Psychological well-being (PWB) is a multidimensional concept encompassing personal growth, autonomy, and positive relationships (Ryff and Keyes, 1995; Schmitt et al., 2014). It emphasizes realizing one’s potential and leading a meaningful life (Houben et al., 2015; Huppert, 2009; Ryff, 2013), contrasting with hedonic well-being, which focuses on happiness and pleasure. In the context of education, PWB influences not only emotional well-being but also cognitive functioning, motivation, and academic performance (Huppert, 2009; Ryff, 2013). University students, who face significant academic demands and life transitions, are particularly vulnerable to fluctuations in PWB (Jackman and Sisson, 2022).

Because of its connection to academic outcomes and resilience, PWB is crucial for university students (Mustafa et al., 2020). Higher PWB not only leads to better academic motivation, emotional regulation, and adaptive coping strategies (Greenier et al., 2021; Li and Hasson, 2020; Tan et al., 2021), but it also strengthens social connections, enabling students to form and maintain meaningful relationships (Tan et al., 2021). Moreover, the COVID-19 pandemic further highlighted PWB’s importance, as students with higher PWB exhibited greater academic persistence and emotional stability during this time of uncertainty (Sun et al., 2020; Morales-Rodríguez et al., 2020).

Several factors predict PWB in university students, including resilience, perceived social support, emotional intelligence, and self-efficacy. Resilience, or the ability to recover from adversity, is critical in helping students manage the emotional and academic stress they encounter (Johnson et al., 2019; Tan et al., 2021). Furthermore, social support from peers, family, and educators enhances emotional well-being by providing a buffer against depressive symptoms and fostering a sense of belonging (Costa et al., 2013; Rueger et al., 2016). Emotional intelligence, which involves the ability to effectively manage emotions, also contributes to PWB by strengthening interpersonal relationships and supporting emotional regulation (Shengyao et al., 2024).

Self-efficacy, or the belief in one’s ability to succeed, is another key predictor of PWB. Students with high self-efficacy tend to be more persistent in the face of academic challenges and exhibit better psychological adjustment to stress (Costa et al., 2013). This relationship was particularly evident during the COVID-19 pandemic, as students with greater self-efficacy, resilience, and social support were better able to maintain their PWB despite the disruptions to their academic and social lives (Marler et al., 2024). Consequently, educational environments that actively foster self-efficacy play a crucial role in supporting students’ well-being and academic success (Tang and Zhu, 2024).

In addition to these individual factors, environmental elements, such as academic stress, financial pressures, and the overall campus environment, can significantly affect students’ PWB (Mazzucchelli and Purcell, 2015). High levels of academic stress, particularly when combined with insufficient institutional support, have been linked to lower well-being and increased psychological distress (Morales-Rodríguez et al., 2020). Conversely, universities that provide accessible mental health resources and create inclusive, supportive environments can enhance students’ well-being, leading to improved academic performance and emotional stability (Hernández-Torrano et al., 2020; Huppert, 2009). These findings highlight the important role that institutions play in shaping students’ well-being by addressing both their academic and emotional needs.

In conclusion, PWB is a multifaceted construct crucial to the academic and personal success of university students, affecting not only academic outcomes but also broader aspects of their social and emotional functioning. By focusing on key predictors such as resilience, social support, emotional intelligence, and self-efficacy, educational institutions can develop strategies to enhance students’ well-being and help them navigate the complexities of higher education. Ultimately, the role of institutions in creating supportive environments that address both academic and emotional needs is essential for promoting long-term PWB and academic achievement.

Gratitude, a core concept in positive psychology, refers to the recognition and appreciation of the positive aspects of life and the contributions of others (Emmons and McCullough, 2003; Gulliford and Morgan, 2017). It can be experienced both as a dispositional trait and a momentary emotional state, each of which enhances well-being by fostering positive affect and strengthening social connections (Emmons and Shelton, 2001). According to Fredrickson’s (2004) broaden-and-build theory, gratitude expands individuals’ cognitive and behavioral repertoires, building lasting personal resources. This approach aligns with the eudaimonic view of well-being, which emphasizes personal growth and meaningful social engagement over short-term pleasure (Wood et al., 2009). Gratitude not only predicts life satisfaction but also enhances psychological well-being (PWB) beyond the influence of other personality traits like the Big Five (Watkins, 2013; Wood et al., 2009).

Gratitude is increasingly recognized as a critical contributor to life satisfaction, happiness, and PWB (Datu, 2014). Ryff and Keyes (1995) conceptualization of PWB includes dimensions such as autonomy, personal growth, and positive relationships, all of which are positively influenced by gratitude. By promoting positive affect, gratitude supports a more optimistic and growth-oriented perspective, which reinforces these dimensions of PWB. Studies consistently show that gratitude is linked to greater resilience, improved emotional regulation, and better coping with stress (Demichelis et al., 2024; Portocarrero et al., 2020). In educational settings, gratitude plays a particularly important role in buffering against stress, thereby enhancing PWB among students (Deichert et al., 2021). Bono et al. (2022) highlight that students who cultivate gratitude experience greater academic success, stronger social relationships, and better emotional health.

The importance of gratitude in well-being became even more evident during crises like the COVID-19 pandemic. Research revealed that students who practiced gratitude were better able to handle the stress of isolation, academic disruptions, and uncertainty (Tolcher et al., 2024). Meta-analyses confirm that dispositional gratitude correlates with higher psychological well-being, greater life satisfaction, and reduced depressive symptoms across diverse populations (Portocarrero et al., 2020). Mason’s (2019) longitudinal study of South African students showed that gratitude acted as a buffer against anxiety and depression, further illustrating its universal benefits.

Gratitude interventions, such as journaling and reflective practices, have shown significant improvements in mental health. Tolcher et al. (2024) found that these interventions led to sustained increases in well-being and reductions in stress by shifting attention from negative to positive experiences, fostering psychological resilience (Dickens, 2017; Wood et al., 2009). Similarly, Zhang et al. (2023) found that gratitude practices increased well-being and reduced emotional exhaustion in Chinese young adults. Several mechanisms explain this link. Gratitude encourages positive cognitive reappraisal, reinterpreting negative events by focusing on benefits or lessons learned, enhancing emotional regulation and stress management (Fredrickson, 2004; Deichert et al., 2021). It also fosters prosocial behavior by increasing empathy and a sense of obligation towards others, strengthening social bonds (Emmons and McCullough, 2003; Kardas et al., 2019). This creates a reinforcing cycle of well-being (Fredrickson, 2004). Gratitude promotes emotional resilience by encouraging a focus on positive aspects, even in difficult situations (Arnout and Almoied, 2021). Huang et al. (2020) found that gratitude moderated the relationship between academic stress and psychological health in students. Finally, gratitude fosters an internal locus of control, enhancing self-efficacy and supporting goal achievement (Portocarrero et al., 2020).

Taken together, gratitude is a powerful resource for enhancing psychological well-being. Its positive effects are mediated through cognitive reappraisal, prosocial behavior, emotional resilience, and an internal locus of control. These findings highlight the importance of incorporating gratitude into mental health interventions and educational programs to improve student well-being across different contexts and populations.

Perceived social support—an individual’s belief in the availability of help from their social network (Procidano and Smith, 1997; Zimet et al., 1988)—differs from objective support, which focuses on actual help received. Perceived support often has a stronger impact on psychological outcomes (Zimet et al., 1988). The Multidimensional Scale of Perceived Social Support (MSPSS) (Zimet et al., 1988) categorizes support into three sources: family, friends, and significant others, each offering distinct types of assistance. The belief in their availability shapes an individual’s well-being (Tonsing et al., 2012).

Perceived social support’s benefits are well-established. Research shows that perceiving supportive relationships significantly improves psychological outcomes, often exceeding the effects of actual support (Feeney and Collins, 2015). Individuals with high levels of perceived support report greater life satisfaction, more positive affect, and lower levels of depression and anxiety (Malkoç and Yalçın, 2015). This is particularly relevant for students facing academic stress. Cohen and Wills’ (1985) buffering hypothesis suggests that social support mitigates stress’s negative effects by providing emotional and practical resources.

In educational settings, perceived social support plays a critical role in enhancing emotional regulation and psychological resilience (Hou et al., 2024). Students who feel supported by their social networks are more likely to experience personal growth, autonomy, and environmental mastery, all of which contribute to well-being (Derakhshan et al., 2024; Malkoç and Yalçın, 2015). Furthermore, social support fosters the development of effective coping strategies, allowing students to better navigate academic and interpersonal challenges (Wang et al., 2023). This is particularly significant for university students transitioning to new academic and social environments, where feelings of isolation can negatively affect their well-being (Tomás et al., 2020).

Empirical studies confirm the link between perceived social support and psychological resilience. Malkoç and Yalçın (2015) found that higher perceived support predicted greater resilience and improved well-being in students. Emadpoor et al. (2016) showed that academic motivation mediated the relationship between social support and well-being, suggesting support networks enhance intrinsic motivation. This is particularly critical for international students facing cultural adjustment challenges. Brunsting et al. (2021) found that higher perceived support in international students led to greater emotional well-being and reduced homesickness.

Several theoretical models explain how social support influences well-being. The stress-buffering model proposes that social support mitigates stress by providing resources (Cohen and Wills, 1985). This is relevant for students facing academic pressures, as support helps them manage stress (Lai and Ma, 2016). The main effects model suggests a direct influence of social support on well-being, independent of stress (Lee and Goldstein, 2016). This model emphasizes the intrinsic value of social relationships for fostering belonging and self-worth (Li and Jiang, 2018). In university students, support from peers, family, and faculty enhances competence and social integration (Rehman et al., 2020). Bandura’s (1986) social cognitive theory posits that perceived support enhances self-efficacy, promoting emotional resilience and well-being (Guzman Villegas-Frei et al., 2024; Yildirim et al., 2023).

In summary, perceived social support is a vital determinant of psychological well-being, particularly in student populations. It acts as a protective factor by buffering against stress, enhancing emotional stability, and fostering a sense of belonging. Theoretical frameworks such as the stress-buffering model, the main effects model, and social cognitive theory provide valuable insights into the mechanisms through which social support positively impacts PWB. Understanding these pathways highlights the importance of fostering supportive social environments in educational settings to promote well-being and academic success.

Prosocial tendencies encompass behaviors like helping, sharing, and cooperating, driven by intrinsic motivation, empathy, and social norms (Carlo and Randall, 2002; Guan et al., 2019). These actions, initially conceptualized as selfless and other-oriented (Eisenberg, 1989), include various forms serving distinct social and psychological functions (Carlo and Randall, 2002; Eberly-Lewis and Coetzee, 2015). The Prosocial Tendencies Measure (PTM) captures this complexity by assessing a range of prosocial behaviors, from public altruism to private, self-sacrificial acts (Carlo and Randall, 2002). A strong link exists between prosocial behavior and psychological well-being. Engaging in these behaviors generates positive emotions, strengthens social bonds, and enhances a sense of purpose and self-worth (Helliwell et al., 2018; Martela and Ryan, 2016), creating a positive feedback loop. Fredrickson’s broaden-and-build theory (2004) explains how positive emotions from prosocial acts expand individuals’ thought-action repertoires, leading to the development of personal and social resources that contribute to higher life satisfaction and reduced stress (Feng and Zhang, 2021).

Empathy is central to driving prosocial behavior by enabling individuals to emotionally connect with others’ needs (Decety et al., 2016). Individuals with higher levels of empathy are more likely to help others, which strengthens their social bonds and emotional well-being (Telle and Pfister, 2016). Additionally, emotional regulation enhances prosocial behavior by allowing individuals to manage their emotions effectively, helping them respond constructively to others’ needs (Bailey et al., 2020; Sharma and Tomer, 2018). Positive emotions, such as gratitude and compassion, further encourage prosocial behavior by fostering a sense of interconnectedness and moral responsibility, benefiting both the helper and the recipient (Aknin et al., 2018).

Social support also plays a key role in fostering prosocial behavior. Individuals who perceive strong social support are more likely to engage in helping behaviors, as supportive relationships provide a sense of security, belonging, and reciprocity (Fu et al., 2022; Rosli and Perveen, 2021). This reciprocal relationship strengthens social networks, enhancing emotional well-being for both the giver and the receiver (Haller et al., 2022). In both educational and organizational settings, fostering positive social environments is essential, as strong interpersonal relationships promote collective well-being and encourage prosocial tendencies (Collie, 2022; Martela and Ryan, 2016).

The development of prosocial tendencies is significantly influenced by environmental factors, particularly during childhood and adolescence (Knafo and Plomin, 2006). Supportive and emotionally nurturing environments in schools and families encourage prosocial behavior in young people (Su et al., 2021). When students’ needs for relatedness are met, particularly in educational settings, their motivation to engage in prosocial acts increases, promoting psychological well-being (Collie, 2022; Liu et al., 2021). These findings highlight the importance of social–emotional learning initiatives that cultivate empathy, emotional intelligence, and social responsibility, fostering personal growth and encouraging positive contributions to others (Martela and Ryan, 2016).

Overall, prosocial tendencies are closely linked to psychological well-being. Empathy, emotional regulation, and social support are key drivers of prosocial behavior, enhancing individuals’ sense of purpose, belonging, and emotional resilience. Given the reciprocal relationship between prosocial behavior and well-being, promoting these behaviors not only benefits those on the receiving end but also enhances the mental health of those who engage in them. Fostering prosocial behavior is therefore essential for improving individual well-being and strengthening social cohesion.

This study draws on key theoretical frameworks to explain the relationships between gratitude, social support, prosocial behavior, and psychological well-being (PWB). Fredrickson’s broaden-and-build theory (2004) posits that positive emotions, such as gratitude, expand cognitive and behavioral capacities, fostering prosocial behaviors that enhance social bonds and improve PWB (Fredrickson, 2004; Huppert, 2009; Wood et al., 2009). Gratitude not only boosts emotional states but also promotes personal growth and social capital through altruism and cooperation (Emmons and McCullough, 2003). Similarly, Bandura’s social cognitive theory (1986) highlights the role of self-efficacy in encouraging individuals with strong social support to engage in prosocial behaviors, enhancing competence and belonging (Rehman et al., 2020; Schwarzer and Knoll, 2007).

The stress-buffering model (Cohen and Wills, 1985) emphasizes that social support reduces stress, promoting resilience and mental health (Wang et al., 2018). Together, these frameworks suggest that gratitude and social support improve PWB by encouraging prosocial actions, which build resilience, strengthen social connections, and foster purpose (Helliwell et al., 2018; Martela and Ryan, 2016).

Empirical evidence supports these links. Gratitude predicts PWB beyond personality traits and strengthens social bonds, enhancing perceived social support and facilitating prosocial behavior (Portocarrero et al., 2020; Feng and Zhang, 2021; Deichert et al., 2021). Social support, both perceived and actual, plays a vital role in reducing stress, particularly for students managing academic pressures (Brunsting et al., 2021; Costa et al., 2013). Students with greater social support report improved emotional well-being, reduced homesickness, and better psychological health (Li et al., 2023). Prosocial behaviors, such as helping others, foster relatedness, competence, and resilience, directly contributing to PWB (Huppert, 2009; Wan et al., 2023). Gratitude and social support promote prosociality by fostering empathy and reciprocity, which strengthen social ties and well-being (Emmons and McCullough, 2003; Su et al., 2021). Prosocial actions also help individuals build psychological resources like emotional regulation, buffering against stress and enhancing mental health (Haller et al., 2022; Zhang et al., 2023). Research demonstrates that prosocial behavior mediates the relationship between social support and mental health, as well as between gratitude and PWB (Li et al., 2023). This highlights the importance of prosociality in translating gratitude and social support into psychological benefits (Deichert et al., 2021; Zhang et al., 2023).

To address these gaps, this study seeks to answer the following research question: How do gratitude and perceived social support, both directly and indirectly via prosocial tendencies, influence psychological well-being among Chinese university students?

The primary objective of this study is to examine the mediating role of prosocial behavior in the relationships between gratitude, perceived social support, and psychological well-being. Specifically, the research aims to identify the mechanisms through which gratitude and social support contribute to psychological health and to test whether these mechanisms are consistent across gender groups.

Based on these theoretical frameworks and empirical evidence, the following hypotheses are proposed:

A total of 703 university students from various academic disciplines in China participated in this study. University students represent a particularly relevant population for this study for several reasons. First, they are at a critical developmental stage marked by significant transitions and challenges that can impact their mental health and well-being. Second, university students are often exposed to high levels of academic stress and social pressures, making them particularly vulnerable to psychological distress (Li and Hasson, 2020). Third, this population is increasingly reliant on social connections and support networks to navigate these challenges, making it crucial to understand the factors that contribute to their psychological well-being in a university context. Furthermore, examining these relationships within a Chinese context provides valuable insights into how gratitude, social support, and prosocial tendencies operate within a collectivist cultural framework.

The participants were recruited over a three-month period, from March to May 2023, using a combination of online and in-person strategies to ensure a diverse and representative sample. The participants ranged in age from 18 to 25 years (M = 21.3, SD = 1.7), with 56% identifying as female (n = 394) and 44% as male (n = 309). The majority of participants were undergraduate students (83%), while the remaining 17% were enrolled in graduate programs.

The sample included participants from various socioeconomic backgrounds, ranging from low-income families to those with higher household incomes. Based on self-reported parental education levels, approximately 35% of participants came from families where neither parent held a college degree, while 42% reported at least one parent with higher education, and 23% indicated both parents had advanced degrees. This distribution highlights the socioeconomic diversity within the sample. Geographically, students were drawn from 15 provinces across China, with a majority (68%) coming from urban areas and 32% from rural regions. This geographic spread ensures representation from diverse cultural and environmental contexts, capturing differences in regional access to educational resources and social support networks. Additionally, participants represented a wide array of academic disciplines, including humanities (26%), sciences (22%), engineering (18%), business (16%), and medical fields (18%). This variety provides a comprehensive view of the student population, avoiding overrepresentation of any single academic focus.

The participants were required to be full-time university students in China. Exclusion criteria included part-time student status or a history of psychiatric disorders that could affect responses to psychological well-being measures. The final sample represented a diverse range of socioeconomic and academic backgrounds, providing a broad cross-section of the Chinese student population. Ethical approval was obtained from the Department of Basic Education at Jiangsu Medical College. Participants were informed about the study’s purpose, assured of voluntary participation, and provided with informed consent forms online before beginning the survey. Anonymity was guaranteed, and participants could withdraw at any time without consequence.

To enhance generalizability, students were recruited from multiple academic disciplines, including humanities, sciences, engineering, and medical fields. Participation was voluntary, with no monetary incentives offered. Contact information for psychological support services was provided to address any emotional discomfort arising from the survey.

The disposition to experience and express gratitude was measured using the Gratitude Questionnaire-6 (GQ-6), developed by McCullough et al. (2002). This six-item scale assesses individual differences in gratitude proneness. Participants responded to items like “I have a lot to be thankful for in my life” on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). A higher average score indicates a greater tendency towards gratitude. The GQ-6 was translated into Chinese following a standardized translation and back-translation protocol. Bilingual experts reviewed the translated version to ensure semantic and conceptual equivalence with the original English version.

In the present study, the GQ-6 exhibited satisfactory internal consistency, with a Cronbach’s alpha of 0.83 for the Chinese sample. To further examine the reliability of the GQ-6, we calculated McDonald’s omega (ω), which was found to be 0.84. Confirmatory factor analysis (CFA) for the GQ-6 demonstrated good construct validity (χ²/df = 2.45, CFI = 0.95, TLI = 0.93, RMSEA = 0.06 [0.04, 0.08], SRMR = 0.04), indicating that the items loaded well onto a single factor of gratitude. Convergent validity was assessed by examining the average variance extracted (AVE), which was 0.60, suggesting that the items explain a substantial proportion of the variance in the gratitude construct. Discriminant validity was assessed using the Fornell-Larcker criterion, which compares the AVE for each construct to the squared correlations between constructs. The results indicated that the AVE for gratitude was greater than its squared correlations with other constructs in the model (see Table 1), supporting discriminant validity.

The perception of social support received from various sources was assessed using the Multidimensional Scale of Perceived Social Support (MSPSS), created by Zimet et al. (1988). This 12-item instrument evaluates perceived support from family, friends, and significant others. Participants rated statements like “I get the emotional help and support I need from my family” on a 6-point Likert scale (1 = strongly disagree to 6 = strongly agree). A higher average score signifies stronger perceived social support. The MSPSS was translated into Chinese using a standardized translation and back-translation procedure, followed by a review by bilingual experts to ensure the semantic and conceptual equivalence of the translated items.

To assess the internal consistency of the MSPSS, we calculated Cronbach’s alpha, which was found to be 0.87, indicating excellent reliability. Additionally, McDonald’s omega was calculated for each subscale of the MSPSS, with values ranging from 0.81 to 0.86, further supporting the internal consistency of the measure. CFA results for the MSPSS indicated a good fit (χ²/df = 3.10, CFI = 0.93, TLI = 0.91, RMSEA = 0.07 [0.05, 0.09], SRMR = 0.05), suggesting that the three-factor structure (family, friends, significant others) is a good representation of the data. Convergent validity was supported with an AVE of 0.63, indicating that the items explain a substantial amount of variance in the perceived social support construct. Discriminant validity checks showed that the AVE for perceived social support exceeded the squared correlations with other constructs in the model (see Table 1).

The inclination to engage in helpful behaviors towards others was measured with the Prosocial Tendencies Measure (PTM), developed by Carlo and Randall (2002). This 26-item scale comprises six subscales: public, anonymous, altruistic, compliant, emotional, and urgent, each capturing a different facet of prosocial behavior. Respondents indicated their agreement with items such as “I feel that helping others is a way to gain people’s approval” on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The PTM was adapted into Chinese using a standardized translation and back-translation protocol. A committee of bilingual experts in psychology reviewed the translated version to ensure that the meaning and nuances of each item were accurately captured in Chinese.

The PTM exhibited good internal consistency in the current study, with a Cronbach’s alpha of 0.81. Internal consistency for each subscale was assessed, with Cronbach’s alpha values ranging from 0.72 to 0.84, indicating acceptable to good reliability for all dimensions. Specifically, the alpha values were as follows: altruistic (0.84), compliant (0.77), emotional (0.79), public (0.72), anonymous (0.80), and urgent (0.81). McDonald’s omega (ω) was also calculated for each subscale, with values all exceeding 0.75, providing further evidence of good reliability. Confirmatory factor analysis for the PTM subscales showed acceptable construct validity (χ²/df = 2.80, CFI = 0.91, TLI = 0.89, RMSEA = 0.07 [0.06, 0.09], SRMR = 0.05). All factor loadings were statistically significant (p < 0.001), and the model fit indices support the six-factor structure of the PTM. In our study, prosocial tendencies were modeled as a higher-order latent construct to capture the shared variance among these dimensions while maintaining parsimony and interpretability in the structural equation model. This higher-order factor showed excellent internal consistency, with ω = 0.88. Convergent validity was indicated by AVEs above 0.50 for all subscales, suggesting that the items within each subscale converge to measure a specific dimension of prosocial tendencies. Discriminant validity was confirmed using the Fornell-Larcker approach, which showed that the AVE for each subscale was greater than the squared correlations with other subscales (see Table 1).

The overall state of psychological health was evaluated using the 18-item version of the Ryff Psychological Well-being Scale (Ryff and Keyes, 1995). This shortened version of the scale encompasses six dimensions of well-being: autonomy, environmental mastery, personal growth, positive relations with others, purpose in life, and self-acceptance. Participants rated statements like “I feel I am in charge of the situation in which I live” on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). A higher mean score across the six dimensions reflects greater psychological well-being. The Ryff scale was translated into Chinese using a rigorous translation and back-translation process, with independent translations by two bilingual individuals. Any discrepancies were resolved through discussion and consensus, and the final translated version was reviewed by a panel of experts to ensure accuracy and cultural appropriateness.

The scale showed high internal consistency in this study, with a Cronbach’s alpha of 0.88. Furthermore, McDonald’s omega for the overall scale was 0.89, and each dimension showed ω values above 0.80, indicating excellent internal consistency. CFA for the Ryff Psychological Well-being Scale confirmed construct validity (χ²/df = 3.00, CFI = 0.92, TLI = 0.90, RMSEA = 0.07 [0.06, 0.09], SRMR = 0.05). All factor loadings were statistically significant (p < 0.001), and the model fit indices support the six-factor structure of the Ryff scale. Construct validity was further supported by AVE = 0.64, suggesting that the items explain a substantial amount of variance in the psychological well-being construct. Discriminant validity testing indicated that each dimension captured a distinct facet of well-being, as the AVE for each dimension was greater than the squared correlations with other dimensions (see Table 1).

To accommodate participant preferences and enhance engagement, data collection was conducted from March to May 2023 using both online and in-person methods. Recruitment was facilitated through a variety of channels, including university-wide emails, digital flyers, social media announcements, and outreach to academic departments. Participants had the option to complete the survey online via a secure Qualtrics link or in person at designated university locations. In-person surveys were conducted in computer labs under the supervision of research assistants, and both methods used identical survey content and presentation to ensure consistency.

Before starting the survey, participants received an overview of the study’s objectives and procedures. Informed consent was obtained electronically for online participants and via signed forms for in-person participants. To ensure voluntary participation, confidentiality was emphasized, with no identifying information collected, and participants were informed that they could withdraw at any time without penalty. No compensation was offered.

To minimize order effects, survey items were randomized. Standardized questionnaires were used to measure gratitude, perceived social support, prosocial tendencies, and psychological well-being. The survey took approximately 15–20 min to complete, with an option to save progress and return later. Participant well-being was prioritized throughout the data collection process. Upon completing the survey, participants were debriefed and given detailed information about the study, including contact details for the research team. Furthermore, participants who experienced any discomfort were informed of available university counseling services and external mental health resources, and research assistants were trained to address immediate concerns during in-person sessions.

This study was carried out by two principal investigators, one specializing in educational psychology and the other focusing on quantitative data analysis. They collaborated closely in designing the research framework, developing the survey instruments, and interpreting the results. A team of four trained research assistants supported data collection and participant recruitment under the investigators’ supervision. The principal investigators held weekly meetings to discuss recruitment progress, address logistical challenges, and review preliminary findings. This regular communication ensured that all team members contributed to refining the research design and maintaining data quality. After data collection, both investigators jointly analyzed the results and worked together on writing and revising the manuscript, integrating feedback from all team members to ensure accuracy and coherence.

Data analysis was conducted using structural equation modeling (SEM) with AMOS 26.0, complemented by assumption testing and data screening in SPSS 26.0. First, to ensure the quality of the data, we conducted preliminary analyses. This included assessing normality through skewness and kurtosis values, multicollinearity via Variance Inflation Factor (VIF) values (Hair et al., 2010), and homoscedasticity by inspecting residual scatter plots. Outliers were identified using univariate and multivariate methods and were excluded to ensure the robustness of the findings.

Next, we performed a confirmatory factor analysis (CFA) to assess the construct validity of our measurement model. This involved examining the fit indices including Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR), following Hu and Bentler (1999) criteria. Good model fit was defined by CFI and TLI values above 0.90, RMSEA below 0.08, and SRMR below 0.08 (Kline, 2016).

Following the CFA, we evaluated the reliability and convergent validity of the measures. We assessed reliability using both Cronbach’s alpha and McDonald’s omega, with values above 0.70 for Cronbach’s alpha and 0.80 for McDonald’s omega considered to indicate good internal consistency. Convergent validity was evaluated by examining the factor loadings of the items on their corresponding latent variables, with high loadings indicating that the items are measuring the same construct. We also calculated the average variance extracted (AVE) for each construct, with AVE values above 0.50 generally considered to indicate good convergent validity (Fornell and Larcker, 1981).

Subsequently, we assessed discriminant validity using the Fornell-Larcker criterion, which compares the AVE for each construct to the squared correlations between constructs. Discriminant validity is supported when the AVE for each construct is greater than the squared correlations with other constructs (Henseler et al., 2015). The heterotrait-monotrait (HTMT) ratio of correlations was also examined as a further test of discriminant validity.

Finally, we conducted structural equation modeling (SEM) to test the hypothesized relationships between the latent variables. Model fit was evaluated using the same fit indices as in the CFA. Hypothesized mediation effects were tested using bootstrapping with 5,000 resamples to generate bias-corrected confidence intervals (Preacher and Hayes, 2008). To further explore the potential moderating effect of gender, we conducted a multi-group SEM analysis. This involved a series of hierarchical invariance tests (configural, metric, scalar, and structural) to determine whether the relationships between the variables differed across genders (Byrne, 2016; Cheung and Rensvold, 2002).

Prior to analysis, data were screened for missing values, outliers, and normality assumptions. Missing data (approximately 3%) were handled using the expectation–maximization (EM) algorithm, which provides unbiased parameter estimates. Little’s MCAR test (χ²(32) = 35.21, p = 0.29) confirmed that data were missing at random, validating the use of EM for imputation (Little and Rubin, 2019). Outliers were identified through Z-scores (±3.29 standard deviations) for univariate outliers and Mahalanobis distance (p < 0.001) for multivariate outliers (Kline, 2016). This process flagged 12 univariate and 8 multivariate outliers. Sensitivity analyses showed minimal impact on results; however, these outliers were excluded to ensure robustness. Normality assumptions were evaluated by examining skewness (−0.25 to 0.35) and kurtosis (−0.45 to 0.82), which indicated univariate normality. Mardia’s coefficient (2.87) confirmed multivariate normality (Byrne, 2016).

To address potential common method bias (CMB) from self-reported measures, several procedural safeguards were implemented, including assurances of anonymity, clear language, and randomized item order to minimize response patterns. Harman’s single-factor test revealed that the first factor accounted for 24% of the variance, below the 50% threshold, indicating minimal CMB (Podsakoff et al., 2003). Additionally, a marker variable unrelated to the main constructs showed no significant correlation (r = 0.03), further supporting the absence of meaningful CMB.

Table 2 presents the descriptive statistics for all key study variables, including gratitude, perceived social support, the six dimensions of prosocial tendencies (altruistic, compliant, emotional, public, anonymous, and urgent), and psychological well-being. The prosocial tendencies construct was modeled as a higher-order factor (HOF) in subsequent analyses, reflecting the shared variance among its six dimensions. Descriptive statistics for these dimensions are provided individually in Table 2. A composite score representing the higher-order factor “prosocial tendencies” was calculated by averaging the z-scores of the six dimensions.

The mean score for gratitude was 5.30 (SD = 0.95), indicating a relatively high level of gratitude among participants, with scores ranging from 2.0 to 7.0, indicating that most participants reported moderate to high levels of gratitude. Perceived social support had a mean score of 4.85 (SD = 0.78), reflecting a moderate to high level of perceived support. Scores on the perceived social support measure ranged from 2.5 to 6.5. For the dimensions of prosocial tendencies, mean scores varied slightly, with altruistic prosocial tendency showing the highest mean (M = 3.85, SD = 0.90), and public prosocial tendency the lowest (M = 3.25, SD = 0.85). Finally, psychological well-being had a mean score of 5.15 (SD = 1.00), reflecting relatively high levels of well-being. Scores on the psychological well-being measure ranged from 2.1 to 6.9 (see Table 2). The skewness and kurtosis values for all variables, including the dimensions of prosocial tendencies, fell within acceptable ranges (skewness: −0.30 to 0.40; kurtosis: −0.50 to 0.85), indicating approximate normality.

Table 3 presents the correlations among the study variables, including the six dimensions of prosocial tendencies and the higher-order factor (HOF) modeled in the structural equation model. Gratitude was positively correlated with perceived social support (r = 0.46, p < 0.001) and psychological well-being (r = 0.48, p < 0.001). Perceived social support showed a strong positive correlation with psychological well-being (r = 0.43, p < 0.001) and a significant relationship with the prosocial tendencies HOF (r = 0.55, p < 0.001).

The six dimensions of prosocial tendencies exhibited varying levels of correlation with psychological well-being. Altruistic (r = 0.47, p < 0.001) and emotional (r = 0.44, p < 0.001) tendencies demonstrated the strongest positive associations, while public prosocial tendency showed the weakest positive association (r = 0.33, p < 0.001). The prosocial tendencies HOF showed moderate to strong correlations with its constituent dimensions, including altruistic (r = 0.71, p < 0.001), emotional (r = 0.72, p < 0.001), and anonymous (r = 0.68, p < 0.001) prosocial tendencies. Overall, the higher-order factor was moderately correlated with psychological well-being (r = 0.42, p < 0.001).

A confirmatory factor analysis (CFA) was conducted to validate the measurement model, which included the latent variables of gratitude, perceived social support, a higher-order factor for prosocial tendencies, and psychological well-being. The six dimensions of prosocial tendencies (public, anonymous, altruistic, compliant, emotional, and urgent) were specified as indicators of the higher-order factor. This model demonstrated a good fit to the data: χ²(180) = 520.35, p < 0.001, CFI = 0.93, TLI = 0.92, RMSEA = 0.06 (90% CI [0.05, 0.07]), SRMR = 0.05. The fit indices indicate that the hypothesized model is a good representation of the data. All factor loadings were statistically significant (p < 0.001). Specifically, the factor loadings of the prosocial tendencies dimensions on the higher-order factor were as follows: Public (0.71–0.81), Anonymous (0.73–0.85), Altruistic (0.66–0.82), Compliant (0.69–0.79), Emotional (0.65–0.79), and Urgent (0.69–0.84). These CFA loadings confirm that all six dimensions contribute meaningfully to the higher-order construct of prosocial tendencies.

This CFA model with a higher-order factor for prosocial tendencies was chosen over a model with separate latent variables for each dimension due to its greater parsimony and conceptual clarity. While both models demonstrated acceptable fit, the higher-order factor model provides a more concise and interpretable representation of the data by capturing the overall construct of prosocial tendencies while still accounting for the unique contributions of its dimensions.

During the CFA process, several diagnostic checks were performed to assess the potential for model improvement. No modifications were required, as the CFA loadings for all items exceeded the recommended threshold of 0.60 (Kline, 2016), and theoretical considerations supported retaining all items. There was no need to correlate error terms or remove any items, as they aligned well with the constructs measured.

In addition to the prosocial tendencies dimensions, the other latent variables also showed strong relationships with their respective indicators. Specifically, the CFA loadings for gratitude items ranged between 0.70 and 0.82, perceived social support items between 0.68 and 0.83, and psychological well-being items between 0.75 and 0.88. These results confirm the adequacy of the hypothesized factor structure and the strong alignment of items with their respective latent constructs.

To assess the reliability and validity of the measurement model, we examined composite reliability, AVE, and discriminant validity. Composite reliability (CR) estimates exceeded the recommended threshold of 0.70, indicating good internal consistency for all constructs (Fornell and Larcker, 1981). The CR values were 0.83 for gratitude, 0.85 for perceived social support, 0.91 for the higher-order factor prosocial tendencies, and 0.89 for psychological well-being (see Table 4). Additionally, the average variance extracted (AVE) for each construct was greater than 0.50, with AVE values of 0.60 for gratitude, 0.63 for perceived social support, 0.65 for “Prosocial Tendencies,” and 0.64 for psychological well-being, confirming that a substantial amount of variance in the indicators was explained by the latent constructs.

Discriminant validity was assessed by comparing the AVE values to the squared correlations between constructs, following the Fornell-Larcker criterion. In all cases, the AVE for each construct was higher than the squared correlations, supporting discriminant validity (see Table 1). For further confirmation, the heterotrait-monotrait (HTMT) ratios were also examined. All HTMT values were below the conservative threshold of 0.85 (Henseler et al., 2015), providing additional support for discriminant validity.