This study uses a cross-sectional quantitative survey design to examine the relationships among Psychological Well-Being, Academic Gratitude, and Academic Buoyancy. Figure 1 presents a conceptual model of the relationships among the variables tested in this study.

This study involved 407 undergraduate students, representing 2.2% of the total 22,572 students enrolled at a state university in West Java. Participants were selected through stratified random sampling based on year of enrollment to ensure a representative sample. Multivariate outlier checks using Mahalanobis Distance (MD) identified three data points as outliers, which were excluded from the analysis.

The final sample consisted of 304 females and 103 males, with distribution based on year of study: first year (88), second year (118), third year (96), and fourth year (105).

Inclusion criteria included active undergraduate students aged 18–25 years enrolled in the current semester who provided consent to participate. Exclusion criteria included questionnaire completion rates below 80%, invalid response patterns (straight-lining), and inactive student status.

The sample size was determined by specific guidelines for covariance-based structural equation modeling (SEM), namely the ratio of estimated parameters to the number of respondents. According to Bentler and Chou (1987), the recommended minimum sample size is 5 times the number of parameters and can be increased to 5–10 times that number, as recommended by Gye-Soo (2016). This research model includes 52 parameters, so the minimum sample size required is 260 respondents. With 407 respondents, the research sample size meets and exceeds the general SEM recommendation of >300 respondents for stable and reliable model estimation (Kline, 2023).

This study uses three main instruments to measure Psychological Well-Being, Academic Gratitude, and Academic Buoyancy. All instruments have undergone linguistic and cultural adaptation in accordance with psychological scale adaptation guidelines.

Data was collected online via Google Forms after obtaining ethical approval from the University Ethics Committee. The questionnaire link was distributed to participants via campus email, official academic communication groups, and social media such as Line and WhatsApp. Before participating, participants received an explanation of the research’s purpose, data confidentiality, and the right to withdraw at any time. Participation was voluntary, and no incentives were given. Online surveys were chosen for their efficiency, although the potential for general and social method biases was still taken into account in interpreting the results.

Data analysis was performed using SPSS version 26 and LISREL version 8.80 through a covariance-based structural equation modeling (SEM) approach. The analysis stages included reliability checks, construct validity evaluation, measurement model testing, and structural model testing.

Construct validity was evaluated through Confirmatory Factor Analysis (CFA) based on standardized factor loadings, Composite Reliability (CR), Average Variance Extracted (AVE), and discriminant validity through the Heterotrait–Monotrait Ratio (HTMT). Internal reliability was evaluated using Cronbach’s Alpha and CR, with values≥ 0.70 as the acceptance criterion. Convergent validity was considered fulfilled if the AVE value was ≥ 0.50.

After the measurement model met these criteria, hypothesis testing was conducted using SEM. Model fit was evaluated using the χ²/df index, RMSEA, CFI, TLI, and SRMR, with thresholds referring to Hair et al. (2019). The list of indices and acceptance criteria is presented in Table 2 in the Results section.

Mediation analysis was performed using the bootstrapping procedure with 5,000 resamples to obtain more stable estimates of indirect effects. Mediation effects were considered significant if the 95% confidence interval did not include zero or if the p-value was < 0.05 (Hayes, 2022).

Evaluation of the measurement model showed that all constructs had adequate validity and reliability. All retained indicators had standard factor loadings (λ) ≥ 0.50, meeting the minimum threshold for indicator validity.

The Average Variance Extracted (AVE) values ranged from 0.50 to 0.67; thus, convergent validity was established for all constructs. In addition, the Composite Reliability (CR) was also strong (0.85–0.97), indicating good internal consistency (Hair et al., 2019).

Discriminant validity was tested using the Heterotrait–Monotrait Ratio (HTMT). The HTMT values for most construct pairs were below 0.85. However, the HTMT value for the PWB–AG pair was slightly above 1.0, suggesting potential construct overlap. In addition, the use of self-report instruments can increase method variance. Both of these issues are discussed further in the limitations section.

A summary of construct validity and reliability is presented in Table 1.

The measurement model shows acceptable fit, although some indices (e.g., SRMR) are slightly above the ideal range, necessitating careful model interpretation. The χ²/df value of 3.26 and RMSEA of 0.075 are within the acceptable range of fit, while CFI and TLI are 0.95, respectively, indicating a good level of model fit. The SRMR value of 0.084 is slightly above the ideal limit (< 0.08), but is still within the acceptable tolerance range in SEM analysis (Hair et al., 2019).

The complete model fit indices are presented in Table 2.

Structural model analysis shows that all paths between variables are statistically significant and consistent with the research hypotheses.

The results of hypothesis testing are as follows:

H1: PWB → AB (β = 0.50, t = 8.37, p < 0.001).

A summary of the structural path results is presented in Table 3, and an illustration of the structural model is shown in Figure 2.

Meanwhile, the coefficients of determination (R²) for the Academic Gratitude and Academic Buoyancy variables are shown in Table 4.

The mediating effect of Academic Gratitude was tested using the bootstrapping procedure (5,000 resampling). The results of the analysis showed that the indirect path from Psychological Well-Being (PWB) to Academic Buoyancy (AB) through Academic Gratitude (AG) was significant (p < 0.05) (Hayes, 2022). The results show that the indirect path PWB → AG → AB is significant (p < 0.05), indicating partial mediation. Thus, Academic Gratitude functions as an additional affective-cognitive pathway that helps explain the relationship between Psychological Well-Being and Academic Buoyancy.

The results of this study indicate a positive association between Psychological Well-Being (PWB) and Academic Buoyancy. These findings are consistent with the literature, which shows that PWB and the ability to cope with daily academic challenges often go hand in hand, particularly through the role of positive emotions. Previous studies have reported that individuals with higher levels of well-being tend to report more positive emotions—such as enjoyment, hope, and optimism—which are also found to be associated with higher levels of buoyancy (Hirvonen et al., 2020; Zhou and Wang, 2024). Additionally, Academic Buoyancy has been associated with more adaptive emotion regulation, fewer negative emotional experiences, and more positive emotions in the learning process; these emotional patterns are associated with more adaptive learning behaviors, such as a tendency to make plans and a decrease in avoidance behaviors (Granziera et al., 2024; Hirvonen et al., 2020; Wang and Hui, 2024).

The Broaden-and-Build Theory framework (Fredrickson, 2001) provides a theoretical basis for understanding the coexistence of PWB, positive emotions, and more adaptive response patterns in an academic context. In this framework, PWB can be viewed as a psychological resource that often accompanies more conducive emotional and cognitive conditions when facing learning demands. These findings also align with the report by Saleem et al. (2022), which shows that positive psychological resources are associated with adaptive response patterns in the work context. However, the associations found in this study remain non-causal given the cross-sectional design.

The second finding shows that students with higher levels of Psychological Well-Being (PWB) tend to be better able to feel and express gratitude in an academic context. Theoretically, this relationship is consistent with the explanation that individuals with high PWB have more adaptive emotional regulation, making it easier for them to recognize and maintain positive emotions, including gratitude. This ability is also supported by a tendency to engage in cognitive reappraisal, which is interpreting experiences more positively, which in turn facilitates the emergence of gratitude for various experiences and academic support (Man and Jing, 2025).

In addition, high PWB provides a psychological foundation that makes individuals more sensitive to positive experiences and the kindness they receive. Students with healthy psychological conditions tend to form more supportive social relationships and engage in prosocial behavior, so they are more likely to experience and appreciate help or support in the academic environment. This pattern of positive interaction also reinforces the tendency for gratitude to emerge (Nawa and Yamagishi, 2024; Prathham Arora et al., 2023).

The Broaden-and-Build Theory framework (Fredrickson, 2001) further clarifies this mechanism. This theory argues that positive emotions broaden the mindset and increase cognitive flexibility, thereby creating internal conditions conducive to the development of appreciation and gratitude. Thus, high PWB can be understood as a foundation that enables students to be more open to positive experiences, more responsive to social support, and ultimately more capable of expressing Academic Gratitude. However, because the research design is cross-sectional, the direction of this relationship cannot be fully ascertained; a reciprocal relationship between PWB and gratitude remains possible and needs to be further tested through longitudinal or experimental studies.

The results of this study show that Academic Gratitude (AG) is positively related to Academic Buoyancy. Students who tend to appreciate the help, learning opportunities, and academic experiences they receive generally also report more adaptive responses when facing daily academic challenges. These findings are in line with previous reports showing that gratitude is associated with stronger perceptions of support and more stable positive emotion regulation, both of which are associated with better academic adaptation patterns (Demichelis et al., 2024; Theiyab Alazemi et al., 2023).

Previous studies have also shown that gratitude often goes hand in hand with resilience, a tendency toward positive reframing, and the use of adaptive coping strategies, enabling individuals to maintain an optimistic attitude in challenging situations (Pohan et al., 2025; Zainoodin et al., 2021). In the context of Indonesian higher education, Academic Gratitude can be reflected in appreciation for faculty support, supportive social relationships, and meaningful learning opportunities (Nisa, 2022; Pohan et al., 2025), all of which are relevant for understanding variations in students’ ability to respond to learning obstacles.

Overall, these patterns support the understanding that AG and Academic Buoyancy often appear together through psychological mechanisms such as positive emotion regulation, perceptions of social support, and a tendency to find positive meaning in academic experiences. However, the relationship found is associative and cannot be interpreted as causal.

The bootstrapping results show that Academic Gratitude (AG) partially mediates the relationship between Psychological Well-Being (PWB) and Academic Buoyancy. These findings indicate that gratitude can be one of the affective-cognitive pathways that help explain the relationship between positive psychological conditions and students’ ability to cope with academic challenges, in line with the Broaden-and-Build Theory framework (Fredrickson, 2001). However, all of the relationships found are correlational.

Previous literature indicates that individuals with higher PWB tend to report more experiences of gratitude in academic contexts (Buenconsejo et al., 2024; Kusumawati et al., 2023). AG itself is often associated with motivation, learning engagement, self-efficacy, resilience, and social support (Buenconsejo et al., 2024; Liu et al., 2024). These patterns are also reported to be related to Academic Buoyancy, including the ability to manage stress, maintain optimism, and use adaptive coping strategies (Cui et al., 2023; Li et al., 2022).

In summary, AG emerges as a mechanism that helps bridge the relationship between PWB and Academic Buoyancy, without implying a causal relationship.

The evaluation of the measurement model indicates that all constructs exhibit adequate convergent validity and internal reliability, as indicated by AVE values ≥ 0.50 and CR values ≥ 0.85 (Hair et al., 2019).

However, the HTMT value for the PWB–Academic Gratitude construction pair is slightly above 1.0, indicating conceptual overlap between the two constructions. This may be due to the theoretical proximity between aspects of PWB (e.g., positive relationships, self-acceptance) and the tendency to feel grateful in an academic context. In addition, the use of self-report instruments may increase the general method variance, thereby strengthening the correlation between constructs. In contrast, the other construct pairs (PWB–AB and AG–AB) showed HTMT values < 0.85, indicating adequate discriminant validity.

These findings confirm that empirical differences between PWB and AG should be interpreted with caution. In future research, multi-trait multi-method approaches, observational data, or third-party reports could be used to minimize single-method bias and clarify conceptual boundaries between constructs.