The proposed model specifies two complementary psychological mechanisms through which CHE relates to CE: a cognitive pathway represented by cultural identity (CI) and an affective pathway represented by cultural pride (CP). Rather than assuming a strict sequential process, CI and CP are theorized to operate in parallel. That is, heritage-related learning experiences may simultaneously strengthen students' sense of cultural self-definition (identity) and their positive emotional attachment to heritage (pride), and both orientations can motivate civic participation.

At the same time, the model allows for the possibility that cultural heritage education may also be associated with civic engagement through additional unmeasured mechanisms (e.g., civic efficacy, perceived social norms, or opportunities for participation). Therefore, in addition to the hypothesized indirect pathways, an alternative structural specification (partial mediation) that includes a direct CHE → CE path is examined empirically to provide a more realistic and reviewer-robust account of the relationship (see Section 4 for model comparison).

Participants were Chinese undergraduate students recruited via online announcements distributed through university learning platforms and student communication channels. Eligibility criteria required current full-time undergraduate enrollment. Participation was voluntary, anonymous, and uncompensated.

Before accessing the questionnaire, participants viewed an online informed consent statement describing the study purpose, confidentiality protections, and their right to withdraw at any time without penalty. Only participants who provided electronic informed consent were permitted to proceed to the survey. Data were collected using Wenjuanxing, a widely used online survey platform in China. Participants were instructed to complete the questionnaire independently and respond as honestly as possible. The average completion time was approximately 15 min.

A pilot study was conducted prior to the main data collection to assess item clarity and preliminary psychometric properties. After data screening, 59 valid responses were retained for pilot analyses. For the main study, a total of 744 valid questionnaires were obtained after excluding responses with unrealistically short completion times, evident patterned responding, or other indicators of insufficient effort.

The analytic sample consisted exclusively of undergraduate students, resulting in a relatively restricted age range with limited variance. Given this homogeneity and the study's theoretical focus on psychological value-internalization mechanisms rather than demographic prediction, demographic variables such as age were not included as exogenous controls in the structural model. This modeling decision was intended to maintain conceptual parsimony and to align the analysis with the primary research objective of examining mediation processes linking cultural heritage education to civic engagement through cultural identity and cultural pride.

The study employed an anonymous, non-interventional online questionnaire and involved no deception. No personally identifiable information (e.g., names, student identification numbers, or contact details) was collected. All responses were analyzed in aggregated form to protect confidentiality. The study targeted adult undergraduate participants and posed no more than minimal risk. Data were stored securely and used solely for research purposes. All participants provided informed consent prior to participation. The study involved anonymous, self-administered questionnaires and posed minimal risk to participants, and ethical procedures were conducted in accordance with institutional guidelines.

All constructs were assessed using self-report items administered in Chinese. The questionnaire included four variables: cultural heritage education (CHE; 5 items), cultural identity (CI; 4 items), cultural pride (CP; 4 items), and civic engagement (CE; 4 items). Items were rated on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Scale scores were computed as the mean of the items for each construct, with higher scores indicating higher levels of the corresponding variable. Detailed information regarding item sources, adaptations, and theoretical rationales for all measurement scales is provided in Supplementary Table S1.

Given that the primary aim of this study was to examine a mechanism-oriented value-socialization model involving multiple latent constructs and parallel mediation pathways, structural equation modeling (SEM) was selected as the primary analytic strategy. SEM is particularly suitable for simultaneously estimating relationships among latent variables, accounting for measurement error, and testing complex mediation structures within a theoretically specified model. All statistical analyses were conducted using SPSS Statistics 27.0, IBM Corp., Armonk, NY, United States and AMOS (version 31.0).

First, descriptive statistics and Pearson correlation analyses were performed in SPSS to examine preliminary associations among cultural heritage education, cultural identity, cultural pride, and civic engagement.

Potential common method bias was assessed using two complementary approaches. In SPSS, Harman's single-factor test was conducted via exploratory factor analysis (principal axis factoring), including all measurement items without rotation. In AMOS, a single-factor confirmatory factor analysis (CFA) model was also estimated by loading all items onto one latent factor. Model fit indices were examined to assess whether a single common factor could adequately account for the covariance among the observed measures.

SEM analyses followed a two-step modeling procedure. First, a CFA was conducted to evaluate the measurement model and assess construct validity for all latent variables. Second, the hypothesized structural model was tested to examine the relationships among cultural heritage education, cultural identity, cultural pride, and civic engagement.

To examine whether cultural identity and cultural pride fully or partially mediated the association between cultural heritage education (CHE) and civic engagement (CE), two theoretically specified structural models were estimated and compared. A fully mediated baseline model (Model A) constrained the direct path from CHE to CE to zero, reflecting a strict value-socialization assumption. A partial mediation model (Model B) additionally included a direct CHE → CE path to test whether any residual association remained after accounting for the two mediators. The two nested models were compared using a chi-square difference test and Akaike Information Criterion (AIC) values to evaluate relative model fit and robustness.

Overall model fit was evaluated using multiple indices, including the chi-square statistic, the comparative fit index (CFI), the Tucker–Lewis index (TLI), and the root mean square error of approximation (RMSEA). Indirect effects were tested using a bias-corrected bootstrap procedure with 5,000 resamples. All statistical tests were two-tailed, with a significance level set at 0.05.

The final sample consisted of 744 Chinese undergraduate students. As shown in Table 1, 42.9% of participants were male (n = 319) and 57.1% were female (n = 425). The sample was relatively evenly distributed across grade levels (Year 1–Year 4). Participants represented a broad range of academic disciplines, with comparable proportions from Humanities/History/Philosophy (24.5%), Economics/Management/Law (25.7%), Arts/Education (24.1%), and Science/Engineering (25.8%).

Table 2 presents descriptive statistics and correlations among the study variables. Participants reported relatively high levels of CHE (M = 4.17, SD = 0.75), CI (M = 4.22, SD = 0.73), CP (M = 4.23, SD = 0.74), and CE (M = 4.21, SD = 0.73). CHE was positively correlated with CI (r = 0.28, p < 0.001), CP (r = 0.31, p < 0.001), and CE (r = 0.28, p < 0.001). CI and CP were also positively correlated with CE (r = 0.30 and r = 0.29, respectively, both p < 0.001). Discriminant validity was supported by the Fornell–Larcker criterion, as the square roots of AVE (diagonal) exceeded the inter-construct correlations.

Common method bias was assessed using Harman's single-factor test and a single-factor CFA. For Harman's test, an unrotated exploratory factor analysis (principal axis factoring) was conducted including all measurement items. Four factors with eigenvalues greater than 1 emerged, and the first factor accounted for 32.45% of the total variance, below the commonly used 40% criterion. In addition, the single-factor CFA model (all items loading on one latent factor) demonstrated very poor fit, χ²(119) = 3, 326.62, χ²/df = 27.96, CFI = 0.457, TLI = 0.379, and RMSEA = 0.190 (90% CI [0.185, 0.196]). Taken together, the results indicate that the covariance among the measures is unlikely to be dominated by a single common-method factor; however, consistent with best practice for cross-sectional self-report designs, common method effects cannot be fully ruled out.

A confirmatory factor analysis (CFA) was conducted to evaluate the four-factor measurement model comprising cultural heritage education (CHE), cultural identity (CI), cultural pride (CP), and civic engagement (CE). The model showed good fit to the data, χ²(113) = 329.416, χ²/df = 2.915, CFI = 0.963, TLI = 0.956, and RMSEA = 0.051 (90% CI [0.044, 0.057], PCLOSE = 0.410), with acceptable absolute fit indices (GFI = 0.947, AGFI = 0.928, RMR = 0.028).

All standardized factor loadings were substantial and exceeded 0.70 (Table 3), supporting indicator reliability. Reliability and convergent validity were further supported (Table 4): Cronbach's α values ranged from 0.844 to 0.875, CR ranged from 0.823 to 0.859, and AVE ranged from 0.538 to 0.570, meeting recommended thresholds for internal consistency (Cronbach's α>0.70, CR >0.70) and convergent validity (AVE >0.50).

Structural equation modeling (SEM) was conducted to test the hypothesized relationships among cultural heritage education, cultural identity, cultural pride, and civic engagement. Based on the latest model comparison results, the primary model was specified as a partial-mediation model (Model B) that included the direct path CHE → CE. Model B demonstrated an acceptable fit to the data, χ²/df = 4.33, CFI = 0.936, TLI = 0.923, and RMSEA = 0.067. Figure 2 presents the standardized structural model for the primary partial-mediation specification (Model B). For clarity, the figure displays only the latent constructs and the structural paths among them. As shown in the model, CHE positively predicts CI and CP, and both mediators are positively associated with CE. In addition, a residual direct path from CHE to CE is retained to test partial mediation, and the residuals of CI and CP are allowed to correlate to account for shared unexplained variance. Detailed estimates and explained variance for each endogenous construct are reported in Table 5.

As shown in Table 5, cultural heritage education positively predicted cultural identity (β = 0.331, p < 0.001) and cultural pride (β = 0.355, p < 0.001), supporting H1 and H2. In turn, cultural identity (β = 0.156, p < 0.001) and cultural pride (β = 0.137, p = 0.003) were positively associated with civic engagement, supporting H3 and H4. Importantly, the direct effect of cultural heritage education on civic engagement also remained statistically significant (β = 0.134, p = 0.002), indicating partial mediation rather than a fully mediated relationship. With respect to explained variance, CHE accounted for 10.9% of the variance in cultural identity and 12.6% in cultural pride, and the full model explained 10.5% of the variance in civic engagement.

Accordingly, Model B was retained as the primary model for hypothesis testing, and Model A served as a nested baseline for comparison. To benchmark the primary partial-mediation model, a fully mediated baseline model (Model A) was also estimated by constraining the direct path CHE → CE to zero. Model A showed a comparable but slightly poorer fit than Model B (Model A: χ²/df = 4.38, CFI = 0.934, TLI = 0.922, RMSEA = 0.067; Model B: χ²/df = 4.33, CFI = 0.936, TLI = 0.923, RMSEA = 0.067). A chi-square difference test indicated that freeing the CHE → CE path significantly improved model fit, Δχ²(1) = 10.00, p = 0.002. Consistent with this improvement, Model B also yielded a lower AIC (571.97 vs. 579.97) and higher explained variance in civic engagement (RCE2=0.105 vs. 0.092). Therefore, Model B was retained as the primary model for inference.

Bootstrapping analyses with 5,000 resamples were conducted to examine the indirect effects of cultural heritage education on civic engagement through cultural identity and cultural pride. As reported in Table 6, the indirect effect via cultural identity was statistically significant (indCI = 0.073, BC 95% CI [0.040, 0.118], p < 0.001), and the indirect effect via cultural pride was also statistically significant (indCP = 0.069, BC 95% CI [0.036, 0.114], p < 0.001). The total indirect effect was significant (indTotal = 0.141, BC 95% CI [0.090, 0.203], p < 0.001), indicating that cultural identity and cultural pride function as parallel mediators linking cultural heritage education to civic engagement. In addition, the direct effect remained statistically significant (direct = 0.190, BC 95% CI [0.114, 0.277], p < 0.001), consistent with partial mediation. The total effect was also significant (total = 0.331, BC 95% CI [0.225, 0.444], p < 0.001).

Three findings merit emphasis.

First, CHE was positively associated with both CI and CP. Students reporting greater exposure to heritage-related learning experiences also reported stronger cultural identity (β = 0.331, p < 0.001) and greater cultural pride (β = 0.355, p < 0.001). This finding aligns with value-socialization theory in that educational experiences can shape both cognitive self-definition (identity) and affective attachment (pride) toward cultural heritage.

Second, both CI and CP were positively linked to CE. Cultural identity (β = 0.156, p < 0.001) and cultural pride (β = 0.137, p = 0.003) each contributed unique variance to civic engagement in the same structural model. Although modest in magnitude, these effects are meaningful in a large undergraduate sample and indicate that civic participation is associated not only with heritage exposure per se, but also with how students incorporate cultural meanings into their self-concept and emotional evaluations.

Third, the mediation evidence supported the proposed parallel mechanisms while also indicating partial mediation. Bias-corrected bootstrap tests showed significant indirect effects through both CI (indCI = 0.073, BC 95% CI [0.040, 0.118], p < 0.001) and CP (indCP = 0.069, BC 95% CI [0.036, 0.114], p < 0.001), with a significant total indirect effect (indTotal = 0.141, BC 95% CI [0.090, 0.203], p < 0.001). At the same time, the direct effect remained significant (direct = 0.190, BC 95% CI [.114, 0.277], p < 0.001), and the total effect was also significant (total = 0.331, BC 95% CI [.225, 0.444], p < 0.001). Together, these results indicate that CI and CP function as reliable parallel mediators, but they do not exhaust the linkage between CHE and civic engagement.

A key theoretical implication of the present findings lies in the conceptual differentiation between cultural identity and cultural pride as two distinct yet complementary value-socialization pathways. In line with the proposed framework, cultural identity is understood as a cognitive process of self-definition, reflecting individuals' perceived cultural belonging and incorporation of cultural meanings into the self-concept. By contrast, cultural pride is explicitly conceptualized as an affective evaluative orientation toward cultural heritage, capturing positive emotional appraisal and attachment rather than cognitive self-categorization.

Importantly, cultural pride is not treated as a cognitive component or sub-dimension of cultural identity. Instead, cultural identity and cultural pride are modeled as parallel mediators that operate through different psychological mechanisms. Whereas cultural identity addresses the question of “who I am” in relation to cultural heritage, cultural pride reflects “how I feel about what I belong to.” By modeling these constructs as parallel rather than hierarchical processes, the present study highlights that educational experiences may foster civic engagement not only by shaping students' cultural self-definitions, but also by cultivating positive emotional attachment and evaluative appreciation toward cultural heritage. In this sense, the present findings empirically operationalize value socialization as a dual-process mechanism involving both cognitive internalization and affective evaluation within the context of higher education.

This study contributes to the literature in several ways. First, it refines value-socialization accounts by identifying two complementary psychological routes through which cultural heritage education relates to civic engagement. Rather than treating cultural heritage education as a diffuse form of cultural exposure, the proposed model specifies a cognitive pathway (cultural identity) and an affective pathway (cultural pride), both of which independently link heritage education to civic engagement. Providing this mechanism-level distinction enhances theoretical precision regarding how educational inputs translate into civic outcomes.

Second, the finding of partial mediation offers a theoretically realistic extension of socialization models. Many value-socialization processes operate through multiple channels, and it is plausible that cultural heritage education also influences civic engagement through additional mechanisms not captured in the present model, such as civic efficacy, moral obligation, social norms, or increased access to civic opportunities within universities and surrounding communities. Retaining a direct path from cultural heritage education to civic engagement acknowledges these possibilities and avoids over-attribution to any single explanatory mechanism.

Third, the results help clarify the joint roles of identity and emotion in civic development models. Demonstrating simultaneous contributions of cognitive identification (cultural identity) and affective evaluation (cultural pride) bridges identity-based perspectives on civic behavior with emotion-based accounts of prosocial motivation. In educational contexts where heritage education aims to foster public-minded orientations and civic participation, both meaning internalization and positive emotional attachment appear to play important roles. Future theory-building may further examine the conditions under which one pathway becomes more salient than the other, such as variations in perceived authenticity of heritage experiences, campus cultural climate, or students' prior engagement with cultural heritage.

Several actionable implications follow for universities and educators.

First, heritage education may be most effective when it is designed to cultivate both identity and pride, rather than focusing narrowly on factual knowledge. To strengthen cultural identity, curricular designs can incorporate reflective and meaning-making components (e.g., guided discussion, narrative analysis, reflective writing) that connect heritage content to students' self-definition and perceived cultural membership.

Second, to foster cultural pride, heritage education should include emotionally engaging and experiential elements. Field visits, hands-on workshops, community-based documentation projects, and collaboration with cultural practitioners can enhance emotional connection and evaluative appreciation, especially when the activities are perceived as authentic and personally relevant.

Third, because civic engagement is ultimately behavioral, universities should translate heritage learning into structured opportunities for civic action. “Heritage-to-civic” pathways can be built through service-learning, volunteering, community education, public cultural events, or project-based courses that require civic contribution. Such designs help ensure that identity and pride are not only internalized, but also expressed through sustained participation.

Several limitations should be considered when interpreting these findings.

First, the cross-sectional design precludes strong causal inference and temporal ordering. It is plausible that civically engaged students are more likely to participate in, attend to, or positively report heritage-related learning experiences. Longitudinal, panel, and intervention designs are needed to test directional effects and change processes.

Second, all constructs were assessed via self-report at a single time point, which may inflate associations due to common method variance and social desirability. Future work should incorporate multi-source data and behavioral indicators of civic engagement (e.g., verified volunteering hours, participation records) and separate measurement occasions for predictors, mediators, and outcomes.

Third, the sample consisted of Chinese undergraduates recruited online, which may limit generalizability to other age groups, educational systems, or cultural contexts. Replication across regions, institutional types, and countries would help establish boundary conditions.

Finally, the explained variance in civic engagement was modest (RCE2=0.105), which is common for broad civic outcomes but indicates that additional predictors and contextual factors remain important. Civic engagement is a distal and multifaceted outcome; therefore, modest explained variance is common and theoretically expected in cross-sectional psychosocial models. Integrating variables such as civic efficacy, perceived social norms, sense of community, and the perceived authenticity/quality of heritage education may improve explanatory power and sharpen theoretical inference.

Future research can extend this work in several directions. First, longitudinal and quasi-experimental studies are needed to test whether CHE predicts changes in CI, CP, and civic engagement over time and to identify which instructional components (e.g., experiential learning, reflection, and community interaction) are most effective.

Second, multi-method approaches should be adopted. Combining self-report measures with behavioral/administrative indicators and using time-lagged measurement can reduce method bias and strengthen mediation inference.

Third, boundary conditions and moderators warrant examination. Prior heritage involvement, perceived authenticity, campus cultural climate, and socioeconomic background may influence whether identity-based or pride-based pathways are more salient. Expanding the model to include additional mechanisms (e.g., civic efficacy, moral identity, and perceived civic norms) may also provide a more comprehensive account of how heritage education connects to civic participation.

In sum, this study provides evidence for a mechanism-oriented value-socialization model in which cultural heritage education is positively associated with civic engagement through two parallel pathways—cultural identity and cultural pride—while also retaining a significant direct association. The findings suggest that heritage education may support civic engagement not only through knowledge and internalized cultural orientations, but also through additional, potentially contextual routes that deserve further investigation.