prevents the evaluation of the long-term sustainability of the intervention effects. Future research should consider employing a longitudinal design to better understand the stability of the intervention outcomes. Second, due to space limitations, this study did not explore the direct effects of online contextual interventions on aggressive behavior in depth, thus limiting a comprehensive understanding of their potential role in reducing aggression. Additionally, the study sample was restricted to university students from Wenzhou University, which means the findings may reflect the characteristics of this specific group. Finally, this study primarily relied on self-report questionnaires. Although measures such as reverse scoring and anonymous reporting were employed to mitigate bias, self-report bias may still exist, potentially affecting the accuracy of the results. Future research could incorporate multiple data collection methods, such as behavioral observations or third-party assessments, to enhance the reliability of the data.

Future research could consider the following: (1) Conduct studies across different regions, cultural contexts, and age groups, particularly focusing on younger adolescents (under 18), to assess the external validity of the findings. This would help determine whether the interventions are equally effective for adolescents of various ages and backgrounds; (2) Design longitudinal studies to evaluate the long-term sustainability of intervention effects, ensuring that improvements in emotion regulation and hostile attribution bias are maintained over time, thus providing stronger evidence for the stability of the interventions; (3) Control potential confounding variables, such as participants’ family backgrounds and mental health conditions, to enhance the internal validity and reliability of the results; (4) Introduce diverse data collection methods, such as behavioral observations and third-party assessments, to reduce self-report bias and increase data accuracy; (5) Conduct baseline measurements of participants’ emotion regulation levels prior to the intervention to allow for a more comprehensive evaluation of intervention effects; (6) Explore the combined effects of multiple psychological intervention strategies, such as mindfulness training and cognitive-behavioral therapy, to identify the most effective integrated intervention model, thus providing a more comprehensive practical basis for the prevention of aggressive behavior and juvenile delinquency.

By exploring these avenues, future research can further deepen our understanding of how hostile attribution bias and emotion regulation abilities influence adolescent aggression. This will provide more effective crime prevention measures, ultimately promoting the healthy development of adolescents and the harmonious stability of society. Additionally, future research should consider conducting studies across different regions, cultural contexts, and age groups, particularly focusing on younger adolescents (under 18), to assess the external validity of the findings. This will help determine whether the interventions are equally effective for adolescents of various ages and backgrounds. Furthermore, future studies should investigate the cultural applicability of these interventions to ensure their effectiveness across diverse cultural settings, thereby enhancing the generalizability and broader applicability of these intervention methods [34].

In conclusion, this study confirms the effectiveness of online interventions, including computer-assisted CBT and cognitive bias modification, in enhancing emotion regulation and reducing hostile attributions among adolescents. These findings are consistent with previous studies [1, 4], which demonstrated the efficacy of CBT delivered through digital platforms and the benefits of cognitive bias modification techniques. Future research should continue to explore and refine these intervention models to better apply them to juvenile crime prevention and correction.

Conclusions

This study was conducted in two parts, systematically exploring the relationships between hostile attribution bias, emotion regulation abilities, and adolescent aggressive behavior, while also evaluating the effectiveness of online contextual interventions. This study yields the following key conclusions: (1) Hostile attribution bias and difficulties in emotion regulation are significant predictors of aggressive behavior in adolescents, with both factors contributing to the likelihood of aggressive actions. (2) Online contextual interventions led to a significant reduction in hostile attribution bias, alleviated difficulties in emotion regulation, and effectively decreased aggressive behavior, suggesting the potential of these interventions in mitigating key risk factors for juvenile delinquency.

In summary, this study underscores the critical role of hostile attribution bias and emotion regulation difficulties in predicting adolescent aggression. The findings also confirm that online interventions, such as computer-assisted CBT and cognitive bias modification, are effective tools for enhancing emotion regulation and reducing aggressive tendencies in adolescents [1, 4]. These findings provide new theoretical foundations and practical strategies for the prevention of adolescent aggression and criminal tendencies. Future research should further validate these findings and explore additional effective intervention methods to promote the healthy development of adolescents and the harmonious stability of society.