Teachers’ psychological well-being and professional self-efficacy have long been recognized as pivotal determinants of educational quality in the Asia-Pacific region, where educational reforms increasingly emphasize both academic excellence and teacher well-being (Zee and Koomen, 2016). Teacher self-efficacy (TSE), defined as teachers’ beliefs in their capabilities to achieve desired instructional outcomes (Tschannen-Moran and Hoy, 2001), has been consistently linked to effective pedagogy, student motivation, and positive classroom climate (Burić and Kim, 2020; Jerrim et al., 2023). Meta-analytic evidence further indicates that TSE is a robust predictor of classroom processes, student engagement, and teachers’ psychological health (Zee and Koomen, 2016).

Beyond its direct association with instructional quality, TSE influences classroom outcomes through multiple mechanisms, including teachers’ instructional persistence and adaptive strategy use, as well as classroom management behaviors and emotional engagement with students (Duan and Zhao, 2024). Teachers with higher TSE are more likely to implement student-centered practices, respond constructively to classroom challenges, and sustain positive emotional climates, which in turn foster student motivation and engagement, particularly in primary school settings (Ma et al., 2025). However, existing evidence remains largely based on cross-sectional studies (Burić and Macuka, 2018; Klassen and Chiu, 2010; Li, 2023), providing limited understanding of how teachers’ efficacy beliefs develop over time within the dynamic sociocultural environments of Asia-Pacific schooling.

Although psychological capital, emotion regulation, and teacher self-efficacy are theoretically related and empirically correlated, they represent conceptually distinct constructs operating at different functional levels. Psychological capital reflects a set of relatively stable psychological resources that teachers draw upon when facing professional challenges (Luthans et al., 2007). Emotion regulation, by contrast, refers to dynamic regulatory processes through which individuals monitor, modify, and manage emotional experiences in situational contexts (Gross, 1998). Teacher self-efficacy represents an outcome-oriented belief system concerning teachers’ perceived capability to organize and execute instructional actions required to attain desired educational goals (Bandura, 1997; Tschannen-Moran and Hoy, 2001). Conceptually distinguishing psychological capital as a psychological resource, emotion regulation as a regulatory process, and teacher self-efficacy as an efficacy belief allows these constructs to be examined simultaneously without conceptual redundancy and provides a coherent basis for modeling a developmental pathway from psychological resources through regulatory processes to efficacy beliefs.

Within this evolving educational landscape, emotional processes play a central role in shaping teachers’ self-beliefs. Emotion regulation (ER)—the ability to monitor and adjust emotional responses to maintain effective functioning (Gross and John, 2003)—is a foundational mechanism that supports teachers’ resilience and instructional confidence (Brackett et al., 2010; Greenier et al., 2021). Teachers with strong ER capacities are more likely to manage stress, prevent burnout, and sustain positive interactions in emotionally demanding classrooms (Chang, 2020). Meanwhile, psychological capital (PsyCap)—a higher-order construct encompassing hope, efficacy, optimism, and resilience (Luthans et al., 2007)—represents a vital psychological resource that fuels professional engagement and well-being (Avey et al., 2011). In Asian educational systems, recent studies have begun to show that PsyCap enhances teachers’ adaptive coping, work engagement, and persistence (Gan and Cheng, 2021; Guo et al., 2022). Yet, empirical work examining how PsyCap and ER jointly contribute to TSE remains scarce, particularly in primary school settings and within the broader Asia-Pacific context. The resource–regulation–belief chain—linking psychological resources (PsyCap) to regulatory processes (ER) and self-efficacy beliefs (TSE)—has seldom been tested using longitudinal data, leaving temporal and developmental mechanisms underexplored.

The present study addresses these gaps by integrating PsyCap, ER, and TSE into a unified explanatory framework and empirically testing their direct and indirect relationships among Chinese primary school teachers. Specifically, this research contributes to the literature in three ways. First, it advances a theoretical model that conceptualizes the development of teacher self-efficacy as a function of teachers’ positive psychological resources and emotion-regulatory capacities, thereby enriching the resource–regulation–belief perspective in teacher psychology. Second, it employs both cross-sectional SEM and a three-wave time-lagged longitudinal design, using CFA/SEM and a time-lagged mediation model to examine whether psychological capital at an earlier time point predicts subsequent emotion regulation and, in turn, later teacher self-efficacy. Third, by focusing on Chinese primary teachers—a key yet underrepresented group in Asia-Pacific research—the study provides regionally grounded evidence that may inform teacher education and professional development initiatives across the region.

One-way analyses of variance (ANOVA) revealed systematic differences across key demographic and professional groups. Age-related differences were observed for psychological capital (PsyCap) and teacher self-efficacy (TSE), including all respective subdimensions (p < 0.01). Teachers aged 50 years and above consistently reported the highest levels of psychological resources and efficacy beliefs, followed by those aged 40–49 years, whereas younger teachers scored significantly lower. This pattern suggests that psychological capital and efficacy beliefs tend to strengthen with accumulated professional experience.

Significant differences were also found across professional title groups (p < 0.001). Teachers holding senior or advanced titles reported higher levels of PsyCap, emotion regulation (ER), and TSE than teachers of lower rank, indicating that professional seniority is associated with stronger psychological resources, greater regulatory capacity, and more robust instructional beliefs.

Differences related to administrative roles and school context further supported this trend. Teachers serving in school-level or middle management positions demonstrated higher teacher self-efficacy—particularly in student engagement—than homeroom and non-administrative teachers (p < 0.05). In addition, teachers from different school locations differed significantly on several indicators of self-efficacy and emotion regulation (p < 0.05). Taken together, these findings suggest that age, professional seniority, leadership experience, and school context constitute important background factors shaping teachers’ psychological resources, emotion regulation competence, and efficacy beliefs in Chinese primary education.

Table 2 presents the descriptive statistics, reliability, and validity indices for psychological capital (PsyCap), emotion regulation (ER), and teacher self-efficacy (TSE) among 606 Chinese primary school teachers. Teachers reported moderately high levels of PsyCap (M = 4.15, SD = 1.00) and ER (M = 4.89, SD = 1.16), whereas TSE (M = 5.94, SD = 1.48) showed the highest mean, indicating generally strong self-beliefs. Given the observed standard deviations (SD_TSE = 1.48; SD_PsyCap = 1.00), teacher self-efficacy exhibited the greatest dispersion, whereas psychological capital showed the least, indicating greater individual variability in efficacy beliefs. All skewness and kurtosis values were within ±2, confirming normality. PsyCap (Sk = −0.77, Ku = −0.60) and TSE (Sk = −1.00, Ku = −0.06) showed mild negative skew and near-normal distributions, whereas ER (Sk = −1.07, Ku = 0.21) was slightly more left-skewed, implying overall high and homogeneous emotion-regulation ability.

The measurement model exhibited strong psychometric properties. Composite reliability (CR = 0.94, 0.88, 0.87) exceeded 0.70, and average variance extracted (AVE = 0.70, 0.56, 0.59) surpassed 0.50, indicating good internal consistency and convergent validity. Maximum shared variance (MSV = 0.31–0.45) was below corresponding AVE values, confirming discriminant validity. MaxR(H) coefficients (0.925, 0.894, 0.879) further supported construct reliability. All inter-construct correlations were positive and significant (p < 0.01): PsyCap correlated with ER (r = 0.52) and TSE (r = 0.66), and ER correlated with TSE (r = 0.55). These results suggest that teachers with higher psychological capital tend to regulate emotions more effectively and hold stronger instructional efficacy beliefs. Overall, the measurement model demonstrated excellent reliability and validity, providing a solid basis for subsequent SEM analyses.

Given the moderate to strong correlations among psychological capital (PsyCap), emotion regulation (ER), and teacher self-efficacy (TSE), it is important to clarify their conceptual and empirical distinctiveness. Conceptually, PsyCap represents a higher-order positive psychological resource encompassing hope, optimism, resilience, and general self-efficacy, whereas ER refers to regulatory processes governing emotional responses, and TSE reflects a domain-specific belief regarding instructional capability. Thus, the three constructs capture related but theoretically separable aspects of teachers’ psychological functioning.

Empirically, discriminant validity was supported through confirmatory factor analysis. For all constructs, average variance extracted (AVE) exceeded the recommended threshold of 0.50, and the square roots of AVE were greater than the corresponding inter-construct correlations. In addition, maximum shared variance (MSV) values were consistently lower than AVE values, indicating that shared variance between constructs did not exceed variance explained by their indicators. These results suggest that PsyCap, ER, and TSE are empirically distinguishable rather than reflections of a single latent factor.

To further reduce the risk of spurious associations and Type I statistical error, mediation effects were tested using bias-corrected bootstrapping with 5,000 resamples, which does not rely on normality assumptions and provides more conservative confidence intervals. Moreover, although self-efficacy is included as a subdimension of PsyCap, it reflects a generalized psychological resource, whereas TSE represents task- and context-specific instructional beliefs; consistent with prior research, these constructs were therefore retained and modeled as theoretically related but non-redundant. Taken together, these analytical safeguards enhance confidence that the observed relationships reflect substantive mechanisms rather than measurement overlap or statistical artifacts.

Table 3 presents the goodness-of-fit indices for the hypothesized structural equation model (SEM). The results indicate that the proposed structural model demonstrates an acceptable fit to the data. The ratio of chi-square to degrees of freedom (χ²/df = 1.509) is well below the recommended threshold of 3.0, indicating an excellent overall model fit. Both the Goodness-of-Fit Index (GFI = 0.913) and the Adjusted Goodness-of-Fit Index (AGFI = 0.901) exceed the conventional cutoff of 0.90, indicating that the model adequately fits the observed data. Furthermore, the Comparative Fit Index (CFI = 0.972), Tucker–Lewis Index (TLI = 0.970), and Incremental Fit Index (IFI = 0.972) all exceed the 0.95 criterion, demonstrating superior fit relative to the baseline model. The Normed Fit Index (NFI = 0.923) also exceeds 0.90, further supporting the model’s robustness. The Root Mean Square Error of Approximation (RMSEA = 0.029) is below the stringent cutoff of 0.05, indicating a minimal discrepancy between the implied and observed covariance matrices and confirming an excellent fit. Taken together, all indices meet or exceed the “excellent” or “acceptable” benchmarks, suggesting that the proposed structural model fits the data well and effectively captures the latent relationships among the constructs.

Figure 1 presents the results of the structural equation modeling (SEM) analysis. Psychological capital exerted a significant positive effect on emotion regulation (β = 0.631, p < 0.001), indicating that teachers with higher levels of psychological resources—namely hope, optimism, resilience, and self-efficacy—are more likely to adopt adaptive emotion-regulation strategies. Moreover, psychological capital had a significant direct effect on teacher self-efficacy (β = 0.566, p < 0.001), suggesting that teachers with richer psychological resources tend to exhibit stronger confidence and a greater sense of control in their instructional practice. In addition, emotion regulation was positively associated with teacher self-efficacy (β = 0.333, p < 0.001), implying that teachers who can effectively manage their emotions perceive themselves as more competent and efficacious in teaching.

The findings provide empirical support for the hypothesized mediation model and identify emotion regulation as a critical link between teachers’ psychological capital and self-efficacy. Results from the bootstrapping mediation analysis (5,000 resamples) demonstrate that psychological capital significantly influences teacher self-efficacy (β = 0.776, 95% CI [0.718, 0.827], p < 0.001), with a notable direct effect (β = 0.566, 95% CI [0.465, 0.663], p < 0.001). Emotion regulation exhibited a significant partial mediating effect (β = 0.210, 95% CI [0.138, 0.290], p < 0.001), indicating that emotional strategies serve as a primary mechanism by which psychological capital enhances self-efficacy. Table 4 presents the detailed results of the mediation analysis.

The three-wave time-lagged mediation model demonstrated satisfactory model fit (χ²/df = 1.742, CFI = 0.958, TLI = 0.952, RMSEA = 0.042). Table 5 summarizes the standardized path coefficients and confidence intervals for the longitudinal mediation model. In the three-wave time-lagged mediation model, PsyCap at T1 significantly predicted ER at T2 (β = 0.48, p < 0.001) and TSE at T3 (β = 0.29, p < 0.001). In addition, ER at T2 significantly predicted TSE at T3 (β = 0.24, p < 0.01), providing longitudinal support for ER as an intervening mechanism. The indirect longitudinal effect of PsyCap_T1 on TSE_T3 via ER_T2 was significant (β = 0.12, p < 0.01; 95% CI [0.06, 0.20]), accounting for approximately 29% of the total longitudinal association. Figure 2 illustrates the corresponding three-wave time-lagged mediation structure linking PsyCap, ER, and TSE across time. These findings provide temporal support for a resource–regulation–belief pathway, whereby teachers’ psychological capital initiates adaptive emotion regulation that fosters stronger efficacy beliefs over time.

This study examined a resource–regulation–belief framework connecting psychological capital (PsyCap), emotion regulation (ER), and teacher self-efficacy (TSE) among Chinese primary school teachers. Structural equation modeling revealed three key findings. First, PsyCap positively predicted ER, suggesting that teachers with higher hope, optimism, resilience, and generalized self-efficacy are more likely to adopt adaptive emotion-regulation strategies (Bing et al., 2022; Gross, 1998). Second, PsyCap directly predicted TSE, supporting the view that psychological resources strengthen efficacy beliefs (Avey et al., 2011; Bandura, 1997; Luthans et al., 2007). Third, ER positively predicted TSE and partially mediated the PsyCap–TSE relationship, accounting for 27% of the total effect (Burić and Macuka, 2018). Longitudinal analyses further indicated that PsyCap predicted later ER and TSE, providing temporal support for its role as an upstream psychological resource that initiates adaptive regulation and contributes to teachers’ subsequent efficacy beliefs.

Beyond replicating prior findings on the positive associations among psychological capital, emotion regulation, and teacher self-efficacy, the present study extends existing literature in several important ways. First, the associations among psychological capital, emotion regulation, and teacher self-efficacy observed in the present study are consistent with prior cross-sectional and longitudinal research indicating that psychological resources are linked to adaptive regulation processes and stronger efficacy beliefs (Burić and Macuka, 2018; Ma, 2023), while further situating these relationships within a unified resource–regulation–belief framework. Second, compared with studies conducted in Western contexts, the relatively strong role of emotion regulation observed in this Chinese primary-school sample may reflect the high emotional labor, hierarchical organizational structures, and strong normative expectations placed on teachers in collectivist educational settings (Day and Gu, 2013; Yin et al., 2016). In such contexts, teachers’ capacity to regulate emotions may be especially critical for transforming internal psychological resources into sustained instructional confidence.

Methodologically, by combining cross-sectional SEM with a three-wave time-lagged mediation model, this study advances prior research that has largely relied on single-wave designs. Although causal inference remains cautious, the temporal ordering of PsyCap, ER, and TSE provides stronger support for the proposed mechanism than cross-sectional evidence alone.

These findings align with prior research demonstrating that TSE underpins classroom processes and teacher well-being (Tschannen-Moran and Hoy, 2001), and that ER supports effective functioning under emotional demands (Yin et al., 2016). By integrating Luthans’ positive psychological capital theory with Gross’s process model of ER, the present study illustrates how PsyCap not only buffers stress but also generates proactive, antecedent-focused strategies that reinforce efficacy beliefs (Luthans et al., 2006; Sweetman and Luthans, 2010). The observed temporal pattern among psychological capital, emotion regulation, and teacher self-efficacy suggests that teachers’ psychological resources function as foundational elements in professional development, rather than merely short-term outcomes of motivational states. In the Chinese primary education context, characterized by high emotional labor, hierarchical structures, and collective norms, PsyCap appears particularly critical for maintaining self-efficacy and emotional resilience (Day and Gu, 2013; Klassen and Chiu, 2010). Practically, the findings highlight a dual-track professional development approach: (1) PsyCap empowerment through goal-setting and pathway training (hope), attribution retraining and learned optimism (optimism), mastery modeling (self-efficacy), and resilience-building programs (resilience) (Luthans et al., 2006). (2) ER skill training emphasizing antecedent-focused strategies—such as cognitive reappraisal, situation modification, and attentional deployment—while reducing dependence on suppression (Gross, 1998; Hulsheger et al., 2013).

School-level systems including mentoring, psychologically safe climates, and recognition of effort and growth can amplify these effects (Collie et al., 2012; Jennings and Greenberg, 2009). In the era of digital education, emerging AI-supported tools may further facilitate reflective emotion regulation and strengthen the connections between teachers’ psychological resources, regulatory processes, and efficacy beliefs (Celik et al., 2022; Holmes et al., 2019). Attention, however, should be given to technostress and digital inequality to ensure that technology enhances rather than erodes teachers’ psychological resources (Lin et al., 2023; Yang and Du, 2024).

Several limitations warrant caution. Although the longitudinal design provides temporal validation, causal inference remains constrained, and unmeasured contextual factors could jointly affect PsyCap, ER, and TSE (Hamaker et al., 2015; Ilies et al., 2007). Future research could adopt multi-wave, multi-level, or intervention designs to test causal directions and contextual moderators. Cross-cultural comparisons within the Asia-Pacific region would further clarify the boundary conditions of the resource–regulation–belief mechanism across diverse educational systems. In summary, this study offers evidence that teachers’ psychological resources are transformed into stronger efficacy beliefs through emotion regulation. Integrating PsyCap cultivation with ER-focused strategy training and embedding these within supportive organizational contexts may sustainably enhance teachers’ instructional confidence and well-being, thereby improving students’ learning experiences.

This study validates a resource–regulation–belief framework among Chinese primary school teachers, demonstrating that psychological capital (PsyCap) enhances teacher self-efficacy (TSE) directly and indirectly through emotion regulation (ER). Teachers with higher levels of hope, optimism, resilience, and self-efficacy are more likely to engage in adaptive emotion-regulation strategies—particularly cognitive reappraisal—when facing classroom stress. Over time, these regulatory processes help sustain and strengthen teachers’ instructional confidence. Longitudinal evidence from the present study indicates that psychological capital operates through emotion regulation to support the development of teacher self-efficacy, highlighting a process through which psychological resources buffer stress and facilitate efficacy growth. Grounded in Luthans’ PsyCap theory and Gross’s process model of emotion regulation, these findings extend positive psychology into the Chinese context, where teachers often face high emotional labor and hierarchical work structures (Erden, 2025).

Practically, teacher development should integrate PsyCap enhancement and ER skill training. Effective modules include goal-setting and pathway planning (hope), attribution retraining and learned optimism (optimism), mastery modeling and resilience-building (resilience), and cognitive reappraisal workshops (ER) (Bertieaux et al., 2024; Liu and Du, 2024). Supportive school climates that promote collegial trust and professional growth further sustain teachers’ emotional well-being and instructional quality (Collie et al., 2012; Jennings and Greenberg, 2009).

In AI-supported classrooms, intelligent dashboards and tutoring systems can serve as emotional scaffolds that reduce cognitive load and enhance self-reflection (Chen et al., 2024; Duan and Zhao, 2024; Oh and Ahn, 2024). Future longitudinal and mixed-method research should explore how psychological capital shapes teachers’ emotion-regulation processes and, in turn, their self-efficacy as these relationships unfold within contexts of digital transformation, thereby generating scalable evidence to inform teacher development across the Asia-Pacific region.

Several limitations should be acknowledged. Although the three-wave longitudinal design represents a methodological strength, causal inferences remain limited. Unmeasured contextual factors—such as teacher burnout, school climate, leadership support, or student characteristics—may simultaneously influence psychological capital, emotion regulation, and teacher self-efficacy. Future studies could incorporate these variables within multi-level or integrative models to more comprehensively capture the contextual conditions under which psychological resources are translated into efficacy beliefs.

In addition, the three-month intervals between measurement waves, while consistent with prior longitudinal studies, may not fully capture longer-term developmental changes in teachers’ emotion regulation and efficacy beliefs. Future research could enhance temporal precision by increasing the number of measurement waves, standardizing or varying time lags, and implementing strategies to reduce participant attrition. Moreover, intervention-based or quasi-experimental designs, as well as multi-level modeling approaches, would help clarify causal pathways and examine how individual psychological resources interact with school-level conditions over time.