This study employed a cluster randomized controlled trial (CRCT) design to evaluate the psychological impact of interactive multimedia legal publicity on vocational college students. Classes were used as the unit of randomization to minimize contamination between intervention and control groups, following established guidelines for educational intervention research (Yang et al., 2024). The study was conducted at a large vocational college in Southwest China from September 10, 2023 to December 15, 2023. Baseline (T0) measurements were collected 1 week before the intervention (September 3–7, 2023), and post-intervention (T1) measurements were obtained 1 week after the completion of the 8-week intervention period (December 18–22, 2023).

The studies involving human participants were reviewed and approved by the Academic Committee of Chongqing Polytechnic University of Electronic Technology (Approval Number: CQUET-EC-2023-001). The study was conducted in accordance with the Declaration of Helsinki and the ethical standards of the institutional research committee. Written informed consent was obtained from all participants prior to enrollment in the study. Participants were informed of their right to withdraw at any time without penalty. All data were anonymized during collection and analysis to protect participant privacy and confidentiality.

Participants were recruited from 14 classes across 7 academic departments within a large vocational college in Southwest China. Eligible participants were full-time students aged 18–25 years who provided informed consent and had active WeChat accounts for participation in class group communications. Students who had received systematic legal publicity activities within the previous 3 months or who had a diagnosed psychological disorder requiring clinical treatment were excluded from the study.

Using stratified random sampling, classes were first stratified by college department to ensure representation across academic disciplines. Within each stratum, classes were randomly assigned to intervention or control conditions using a computer-generated random number sequence created by a researcher not involved in participant recruitment or assessment. The randomization was performed at the class level rather than the individual level to minimize contamination between conditions, as students within the same class share common WeChat groups and daily interactions. Allocation concealment was maintained by using sequentially numbered, opaque, sealed envelopes that were opened only after all classes within each stratum had been enrolled. Class advisors and students were informed of their group assignment only after baseline data collection was completed.

Based on power analysis using G*Power software (version 3.1.9.6; α = 0.05, power = 0.90, expected effect size d = 0.50 derived from similar multimedia intervention studies), and accounting for the design effect of cluster randomization (DEFF = 1 + δ(m−1), where δ represents the assumed intraclass correlation coefficient of 0.05, consistent with typical values reported in educational intervention research (Hedges and Hedberg, 2007), and m represents the average cluster size of 36 students per class), the required sample size was determined to be 154 students per group. To accommodate potential attrition (estimated at 10%–20% based on similar educational intervention studies), 14 classes with approximately 500 students were recruited. The final sample comprised 500 students with 258 in the intervention group and 242 in the control group.

Data were collected using validated Chinese instruments at baseline (T0) and post-intervention (T1), selected based on their established psychometric properties in Chinese college student populations.

Originally developed by Zung (1971) and validated in Chinese contexts, the SAS assesses anxiety symptoms through 20 items rated on a 4-point scale (1 = none or a little of the time, 4 = most or all of the time). Sample items include “I feel more nervous and anxious than usual” and “I feel afraid for no reason at all.” Items 5, 9, 13, 17, and 19 were reverse-scored before analysis. Raw scores were multiplied by 1.25 to obtain standard scores, with scores ≥50 indicating clinical anxiety levels. The Chinese version demonstrates good internal consistency in college student populations, with previous studies reporting Cronbach’s α coefficients ranging from 0.82 to 0.89 (Yang et al., 2024). Confirmatory factor analysis supports a two-factor structure distinguishing cognitive and somatic anxiety symptoms. In the current study, internal consistency was adequate at both time points (T0: α = 0.78; T1: α = 0.76).

Developed by Xie (1998) specifically for Chinese populations, the SCSQ measures coping strategies through 20 items divided into positive coping and negative coping subscales, each rated on a 4-point scale (0 = never use, 3 = frequently use). The positive coping subscale (14 items) includes problem-solving, seeking social support, and cognitive reframing strategies, with sample items such as “Try to see the positive side of things” and “Seek help from professionals.” The negative coping subscale (6 items) encompasses avoidance, denial, and potentially maladaptive strategies, including items like “Fantasize how much better it would be if things were different” and “Make yourself feel better through smoking, drinking, taking medicine, or eating.” Higher scores indicate greater utilization of the respective coping style. The original validation study reported acceptable internal consistency for both subscales (positive coping: α = 0.78; negative coping: α = 0.72), and subsequent studies in Chinese college samples have demonstrated similar reliability patterns (Xu et al., 2024). Factor analysis confirms the two-factor structure with clear distinction between adaptive and maladaptive coping strategies. In the present study, reliability was acceptable for the positive coping subscale (T0: α = 0.67; T1: α = 0.73), while the negative coping subscale showed lower internal consistency consistent with previous findings in similar populations.

A structured questionnaire collected information on age, gender, ethnicity, academic major, academic performance ranking (categorized as top 25%, 25%–50%, 50%–75%, or bottom 25%), residential background (urban vs. rural), and paternal education level (ranging from no formal education to doctoral degree) to characterize the sample and identify potential confounding variables.

Prior to main analyses, the factor structure of all instruments was examined in the current sample. Both the SAS and SCSQ demonstrated factor structures consistent with their established theoretical models and previous validation studies in Chinese college student populations. The SAS supported the two-factor structure distinguishing cognitive and somatic anxiety components, while the SCSQ confirmed the distinction between positive and negative coping strategies. Measurement invariance across intervention and control groups was established for both instruments, supporting the validity of between-group comparisons.

Data collection occurred in standardized classroom environments under supervision of trained research assistants, with class teachers absent to minimize response bias. Participants received identical instructions, with approximately 30 min allocated for questionnaire completion. Participant anonymity was maintained through coded identification systems.

Data analysis was conducted using R (version 4.1.0). Data preprocessing included multiple imputation for missing values (<5%) and outlier detection. Normality was assessed via Shapiro–Wilk tests. Baseline group comparisons utilized chi-square tests for categorical variables and Mann–Whitney U tests for continuous variables.

Primary analyses used linear mixed-effects models (lme4 package) with class-level clustering as a random effect. Fixed effects included group, time, group × time interaction, and covariates (gender, academic ranking, residence). The group × time interaction served as the primary parameter of interest. Moderation analyses examined three-way interactions for gender, major type, and residence.

Clinical significance was evaluated by calculating improvement rates for participants with baseline clinical anxiety (SAS ≥ 50) and associated risk ratios. Effect sizes were calculated using Cohen’s d. Mediation analysis employed the mediation package with 5,000 bootstrap resamples to examine whether positive coping mediated the intervention-anxiety relationship.

All tests were two-sided with α = 0.05. Bonferroni corrections were applied for multiple comparisons. Missing data were addressed through multiple imputation when appropriate.

A total of 500 students from 14 classes across 7 academic departments at a large vocational college in Southwest China. The intervention group comprised 258 students from 7 classes, while the control group included 242 students from 7 classes. Table 1 presents the baseline characteristics of both groups.

The groups were comparable in most demographic variables, including gender (intervention: 56.2% male; control: 57.0% male; χ²(1) = 0.01, p = 0.924), ethnicity (both groups: 95.0% Han; χ²(1) = 0.00, p = 1.000), academic ranking (χ²(3) = 3.92, p = 0.270), and Family education background (based on the highest education level of either parent) level (χ²(6) = 8.44, p = 0.208). However, a significant difference was observed in residence type, with a higher proportion of urban residents in the intervention group (53.5%) compared to the control group (38.8%; χ²(1) = 10.69, p = 0.001). To account for this difference, we included residence as a covariate in all primary analyses and conducted sensitivity analyses with and without this covariate. The inclusion of residence did not substantially alter the pattern or significance of results, suggesting that the observed intervention effects were robust to this baseline difference.

At baseline, no significant differences were found in anxiety levels (SAS) (intervention median = 56.00, IQR = [50.00, 64.00]; control median = 56.00, IQR = [49.25, 65.00]; U = 30672.00, p = 0.790), positive coping strategies (SCSQ+) (intervention median = 23.00, IQR = [19.00, 28.00]; control median = 23.00, IQR = [19.00, 27.00]; U = 30884.00, p = 0.813), and negative coping strategies (SCSQ−) (intervention median = 14.00, IQR = [10.00, 17.00]; control median = 14.00, IQR = [11.00, 17.75]; U = 30950.50, p = 0.879). These results indicate that the intervention and control groups were well-balanced in terms of psychological outcomes at baseline.

Prior to hypothesis testing, we examined the intraclass correlation coefficients (ICCs) to assess the degree of clustering within classes and justify the use of mixed-effects models. The ICCs for the primary outcome variables were as follows: anxiety levels showed an ICC of 0.048, positive coping strategies showed an ICC of 0.052, and social support showed an ICC of 0.061. These values indicate that approximately 5%–6% of the variance in outcomes was attributable to class-level differences, confirming the appropriateness of multilevel modeling to account for the nested data structure. These ICC values, combined with the cluster sizes, yielded design effects greater than 1.5, further supporting the decision to use class as the unit of randomization and to include random intercepts for class in all analyses.

The mixed linear model analysis controlling for class-level clustering (Table 2) revealed significant intervention effects across all psychological outcomes, providing support for Hypotheses 1 and 2. Both groups showed improvements over time, but the intervention group demonstrated significantly better outcomes compared to the control group.

For anxiety levels (SAS), both groups showed reduced anxiety at post-intervention (β = −3.39, 95% CI [−4.8, −2.0]), but this improvement was significantly attenuated in the control group as indicated by the significant group × time interaction (β = 2.76, 95% CI [0.7, 4.8]). This result supports Hypothesis 1, suggesting that the interactive multimedia legal education was more effective in reducing anxiety compared to traditional legal education.

For positive coping strategies (SCSQ+), there was an overall improvement over time (β = 5.41, 95% CI [4.4, 6.4]), with a significant group × time interaction (β = −4.92, 95% CI [−6.3, −3.5]) indicating that the intervention group showed greater improvement in positive coping than the control group. For negative coping strategies (SCSQ−), post-intervention scores were lower overall (β = −3.99, 95% CI [−4.8, −3.2]), with a significant group × time interaction (β = 3.55, 95% CI [2.4, 4.7]), showing that the reduction in negative coping was more pronounced in the intervention group. These findings support Hypothesis 2.

Academic ranking showed a significant effect on some outcomes, with students in the top 25% showing higher anxiety levels (β = 1.35, p < 0.05, 95% CI [0.3, 2.4]) and lower positive coping (β = −0.80, p < 0.05, 95% CI [−1.5, −0.1]) compared to those in other ranking categories.

The moderation analysis (Table 3) examined whether the intervention effects varied by gender or urban–rural residence through three-way interactions (group × time × moderator). Contrary to our exploratory expectations, none of the moderation effects reached statistical significance.

For anxiety levels, the three-way interaction with gender was not significant (β = −0.056, SE = 2.108, t = −0.027, p = 0.979), nor was the interaction with residence (β = −0.019, SE = 2.113, t = −0.009, p = 0.993). Similarly, for positive coping strategies, neither gender (β = 0.299, SE = 1.433, t = 0.209, p = 0.835) nor residence (β = 0.228, SE = 1.437, t = 0.159, p = 0.874) significantly moderated the intervention effect. The same pattern was observed for social support outcomes, with non-significant interactions for both gender (β = 0.595, SE = 2.151, t = 0.276, p = 0.782) and residence (β = 0.233, SE = 2.165, t = 0.107, p = 0.915).

These null findings suggest that the intervention was equally effective across gender and residential background subgroups. Rather than indicating a limitation, this pattern supports the broad applicability of the interactive multimedia legal publicity approach across diverse student populations within the vocational college context.

Analysis of clinical outcomes revealed that among students with clinical anxiety at baseline (SAS ≥ 50), the improvement rate to subclinical levels was higher in the intervention group (18.4%) compared to the control group (11.6%). The intervention yielded a risk ratio of 1.59 and a number needed to treat of 14.7, indicating that approximately 15 students need to receive the intervention for one additional student to show clinically meaningful improvement.

Mediation analysis examined whether positive coping strategies mediated the relationship between the intervention and anxiety reduction, providing a test of Hypothesis 3 (Figure 1). Following Baron and Kenny’s approach with bootstrap confidence intervals, we tested the indirect effect of the intervention on anxiety through positive coping strategies.

The results supported complete mediation. The intervention had a significant effect on positive coping (path a: β = −4.919, SE = 0.709, p < 0.001), indicating that participants in the intervention group showed substantially greater improvements in positive coping strategies compared to controls. Positive coping, in turn, significantly predicted anxiety levels (path b: β = −0.429, SE = 0.044, p < 0.001), with higher positive coping associated with lower anxiety. The total effect of the intervention on anxiety was significant (path c: β = 2.759, SE = 1.043, p = 0.008), confirming the overall effectiveness of the intervention in reducing anxiety.

Critically, after controlling for positive coping, the direct effect of the intervention on anxiety became non-significant (path c′: β = 0.650, SE = 1.018, p = 0.524). The indirect effect through positive coping was significant (a × b = 2.110, bootstrap 95% CI [1.42, 2.85], p < 0.001), with the proportion of the total effect mediated being 76.5%. According to Baron and Kenny’s criteria, when the direct effect becomes non-significant while the indirect effect remains significant, this constitutes complete mediation.

These findings strongly support Hypothesis 3, indicating that the intervention’s anxiety-reducing effect operates primarily through enhanced positive coping strategies. These findings suggest that interactive multimedia legal publicity reduces anxiety by enhancing students’ positive coping strategies rather than through direct effects on anxiety per se. This mechanism highlights the importance of targeting coping skill development in legal publicity initiatives rather than focusing exclusively on knowledge acquisition.

The significant reduction in anxiety levels observed in the intervention group (Cohen’s d = 0.64) represents a robust effect that exceeds typical outcomes reported in educational anxiety interventions and approaches the effectiveness of specialized clinical interventions. Our results substantially surpass the average effect sizes found in virtual reality-based interventions for college student anxiety (d = 0.344–0.492) documented by Man et al. (2024) and Kulkarni and Weir (2024), while remaining within the range of credible psychological interventions. This enhanced effectiveness may be attributed to our theory-driven integration of micro-learning, narrative psychology, and social cognitive approaches, which simultaneously addresses cognitive load reduction, emotional engagement, and social learning processes.

The micro-learning component of our intervention, delivering content in 3–5 min segments, appears to have effectively reduced cognitive overwhelm that often accompanies legal information processing. This aligns with theoretical predictions that breaking complex information into manageable units enhances comprehension while reducing anxiety-provoking cognitive load (Laure et al., 2025). The narrative-based presentation of legal cases through first-person storytelling likely facilitated emotional connection and identification, helping students develop cognitive frameworks for understanding and managing legal stressors. This finding supports narrative psychology principles suggesting that story-based learning enhances emotional processing and reduces uncertainty-related anxiety (Zulhemay et al., 2024).

The substantial improvement in positive coping strategies (Cohen’s d = 0.52) and reduction in negative coping patterns (Cohen’s d = 0.40) demonstrate that the intervention successfully enhanced students’ psychological resources for managing legal challenges. These improvements align with social cognitive theory predictions that observational learning through peer interactions and multimedia modeling can enhance self-efficacy and promote adaptive coping responses (Pan et al., 2024). The WeChat group discussions and peer sharing components likely provided vicarious learning experiences that expanded students’ repertoire of coping strategies while reducing their reliance on avoidance and other maladaptive responses.

Our mediation analysis provides critical insight into the psychological mechanisms underlying the intervention’s effectiveness. The finding that positive coping strategies completely mediated the relationship between the intervention and anxiety reduction (76.5% of total effect) reveals the central role of enhanced psychological resources in promoting mental health. This result extends beyond previous research by demonstrating not merely that multimedia interventions reduce anxiety, but how they accomplish this effect through enhanced adaptive coping capabilities.

This complete mediation pattern aligns with theoretical frameworks emphasizing the importance of coping resources in psychological adjustment. The intervention’s multiple components likely contributed synergistically to coping enhancement: micro-learning modules provided structured knowledge that enhanced problem-solving confidence, narrative case presentations offered emotional frameworks for understanding legal stressors, and peer discussions facilitated social learning of diverse coping strategies. The WeChat group environment, familiar and comfortable for Chinese college students, created a supportive social context that encouraged coping skill practice and mutual support (Luo and Mohammed, 2023).

The complete mediation finding carries important theoretical and practical implications. The non-significant direct effect after controlling for positive coping indicates that the intervention’s anxiety-reducing benefit operates primarily through enhanced adaptive coping capabilities rather than through alternative pathways such as direct knowledge acquisition or reduced legal uncertainty. This pattern suggests that coping skill development represents the critical mechanism through which multimedia legal publicity promotes psychological wellbeing among vocational students. From a practical perspective, these findings underscore the importance of explicitly incorporating coping strategy training into legal education programs, as knowledge transmission alone may be insufficient to achieve meaningful reductions in law-related anxiety.

Contrary to our exploratory expectations, the moderation analyses revealed no significant three-way interactions between group, time, and demographic variables including gender and residential background. The intervention effects on anxiety reduction, positive coping enhancement, and social support did not differ significantly across these subgroups. While this null finding might initially appear disappointing, it actually carries important practical implications for intervention implementation.

The absence of significant moderation effects suggests that the interactive multimedia legal publicity intervention was equally effective for male and female students, as well as for students from urban and rural backgrounds. This broad applicability is a strength rather than a limitation, as it indicates that the intervention can be implemented across diverse student populations without requiring tailored modifications for specific demographic groups (Cuijpers et al., 2016). From a practical standpoint, this finding simplifies implementation decisions and suggests that educational institutions can deploy the intervention universally rather than targeting specific subpopulations.

Several factors may explain why the intervention demonstrated consistent effectiveness across demographic subgroups. First, the digital delivery platform through WeChat is ubiquitous among Chinese college students regardless of gender or residential background, ensuring equal access and familiarity with the intervention medium (Pan et al., 2024). Second, the legal topics addressed in the intervention, including campus loans, online fraud, employment rights, and rental disputes, represent universal concerns facing vocational college students irrespective of their demographic characteristics. Third, the combination of multiple theoretical approaches including micro-learning, narrative psychology, and social cognitive theory may have created a sufficiently comprehensive intervention that resonated with diverse learning preferences and needs (Laure et al., 2025).

These findings contrast with some previous research suggesting differential responses to educational interventions based on gender or socioeconomic background. While gender differences in psychological processes have been documented across various domains (Hyde, 2014), such differences may be attenuated when interventions are designed with universal accessibility and relevance in mind. Our results align with studies demonstrating that well-designed digital interventions can achieve equitable outcomes across diverse populations when the content is universally relevant and the delivery platform is equally accessible. A systematic review of digital mental health interventions for college students found that technology-based approaches can effectively reach diverse student populations and produce comparable outcomes across demographic groups (Lattie et al., 2019).

The intervention’s effectiveness in the Chinese vocational college context reflects important cultural and educational factors that enhance its relevance and impact. Traditional Chinese educational approaches, influenced by Confucian-heritage pedagogical traditions, have historically emphasized knowledge transmission through direct instruction and memorization (Yang et al., 2024). Our intervention’s success suggests that contemporary Chinese students, particularly digital natives, benefit from more interactive and socially-mediated learning approaches that complement rather than replace traditional methods.

The prominence of WeChat in Chinese digital culture provided an ideal platform for intervention delivery, as students were already comfortable with and actively engaged in class group communications (Pan et al., 2024). This cultural familiarity likely enhanced intervention engagement and reduced barriers to participation that might exist with less familiar platforms or technologies. The integration of peer discussion and collective problem-solving also aligns with Chinese cultural values emphasizing group harmony and mutual support in learning contexts.

The specific legal topics addressed in the intervention—campus loans, online fraud, employment rights, and rental disputes—reflect the most pressing concerns facing Chinese vocational college students as they transition toward independent adult roles. By addressing these authentic, personally relevant challenges, the intervention achieved both educational and psychological objectives that resonate with students’ immediate needs and long-term career preparation goals.

The intervention’s effectiveness in the Chinese vocational college context reflects important cultural and educational factors that merit consideration for understanding the underlying mechanisms and potential generalizability (Yang et al., 2024). Chinese cultural values and communication patterns likely influenced both the intervention design and participant responses in several important ways.

The collectivist orientation prevalent in Chinese culture may have enhanced the effectiveness of the peer-mediated learning components. Hofstede’s cultural dimensions framework identifies China as a highly collectivist society where group harmony, interdependence, and social relationships take precedence over individual autonomy (Hofstede, 2011). The WeChat group discussions fostered a sense of shared experience and mutual support that aligns with Chinese cultural emphases on group harmony and collective problem-solving (Luo and Mohammed, 2023). Students may have been more willing to engage in discussions and share experiences within the familiar context of their class WeChat groups, where established relationships and social norms facilitated open communication about potentially sensitive legal concerns.

Confucian educational traditions emphasizing respect for knowledge and authority may have facilitated acceptance of legal information delivered through institutional channels. Li’s (2012) comparative analysis of Eastern and Western learning models highlights how Chinese learners are socialized to value knowledge acquisition as a moral endeavor and to respect teachers and educational institutions as sources of wisdom. The intervention’s integration with existing class structures and delivery by class advisors positioned the content within established educational hierarchies that Chinese students are accustomed to respecting. This alignment with Confucian educational values, which emphasize hierarchical relationships between teachers and students alongside the moral purposes of education (Tan, 2017), may have enhanced the perceived credibility and relevance of the legal information provided.

The widespread adoption of WeChat as a primary communication platform among Chinese young adults provided a familiar and comfortable medium for intervention delivery (Pan et al., 2024). Unlike interventions requiring adoption of new technologies, our approach leveraged existing digital habits and social networks, reducing barriers to engagement. The integration of multimedia content with social interaction features aligned with Chinese students’ established patterns of digital communication and information consumption (Shi and Zhang, 2024).

These cultural considerations suggest that while the specific intervention components demonstrated effectiveness in the Chinese context, adaptation may be necessary for implementation in cultures with different communication norms, educational traditions, or digital platform preferences. A comprehensive review of meta-analyses on culturally adapted mental health interventions confirms that cultural adaptation of evidence-based interventions produces moderate to large effect sizes, and that interventions should be tailored to specific cultural contexts to maximize engagement and outcomes (Rathod et al., 2018). Future research should examine whether similar interventions produce comparable effects in other cultural contexts and identify which intervention components are culturally universal versus culturally specific.

Several methodological limitations warrant consideration in interpreting these findings. First, the single-institution design limits generalizability across China’s diverse vocational education landscape, as different institutional cultures and regional characteristics may influence intervention effectiveness. Second, the immediate post-intervention assessment precludes evaluation of long-term effects, particularly regarding coping skill maintenance and legal knowledge retention. Third, the reliance on self-report measures introduces potential social desirability biases, though this is mitigated by the use of validated instruments and anonymous data collection.

Future research should address these limitations through multi-site replication studies to establish broader applicability and extended follow-up assessments to evaluate sustained benefits over 6–12 months. Additionally, incorporating behavioral measures of coping effectiveness and objective indicators of legal knowledge application would strengthen the evidence base. Component analysis studies examining the relative effectiveness of micro-learning, narrative presentation, and peer discussion elements would inform intervention optimization and resource allocation for diverse educational contexts.

These findings have important implications for legal education policy in Chinese higher education. The demonstrated effectiveness of theory-based multimedia approaches suggests that institutions should integrate such methods into legal publicity programs, particularly given their cost-effectiveness and scalability.

The finding that positive coping strategies fully mediated the intervention’s effect on anxiety suggests that legal education programs should explicitly incorporate coping skill development rather than focusing solely on knowledge transmission. Additionally, the intervention’s consistent effectiveness across demographic subgroups supports universal implementation strategies without requiring costly demographic-specific adaptations.

The intervention’s success through existing institutional structures demonstrates that educational innovation need not require extensive new resources. Strategic enhancement of familiar platforms and established procedures can yield substantial improvements in both educational and psychological outcomes, making this approach suitable for widespread implementation in resource-constrained educational environments. These results also highlight opportunities for interdisciplinary collaboration between legal education, student affairs, and mental health professionals.