Digital Leadership and Teacher Well-Being Across Educational Levels: A Quantitative Study Based on the Job Demands-Resources Model

• In Event: Teacher Agency, Well-Being, and Equity: Professional Development and Leadership in Diverse Contexts

Proposal

Teacher well-being (TWB) has become a central concern in global education setting, especially under the pressures of digital transformation and increasingly complex professional demands. While digital reforms have modernized teaching, they also introduced new stressors, positioning TWB as a key factor for sustainable education. TWB encompasses emotional, cognitive, social, and psychological dimensions, including both positive indicators (e.g., satisfaction, resilience) and negative ones (e.g., stress, burnout). Understanding the mechanisms that influence TWB is essential for improving education quality and retaining a stable teaching workforce.

This study adopts the Job Demands-Resources (JD-R) model to examine whether digital leadership (DL) that defined as the ability of school leaders to lead digital transformation through strategic vision, technical competence, and supportive management that may affects TWB. DL is hypothesized to promote TWB by increasing resources and reducing demands, with the JD-R model offering a more comprehensive framework than alternative theories.
Four hypotheses are proposed: H1: DL is positively associated with TWB. H2: Job resources mediate the relationship between DL and TWB. H3: Job demands negatively mediate the relationship between DL and TWB. H4: Educational level moderates the relationship between DL and TWB.

Methodology
A sequential quantitative design was employed with snowball sampling, involving 268 teachers from kindergartens, primary schools, and middle schools. DL was measured using the Digital Leadership Scale (DLS), and TWB was assessed via the PANAS and SWLS scales. Job resources (JR) and job demands (JD) were treated as mediators, while educational level served as a moderator.

Measurement Validation
Prior to structural modeling, exploratory factor analysis (EFA) was conducted to assess the construct validity of all measurement scales. The Kaiser-Meyer-Olkin (KMO) values ranged from 0.87 to 0.95 across all constructs, and Bartlett’s tests of sphericity were highly significant (p < .001), confirming sampling adequacy and factorability. Although parallel analysis suggested two to three factors for some scales, single-factor solutions were retained based on theoretical grounds and strong item loadings. All items demonstrated substantial loadings (mostly > 0.70), and the proportion of variance explained ranged from 45.4% (JD) to 78.3% (DL), indicating robust unidimensionality and internal consistency.

Confirmatory factor analysis supported the proposed measurement model, with all items loading significantly onto their respective latent constructs (standardized loadings mostly > 0.80). Although model fit indices were slightly below conventional thresholds (CFI = 0.807, RMSEA = 0.092, SRMR = 0.112), the factor structure remained theoretically coherent and statistically robust.

Findings
Structural equation modeling (SEM) revealed that DL significantly predicted job resources (β = 0.533, p < .001) and job demands (β = 0.183, p < .001). In turn, job resources positively influenced TWB (β = 0.848, p < .001), while job demands had a negative effect (β = –0.547, p < .001). SEM results showed that DL significantly predicted TWB (β = 0.429, R² = 0.227, p < .001), as well as JR and JD, supporting both direct and mediated pathways.

Regression analysis confirmed that DL significantly predicted TWB (β = 0.78, R² = 0.227, p < .001), supporting H1. Mediation analysis showed that JR positively mediated the DL–TWB relationship (ACME = 0.379, p < .001), while JD negatively mediated it (ACME = –0.067, p = .03), supporting H2 and H3.

To test H4, multi-group SEM was conducted across educational levels. Although the moderation effect was not statistically significant (p = .597), subgroup regression analyses revealed meaningful differences. DL had the strongest impact on TWB among middle school teachers (β = 0.83), followed by primary school teachers (β = 0.72). No significant effect was found among kindergarten teachers (β = –0.16), possibly due to lower digital engagement or limited leadership structures in early childhood settings. In addition, to better capture the multidimensional nature of TWB, this study modeled TWB as a second-order latent construct composed of three first-order factors: positive well-being (WBP), negative well-being (WBN), and satisfaction (WBS). This hierarchical structure conducted excellent model fit and aligns with prior frameworks (CFI = .990, RMSEA = .046, SRMR = .057), supporting the validity of the measurement model.

Discussion
These findings align with the JD-R model, highlighting DL’s dual role in enhancing resources and mitigating demands to promote well-being. While the moderating effect of educational level was not statistically significant, contextual differences suggest that DL may be more effective in primary and secondary schools. This may be due to greater digital maturity, stronger organizational structures, and more frequent interaction with school leaders in these settings. These findings underscore the importance of leadership strategies that are both digitally competent and emotionally supportive. As schools navigate the challenges of digital transformation, DL offers a pathway toward more equitable, stable, and humane educational systems.

Theoretical and Practical Implications
Theoretically, this study integrates DL into the JD-R model and expands its application by introducing digital-specific demands and resources, such as technological overload and online collaboration. It also highlights the contextual role of educational level in shaping leadership effectiveness.
Practically, the findings offer actionable insights for school leaders and policymakers. Promoting DL is not merely about technology adoption but about cultivating a supportive work environment. Leadership strategies that balance resource-building and stress reduction can enhance teacher morale, reduce turnover, and strengthen the resilience of the education system.

Conclusion
This study aligns closely with the theme of CIES2026, “Re-examining Education and Peace in a Divided World.” As digital transformation continues to reshape education, leadership that supports TWB is essential for building peaceful and resilient learning environments. By focusing on DL, TWB, and contextual adaptation, this research contributes to the broader conversation on educational equity, sustainability, and institutional stability in the digital age.

Authors

• Yongyi Rayna CHEN, the Education University of Hong Kong

• Kwok Kuen Tsang, The Education University of Hong Kong