Impact of 3D learning resources on learning resilience: mediating roles of positive emotion and cognitive load

The study investigates whether and how 3D learning resources influence college students' learning resilience. Situated within the context of growing adoption of 3D technologies (e.g., 3D videos, animations, VR/AR) in higher education, the paper posits that these resources may not only enhance cognitive understanding but also affect emotional and psychological factors linked to resilience. The research addresses two questions: (RQ1) whether widespread use of 3D learning resources impacts students' learning resilience, and (RQ2) through what mechanisms this impact occurs. Drawing on broaden-and-build theory, the authors hypothesize that positive emotions broaden thought-action repertoires and help build resilience, while cognitive load theory suggests that well-designed 3D resources can reduce extraneous load, thereby supporting resilient learning. The purpose is to construct and test a structural model linking 3D learning resources, positive emotion, cognitive load, and learning resilience, clarifying direct and mediated effects and offering guidance for improving teaching quality and student persistence.

Theoretical background covers: (1) Learning resilience: defined as the learner's capacity to adapt and succeed when facing academic challenges. Prior work identifies multiple determinants, including emotions, family environment, teacher support, socioeconomic status, classroom characteristics, and collaborative learning. Table 1 summarizes empirical findings that social, teacher, and peer support, personality, SES, family communication, GPA, class characteristics, and collaborative learning positively relate to resilience. (2) 3D learning resources: 3D representations (3D videos/animations, VR/AR) can enhance visualization, comprehension of complex content (e.g., anatomy), and collaboration/engagement, potentially fostering resilience. Hypotheses include H1 (3D resources → resilience positive) and H2 (3D resources → positive emotion positive). (3) Broaden-and-build theory: positive emotions broaden cognition and build enduring resources like resilience. Empirical work shows associations between positive affect and resilience; thus H3 (positive emotion → resilience positive). (4) Cognitive load theory: working memory limits imply excessive load hinders learning and resilience. Well-designed 3D resources may reduce extraneous load; thus H4 (3D resources → cognitive load negative) and H5 (cognitive load → resilience negative). The study proposes a model where 3D resources influence resilience directly and indirectly via positive emotion and cognitive load.

Design: Cross-sectional survey with structural equation modeling (SEM). Setting: Universities in Shandong Province, China, during a province-wide classroom reform emphasizing digital/3D resources. Measurement: Seven-point Likert scale (1 = strongly disagree, 7 = strongly agree). The model includes eight variables and 35 items. 3D learning resources is a second-order construct with three first-order dimensions: information quality, system quality, and service quality (items adapted from Chiu et al., Lin & Wang, Wang & Chiu). Learning resilience is a second-order construct with three first-order dimensions: perseverance; reflecting and adaptive help-seeking; and negative affect and emotional response (items from ARS-30 by Cassidy). Positive emotion items adapted from Davis et al. Cognitive load items adapted from Hong et al. Pilot: Conducted with 5 teachers and 20 students to verify clarity. Data collection: Convenience sampling via Academic Affairs Offices; anonymous online survey administered to students in classes using 3D digital resources from 10/18/2023 to 12/06/2023. Responses: 1,253 collected; 963 valid retained. Sample demographics (N=963): 49.53% male, 50.47% female; ages mostly 18–20; academic years spread across freshman to senior; majors include management, economics, arts, philology, history, others. Analysis: SPSS 18.0 used for factor analysis, reliability, validity, and ANOVA homogeneity checks; Harman's single-factor test for common method bias. LISREL 8.7 used for SEM model estimation and fit assessment. Mplus 7.0 with 3,000 bootstrap samples used for mediation (95% bias-corrected CI). Data quality and validity: KMO = 0.90; one item (NA4) removed due to low loading (<0.7). Reliability: Cronbach's alpha and composite reliability (CR) for all constructs > 0.7. Convergent validity: AVE > 0.5 for all constructs. Discriminant validity: Square roots of AVE exceeded inter-construct correlations. Homogeneity: No significant differences in item understanding across gender, grade, and major (ANOVA). Common method bias: Largest single factor explained 34.16% variance, indicating no severe bias. Model fit: χ²/df = 2.931, GFI = 0.92, AGFI = 0.89, RMSEA = 0.075, PNFI = 0.83, PGFI = 0.68, CFI/NFI/IFI = 0.99.

- All five hypothesized paths supported with significant standardized coefficients: H1 (3D learning resources → learning resilience): β = 0.25, p < 0.001; H2 (3D learning resources → positive emotion): β = 0.89, p < 0.001; H3 (positive emotion → learning resilience): β = 0.68, p < 0.001; H4 (3D learning resources → cognitive load): β = −0.15, p < 0.001; H5 (cognitive load → learning resilience): β = −0.06, p < 0.001. - Mediation: Positive emotion and cognitive load significantly mediate the effect of 3D learning resources on learning resilience. Bootstrapped 95% CIs (3,000 samples) for indirect effects: positive emotion [0.067, 0.121], cognitive load [0.042, 0.090] (both exclude zero). - Reliability/validity: KMO = 0.90; all factor loadings ≥ 0.79 except NA4 (0.58, removed); Cronbach's alpha and CR > 0.7; AVE ≥ 0.68; discriminant validity satisfied. - Model fit indices indicate good fit: χ²/df = 2.931, GFI = 0.92, AGFI = 0.89, RMSEA = 0.075, PNFI = 0.83, PGFI = 0.68, CFI = 0.99, NFI = 0.99, IFI = 0.99. - Common method bias not critical: largest single factor variance = 34.16%. - Demographic attributes (gender, grade, major) did not significantly affect item understanding (homogeneity). - Descriptive trends: High means for 3D resource quality and resilience items (~5.7–5.97); moderate cognitive load (~4.2–4.6).

Findings confirm that 3D learning resources are associated with higher learning resilience directly and through two mechanisms: (a) by enhancing positive emotions, which broaden students' cognitive and behavioral repertoires in line with broaden-and-build theory, and (b) by reducing cognitive load, which supports efficient processing and resilient coping, per cognitive load theory. The results address RQ1 by demonstrating a positive overall impact of 3D learning resources on resilience and address RQ2 by identifying positive emotion and cognitive load as mediators. Although the correlations involving cognitive load and resilience are relatively weak, this is consistent with the multifactorial nature of cognitive load and resilience, which are influenced by course characteristics, individual abilities, and social-contextual factors. Practically, the study suggests designing 3D learning environments to minimize extraneous cognitive load and to incorporate interactive elements that stimulate positive emotions, thereby supporting persistence, adaptive help-seeking, and management of negative affect.

The study contributes a validated structural model showing that 3D learning resources improve college students' learning resilience both directly and indirectly via increased positive emotion and reduced cognitive load. It extends broaden-and-build theory into educational technology contexts and highlights the importance of cognitive load management in 3D learning design. The evidence supports broader adoption and thoughtful design of 3D resources to promote engagement and resilient learning. Future research should include longitudinal and mixed-methods designs, consider additional psychological and dispositional variables, and reassess these relationships as 3D technologies evolve.

- Cross-sectional survey limits causal and temporal inference; longitudinal tracking and time-series data are recommended. - Focus restricted to 3D learning resources, positive emotions, and cognitive load; other factors (e.g., cognitive styles, personality traits) may also shape resilience. - Findings are contextualized within current 3D technologies; impacts may change as technologies diversify and mature. - Convenience sampling in one province may limit generalizability.