Digital shadow of childhood trauma: Associations between childhood maltreatment, moral disengagement, and online aggressive behavior in college students

Abstract

BACKGROUND

Online aggressive behavior (OAB) has emerged as a significant mental health concern with the proliferation of digital communication.

AIM

To examine the dual roles of moral disengagement as both a mediator and a moderator in the relationship between childhood maltreatment and OAB among college students.

METHODS

A cross-sectional survey was conducted among college students at two comprehensive universities in Chengdu, China. Of 512 questionnaires distributed, 482 (94.14%) were retained for analysis after excluding responses with extensive missing data or invalid response patterns. Participants completed validated self-report measures of childhood maltreatment (Childhood Trauma Questionnaire-Short Form), moral disengagement (Bandura’s Moral Disengagement Scale), and the OAB Scale. Mediation was examined using structural equation modeling with bootstrap estimation, and moderation was evaluated using hierarchical regression analysis.

RESULTS

In the final sample (n = 482), childhood maltreatment, moral disengagement, and OAB were significantly and positively intercorrelated (r = 0.43-0.52, P < 0.01). Moral disengagement partially mediated the association between childhood maltreatment and OAB (standardized indirect effect, 95% confidence interval: 0.14-0.34). Moral disengagement also significantly, though modestly, moderated this association (β = 0.09, P = 0.03): Childhood maltreatment was a strong correlate of OAB at high levels of moral disengagement (β = 0.53, P < 0.01), whereas the association was not statistically significant at low levels (β = -0.11, P = 0.28).

CONCLUSION

Moral disengagement was statistically associated with both the transmission and the strength of the relationship between childhood maltreatment and OAB among college students, consistent with a dual mediating and moderating pattern. These preliminary findings suggest that moral disengagement may be a worthwhile target for school-based interventions aimed at reducing cyber-aggression in this population, pending confirmation in longitudinal research.

Key Words: Childhood maltreatment; Moral disengagement; Online aggressive behavior; Mediation; Moderation; College students

Core Tip: This cross-sectional study examined whether moral disengagement simultaneously mediates and moderates the association between childhood maltreatment and online aggressive behavior in college students. The findings suggest that strengthening moral self-regulation could be a promising direction for reducing the behavioral correlates of early maltreatment, although longitudinal evidence is needed to confirm this.

INTRODUCTION

Online aggressive behavior (OAB), defined as deliberate harmful acts directed at others via the internet using electronic devices such as computers and mobile phones[1], has emerged as a significant mental health concern with the proliferation of digital communication. Epidemiological surveys indicate that approximately 60% of college students engage in OAB and around 55% have been victimized by it[2]. The consequences are substantial: For victims, OAB is associated with anxiety, depression[2], and more severe psychiatric outcomes including suicidal ideation and self-harm[3]; for perpetrators, repeated cyber-aggression may result in personality distortions, diminished self-awareness, impaired interpersonal functioning, and legal repercussions[4]. Given the breadth of its impact on academic performance, social development, and psychological well-being, identifying the antecedents and underlying mechanisms of OAB in college students is a research priority[5].

Childhood maltreatment, which encompasses physical, emotional, and sexual abuse as well as physical and emotional neglect perpetrated by adults in positions of trust or authority over the child[6,7], is among the most consistently identified distal risk factors for aggressive behavior[8-10]. In China, systematic reviews estimate the prevalence of childhood maltreatment at approximately 54%[11], placing it as a major public health issue. Neurobiologically, maltreatment disrupts the development of the prefrontal cortex and limbic system[12], impairing the emotion regulation and executive control capacities that normally inhibit aggressive responding. From a social-learning perspective, Script Theory proposes that children exposed to violence acquire cognitive scripts (internalized templates for aggressive behavior) that are rehearsed and reinforced over developmental time, increasing their later propensity for aggression[13]. Consistent with this framework, empirical studies have documented significant positive associations between childhood maltreatment and OAB[14,15], mediated by variables such as self-esteem[16], anger rumination[17], and neurobiological alterations[18]. The present study therefore posits hypothesis H1: Childhood maltreatment positively predicts OAB among college students.

The mechanisms connecting early adversity to later online aggression are likely cognitive as well as neurobiological. A particularly relevant construct is moral disengagement, which refers to the set of cognitive processes through which individuals selectively disengage the self-regulatory functions that normally inhibit conduct that violates one’s moral standards, thereby allowing themselves to engage in harmful behavior without experiencing guilt or self-censure[19]. Bandura et al[19] identified eight distinct mechanisms of moral disengagement: Moral justification, euphemistic labeling, advantageous comparison, displacement of responsibility, diffusion of responsibility, distortion of consequences, attribution of blame, and dehumanization. There is compelling evidence that childhood maltreatment promotes elevated moral disengagement by undermining moral development and normalizing harmful conduct[20-22]. In turn, higher moral disengagement is associated with greater frequency and intensity of OAB in college students[23]. Furthermore, moral disengagement has been identified as a mediator between a range of antecedents, including harsh parenting, hostile attribution, and exposure to violent media, and various forms of aggressive behavior[24-27]. Taken together, these findings suggest a pathway in which maltreatment fosters moral disengagement, which in turn facilitates OAB. We therefore propose hypothesis H2: Moral disengagement mediates the effect of childhood maltreatment on OAB among college students.

Although childhood maltreatment confers elevated risk for OAB, not all maltreated individuals display aggressive behavior[28-30], suggesting that important moderating factors exist. The General Aggression Model posits that provocative stimuli activate aggression-related cognitions and affect; whether a hostile response ensues depends on the psychological resources available to inhibit impulsive action[31]. Moral cognition normally serves as such a resource, providing an internalized brake against harmful conduct. However, moral disengagement undermines this regulatory function by constructing justifications for aggression[32,33]. Consistent with a person-context interaction perspective[34], individual differences in moral disengagement may therefore modulate the extent to which maltreatment history translates into OAB. Empirically, moral disengagement has been shown to amplify the relationship between trait anger and physical aggression, such that the association is strong in high-moral-disengagement individuals but non-significant in those with low moral disengagement[35]. Analogous moderating effects have been reported for empathy, anger rumination, psychopathic traits, and moral judgment[27,36,37]. We thus propose hypothesis H3: Moral disengagement moderates the association between childhood maltreatment and OAB among college students, with a stronger association at higher levels of moral disengagement.

Although mediation and moderation are conceptually distinct roles, they are not mutually exclusive for a single construct. As an intervening mechanism, moral disengagement may transmit part of the association between maltreatment and OAB, insofar as maltreatment tends to elevate this cognitive style on average. As a boundary condition, moral disengagement may separately determine how readily an individual's aggressive impulses translate into aggressive behavior, regardless of maltreatment history. This pattern parallels prior work in which a single social-cognitive variable has been reported to carry part of a predictor’s association with an outcome while also conditioning the strength of the remaining association[24-27,35-37]. We therefore examined moral disengagement in both capacities rather than assuming a priori that only one framework applies.

MATERIALS AND METHODS

Participants

Participants were recruited using a convenience sampling approach: Recruitment notices were posted on the campus online bulletin board system at two comprehensive universities in Chengdu, and students who volunteered to participate were subsequently assembled at a scheduled time in a large classroom to complete a paper-based questionnaire under standardized conditions. Eligible participants were currently enrolled undergraduate students aged 18 years or older who gave informed consent; students who did not complete the survey within the allotted classroom period, or who withdrew consent, were excluded. Of 512 questionnaires distributed, 30 were excluded before analysis because of extensive missing data (> 10% of items), straight-line response patterns, or completion times inconsistent with attentive reading, leaving 482 valid questionnaires (effective response rate: 94.14%). The final analytic sample comprised 256 males and 226 females aged 18-22 years (mean ± SD: 19.11 ± 1.66 years), representing diverse academic disciplines including Chinese language and literature, e-commerce, environmental engineering, and education. Prior to participation, all individuals provided written informed consent and received a nominal compensation of RMB 10. The Ethics Committee of the authors’ institution approved the study protocol in accordance with the Declaration of Helsinki and American Psychological Association ethical guidelines.

Instruments

Childhood Trauma Questionnaire-Short Form: The Childhood Trauma Questionnaire-Short Form[7] is a 28-item retrospective self-report measure assessing five dimensions of childhood maltreatment: Physical neglect, physical abuse, emotional abuse, sexual abuse, and emotional neglect. Items are rated on a 5-point scale from 1 (never) to 5 (always). The Childhood Trauma Questionnaire-Short Form has been validated in Chinese samples[16] and demonstrated acceptable internal consistency in the present study (Cronbach’s α = 0.74-0.82 across subscales).

Moral Disengagement Scale: Developed by Bandura et al[19], the Moral Disengagement Scale comprises 32 items covering eight mechanisms of moral disengagement (four items each). Responses are recorded on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree); higher total scores (range 32-160) indicate greater moral disengagement. The Chinese version of the Moral Disengagement Scale has demonstrated strong reliability and validity[35]. In the current study, Cronbach’s α was 0.87.

OAB Scale: The OAB Scale[38] assesses two dimensions of OAB: Overt aggression (7 items) and relational aggression (8 items), rated on a 4-point scale (1 = never to 4 = always). Higher scores indicate greater OAB. The scale has demonstrated good reliability and validity in Chinese college student populations[39,40]. In the present study, Cronbach’s α coefficients were 0.83 and 0.85 for the overt and relational aggression subscales, respectively.

Statistical analysis

Common method bias was assessed using Harman’s single-factor test; because this test alone cannot rule out common method variance when all constructs are measured by self-report, it is treated here as a preliminary screening step only. Descriptive statistics and Pearson correlation analyses were conducted using SPSS version 17.0 (SPSS Inc., Chicago, IL, United States). Mediation analysis followed the two-step procedure of Anderson and Gerbing[41], with structural equation modeling (SEM) conducted in AMOS 17.0 using maximum likelihood estimation. Mediation and moderation were tested in separate SEM and hierarchical regression models, consistent with the classic mediator-moderator distinction, because our hypotheses concerned two conceptually distinct roles of moral disengagement (an intervening mechanism vs a boundary condition) that are most directly tested with methods tailored to each; an integrated moderated-mediation model would nonetheless be a valuable extension in future work with a design powered for that analysis. Model fit was evaluated using χ²/df, standardized root mean square residual (SRMR), root mean square error of approximation (RMSEA), and comparative fit index (CFI), with acceptable fit defined as χ²/df < 5, SRMR ≤ 0.08, RMSEA ≤ 0.08, and CFI ≥ 0.90[42]; significance of indirect effects was evaluated via bootstrap estimation (1500 samples, 95%confidence interval). Moderation was examined using hierarchical regression with all predictors centered prior to analysis to address multicollinearity[43].

RESULTS

Common method bias and descriptive statistics

Harman’s single-factor test extracted 16 factors with eigenvalues > 1, with the first factor accounting for 25.51% of total variance, well below the 40% threshold, indicating that common method bias was unlikely to be a major concern; as noted in the Methods, this single-factor test provides only a preliminary indication rather than definitive evidence against common method bias, given that all variables were assessed via self-report. Pearson correlation analysis revealed significant positive associations among childhood maltreatment, moral disengagement, and OAB (Table 1).

Table 1 Correlations among childhood maltreatment, moral disengagement, and online aggressive behavior (n = 482).

Variable	Mean	SD	1	2
Childhood maltreatment	1.73	0.59	-	-
Moral disengagement	1.76	0.62	0.52b	-
OAB	2.48	0.58	0.43b	0.48b

^bP < 0.01.

OAB: Online aggressive behavior.

Mediation analysis: Moral disengagement as mediator

Following the two-step SEM procedure, the measurement model demonstrated adequate fit (χ²/df = 4.59, CFI = 0.94, SRMR = 0.03, RMSEA = 0.08), confirming that the 15 observed indicators adequately represented the three latent constructs (childhood maltreatment, moral disengagement, and OAB); the RMSEA of 0.08 indicates acceptable, rather than excellent, fit. The structural model also achieved satisfactory fit (χ²/df = 3.42, CFI = 0.96, SRMR = 0.03, RMSEA = 0.07). Bootstrap estimation (1500 resamples) yielded a significant standardized indirect effect of childhood maltreatment on OAB through moral disengagement (95% confidence interval: 0.14-0.34), with zero excluded from the interval, confirming partial mediation; this estimate did not adjust for gender, age, discipline, or time spent online. The standardized path coefficients are depicted in Figure 1.

Figure 1 Mediating model of childhood maltreatment, moral disengagement, and online aggressive behavior. ^bP < 0.01, all coefficients were standardized. Childhood maltreatment 1-5 were physical neglect, emotional abuse, emotional neglect, physical abuse, and sexual abuse; moral disengagement 1-8 were moral justification, euphemistic language, advantageous comparison, displacement of responsibility, diffusion of responsibility, distortion of consequences, attribution of blame, and dehumanization. OAB: Online aggressive behavior; Childhood maltreatment; MD: Moral disengagement.

Moderation analysis: Moral disengagement as moderator

Hierarchical regression analysis (Table 2) was conducted with OAB as the outcome variable. In step 1, childhood maltreatment significantly predicted OAB (β = 0.43, P < 0.01, R² = 0.19). Step 2 added moral disengagement (β = 0.33, P < 0.01), significantly improving model fit (ΔR² = 0.08, ΔF = 51.41, P < 0.01). In step 3, the childhood maltreatment × moral disengagement interaction term was significant (β = 0.09, P = 0.03; ΔR² = 0.01, ΔF = 5.07, P = 0.03), confirming significant moderation. Simple slopes analysis illustrated the moderation pattern: At high moral disengagement (+1 SD), childhood maltreatment was a strong and significant predictor of OAB (β = 0.53, P < 0.01). Conversely, at low moral disengagement (-1 SD), the association between childhood maltreatment and OAB was not statistically significant (β = -0.11, P = 0.28).

Table 2 Hierarchical regression analysis predicting online aggressive behavior from childhood maltreatment and moral disengagement (n = 482).

Variable	Step 1	Step 2	Step 3
Childhood maltreatment	0.43b	0.26b	0.25b
Moral disengagement	-	0.33b	0.32b
Childhood maltreatment × moral disengagement	-	-	0.09a
R2	0.19	0.26	0.27
ΔR2	0.19	0.07	0.01
ΔF	110.99b	51.41b	5.07a

^aP < 0.05.

^bP < 0.01.

All variables were standardized prior to analysis.

DISCUSSION

Principal findings

The present study examined whether moral disengagement mediates and moderates the association between childhood maltreatment and OAB in a sample of Chinese college students. All three hypotheses were supported. First, consistent with prior research[8-10,44,45], childhood maltreatment significantly and positively predicted OAB, confirming hypothesis H1. Second, moral disengagement partially mediated this association, accounting for a significant proportion of the indirect effect, thereby supporting hypothesis H2. Third, moral disengagement moderated the maltreatment–OAB relationship such that the association was pronounced at high levels of moral disengagement but negligible at low levels, confirming hypothesis H3. It should be noted that because the design was cross-sectional and relied on retrospective self-report, the interpretations above just describe statistical associations and should not be read as evidence of causal or temporal sequences among childhood maltreatment, moral disengagement, and OAB.

Childhood maltreatment and OAB

The positive association between childhood maltreatment and OAB replicates findings from both offline and online aggression literature[8-10,14,15]. From a social learning perspective, college students who were maltreated are more likely to have internalized aggressive behavioral scripts modeled by caregivers or significant adults during formative developmental periods[13,46]. These scripts, reinforced over time and now expressed through digital channels, increase susceptibility to OAB. The present data extend this framework to the online domain, underscoring that the “digital shadow” of early adversity is a meaningful clinical concern.

Mediating role of moral disengagement

The partial mediation finding is consistent with the General Aggression Model[31], which posits that environmental influences on aggression operate through internal cognitive states. Maltreatment environments may condition attitudes that minimize personal responsibility for harm, rationalize aggressive retaliation, and diminish empathy toward potential victims, which are precisely the cognitions captured by moral disengagement. This interpretation aligns with prior evidence linking maltreatment to impaired moral development[20-22] and with studies documenting moral disengagement as a mediator of various social influences on aggression[24-27]. That mediation was partial, rather than complete, suggests that additional psychological pathways, such as emotion dysregulation, hostile attribution bias, or impaired self-control, also contribute to maltreatment’s effect on OAB and warrant investigation in future research.

Moderating role of moral disengagement

The moderation finding is theoretically significant because it identifies a psychological boundary condition on the maltreatment-OAB pathway. Among students with high moral disengagement, a history of maltreatment was a strong predictor of OAB, presumably because the internal moral self-regulatory system that would normally inhibit cyber-aggression has been dismantled by the rationalizing mechanisms of moral disengagement[19,32]. Specifically, moral justification and displacement of responsibility may allow these individuals to frame OAB as deserved or consequence-free, rendering the behavioral threshold for aggression very low[47]. In contrast, students with low moral disengagement maintained a robust internal moral compass: Even those with significant maltreatment histories were no more likely to engage in OAB than their peers, likely because anticipated guilt and self-reproach functioned as effective deterrents[48]. These findings parallel reports of moral disengagement moderating the effects of trait anger, empathy, and anger rumination on aggression[27,35,36], and extend them to the domain of OAB driven by early-life adversity.

The mediation and moderation patterns observed here are consistent with, but do not uniquely establish, a moral-disengagement pathway. Other cognitive-affective and social processes, such as emotion dysregulation, hostile attribution bias, moral identity, empathy, impulsivity, peer influence, online disinhibition, and cyber-victimization, could plausibly account for part of the association we observed, either instead of or alongside moral disengagement. Future multi-mediator designs that assess these constructs together would help clarify their relative and joint contributions.

Clinical and educational implications

Practically, these findings point to moral disengagement as a promising intervention target for reducing OAB in college populations. Programs designed to strengthen moral reasoning, foster perspective-taking, and counteract mechanisms such as dehumanization and diffusion of responsibility may directly reduce OAB and also attenuate the downstream effects of childhood maltreatment. Such interventions could be incorporated into existing university mental health and character education curricula. Screening for childhood maltreatment in campus counseling services, and providing targeted psychological support that addresses maladaptive moral cognitions, may further reduce the risk of OAB in students with adverse childhood histories. Because our design was cross-sectional, these intervention implications should be regarded as hypotheses to be tested in intervention or longitudinal studies, rather than as an established basis for practice change.

Limitations

Several limitations should be noted. First, the cross-sectional design precludes causal inference; longitudinal or experimental designs are needed to establish directionality. Second, the model considered only moral disengagement as a mediator and moderator; other relevant cognitive-affective variables, such as self-control, emotion regulation, maladaptive schemas, and moral identity, may also play important roles[49-52] and should be incorporated in future multi-mediator models. Third, participants were recruited via convenience sampling through a campus bulletin board system platform at two universities in Chengdu, and academic year, socioeconomic background, and internet-use patterns were not collected; these features limit generalizability to other developmental stages, cultural contexts, or clinical populations, and preclude examining these unmeasured variables as covariates. Fourth, reliance on self-report measures raises the possibility of socially desirable responding, particularly for sensitive constructs such as childhood maltreatment and online aggression.

CONCLUSION

Moral disengagement was statistically associated with childhood maltreatment and OAB in this sample of college students, showing a pattern consistent with both partial mediation and moderation of this relationship. Students who reported childhood maltreatment and who also scored high in moral disengagement showed the highest levels of OAB, whereas those with low moral disengagement did not show a significant maltreatment-OAB association in this sample; this difference should not be read as evidence that low moral disengagement “protects” individuals, since the cross-sectional design cannot support such a causal claim. These findings support incorporating moral cognition into models of cyber-aggression and suggest morality-focused prevention and intervention programs within university settings as one plausible direction warranting further study.