Identifying Core and Bridge Nodes of Psychological Capital and Social-Emotional Competence Among Medical Students: A Network Analysis Study

Abstract

Purpose

Most existing studies adopt traditional regression models to examine either psychological capital or social‑emotional competence in isolation. Few integrate these two sets of multi‑dimensional positive psychological resources within a unified network framework, hindering the exploration of their complex patterns of covariation. This study therefore aimed to identify core and bridge nodes of psychological capital and social‑emotional competence among medical students using network analysis, so as to provide potential targets for positive‑psychological interventions.

Methods

A cross-sectional questionnaire survey was conducted among 1,217 undergraduate medical students. After adjusting for confounding factors, a regularized partial correlation network was constructed. Strength centrality and bridge strength were calculated, and network robustness tests were performed.

Results

1. The various psychological dimensions were closely interrelated, and the network stability met the criteria. 2. Self-efficacy exhibited the highest strength centrality and served as the network’s core. 3. Self-relationship exhibited the highest bridge strength, serving as a core mediating node between the two systems. Hope, responsible decision-making, and group relationships also exerted significant influence.

Conclusion

Close concurrent covariation exists between the two positive-psychological systems of psychological capital and social-emotional competence among medical students Targeted interventions focused on core variables such as self-efficacy and self-relationship competence can effectively improve medical students’ psychological well-being and optimize programs for cultivating the psychological qualities of medical professionals.

Introduction

The healthcare industry is characterized by both high stress and heavy workloads. The development of the healthcare sector, changes in doctor-patient relationships, and public health challenges have placed higher demands on the comprehensive competencies of healthcare professionals. Medical students represent the future workforce of the healthcare sector. Their psychological resilience and professional adaptability not only influence their long-term personal development but also impact the quality of healthcare services and the stability of the healthcare workforce. Due to multiple stressors—including extended program duration, demanding coursework, clinical practice, and professional ethics—medical students are under chronic psychological strain. The prevalence of psychological issues such as burnout, anxiety, and depression is significantly higher among medical students than in the general population.^1,2 They are also prone to issues such as academic burnout and a tendency to drop out.³ Consequently, enhancing medical students’ psychological resilience and safeguarding their mental health has become a key research focus in the field of medical education.

Social-emotional competence refers to an individual’s capacity to regulate one’s own emotions, recognize the emotions of others, cooperate and build positive interpersonal relationships, and resolve social problems during social interactions.⁴ It comprises four dimensions: self-relationship, other-relationship, group-relationship, and responsible decision-making,⁵ which run through medical students’ academic training, clinical practice, and professional development. In recent years, global medical-education reforms have increasingly emphasized social-emotional learning (SEL) as an essential component of physician competency. Capacities including emotion regulation, teamwork, ethical decision-making and empathy are regarded as core professional attributes for future clinicians, which are closely linked to doctor-patient communication, interprofessional collaboration, and human-centered clinical care.^6,7 For medical education, fostering social-emotional competence enables medical students to cope with academic stress and interpersonal conflicts and prevents the accumulation of adverse emotions; it also supports empathic patient-provider interactions in clinical contexts and constitutes an indispensable component of medical students’ professional literacy.⁶ Systematic-review evidence has demonstrated that structured training targeting social-emotional competencies can improve stress-management skills and subjective well-being among medical students, although effect heterogeneity across intervention programs has been noted.^8,9 Despite its well-recognized educational importance, existing empirical studies on medical students have certain limitations. Most prior work has examined unidirectional predictive effects of social-emotional competence on mental-health or academic outcomes mainly through aggregate-score regression approaches. Few studies have adopted a systemic perspective to unpack inter-dimensional dynamics within this construct. The interplay between social-emotional competence and other positive-psychological resources remains under-explored, and empirical evidence tailored to domestic medical-education contexts is still insufficient.

Psychological capital is a set of positive psychological resources centered on self-efficacy, hope, resilience, and optimism. It focuses on tapping into an individual’s inner positive psychological potential to help them withstand external stressors, alleviate psychological exhaustion, and enhance their ability to cope with stress.¹⁰ Existing research also confirms that psychological capital not only directly and positively predicts medical students’ mental health levels but also mediates the relationship between variables such as interpersonal relationships and mindfulness literacy and psychological state, serving as a core element in maintaining psychological homeostasis among medical students.^11–13

Traditional regression-based statistical methods often rely on the underlying assumptions of variable independence and unidirectional causality, making it difficult to capture the complex interrelationships among multidimensional psychological traits. The network analysis paradigm effectively addresses this research shortcoming. This theoretical paradigm breaks down abstract psychological systems into multiple visualizable nodes and constructs a network of node connections using regularized partial correlation coefficients. By using metrics such as centrality and bridge strength to identify core variables and key pathways within psychological systems, this approach better aligns with the nonlinear characteristics of complex psychological systems. This method has now become a mainstream research approach in the fields of positive psychology and clinical psychology.^14,15 At present, this method has been widely applied in research on medical students’ mental health. Existing studies have constructed network models around dimensions such as psychological stress, negative emotions, burnout levels, and interpersonal relationships among medical students, accurately identifying core symptoms and protective factors within different psychological systems, thereby providing a mature research paradigm for this study.^16–19 Furthermore, network analysis has been increasingly applied to medical-student populations, and several empirical studies with closely-related themes have performed network modelling for positive-psychological resources. Wen, Zhang, Zhu, Wang, Feng, Tao, Chen, Chen, Jin, Zhu and Chang²⁰ adopted network analysis to explore associations between loneliness and mental health among first-year medical undergraduates, and examined the protective function of positive-psychological indicators for adverse mental outcomes. Using the same analytical technique, Zhang, Zhao, Lu, Wu and Yang²¹ conducted network-modelling analyses on positive-psychological traits including gratitude, grit, and character strengths among military medical undergraduates, achieving the identification of core and bridge nodes within positive-psychological systems. These studies have carried out empirical investigations on medical-student positive-psychology topics via network-analysis approaches and offer referable analytical paradigms for this line of research.

Although both social-emotional competence and psychological capital contribute substantially to medical students’ psychological adaptation, most existing studies investigate these two psychological systems in isolation. From the conceptual framework of Conservation of Resources Theory (COR), psychological capital and social-emotional competence are categorized as individual personal psychological resources: psychological capital refers to action-oriented internal agentic resources, while social-emotional competence represents interpersonal emotion-regulation resources. The theory posits that psychological resources are not mutually independent and may produce gain spirals or experience resource loss.²² Conventional regression analyses are typically applied to test unidirectional predictive hypotheses and struggle to simultaneously depict complex patterns of covariation among multiple resources. The network-analysis paradigm provides a valuable extension to COR. By avoiding preset causal directions, it models pairwise associations among diverse resources and conforms to the theoretical assumption of interrelated psychological resources.²³ Existing network-analysis research on medical students largely centers on negative symptoms including depression and burnout. Few empirical efforts integrate the two multi-dimensional positive-psychological systems within one analytical framework, which hinders the detection of their internal correlational features as well as the identification of in-system core nodes and cross-module bridge nodes. This forms the core conceptual gap motivating the present study.

In summary, most existing studies examine the impact of either social-emotional competence or psychological capital on medical students’ psychological well-being in isolation, and primarily rely on traditional linear analysis, which fails to capture the complex dynamic relationships among the various dimensions of these two psychological systems. To date, few studies have employed network analysis methods to simultaneously explore the overall structure and interaction mechanisms between the two. Based on Resource Conservation Theory, this study focuses on medical students. After controlling for demographic variables such as age and grade level, we constructed a regularized partial correlation network model and proposed the following hypotheses: H1: There are extensive associations among the various dimensions of social-emotional competence and psychological capital in medical students, and the overall network exhibits stable structural integrity. H2: Prominent core nodes exist within the integrated network, and these nodes play a key driving role in the entire positive psychological system. H3: Specific dimensions serve cross‑module bridge functions and play an important associative‑bridging role between the two systems of social‑emotional competence and psychological capital.

Materials and Methods

Participants

This study adopted convenient cluster sampling to recruit medical students from an undergraduate medical university in Anhui Province, China. Natural classes were used as sampling units, and classes across junior and middle grades from various medical majors were selected for group data collection. The online questionnaire was developed on an online platform and included demographic information, the Social-Emotional Competence Scale, and the Psychological Capital Scale. To reduce social desirability bias, the front page of the questionnaire clearly informed participants that the survey was anonymous, no personally identifiable information would be collected, and the data would only be used for scientific research. Participants were asked to respond according to their actual experiences. Uniformly trained researchers administered the questionnaire in classrooms and read standardized instructions on site. All students provided informed consent before completing the survey. All items were set as mandatory to minimize missing data.

Strict data quality screening was conducted after questionnaire collection. Invalid responses characterized by patterned answering, invariant selections, or extreme repetitive responses were excluded. Ultimately, 1217 valid questionnaires were retained, corresponding to an effective response rate of 93.6%. Among the participants, there were 391 males (32.1%) and 826 females (67.9%). The distribution by grade level was as follows: 624 first-year students (51.3%), 241 second-year students (19.8%), 218 third-year students (17.9%), and 134 fourth-year students (11.0%). The mean age was 19.33 years (SD = 1.38). There were 312 only children (25.6%) and 905 non-only children (74.4%). This study was reviewed and approved by the local ethics committee (Approval No.: LCYXY00024), and all participating students provided informed consent. This study complies with the Declaration of Helsinki.

Materials

Social-Emotional Competence

The social-emotional competence scale for Chinese college students developed by Chen Zhaojun⁵ was used. It consists of 26 items and is divided into four dimensions: self-relationship competence, interpersonal relationship competence, group relationship competence, and responsible decision-making competence. Participants rated each item on a 5-point scale (1=strongly disagree, 5=strongly agree), with higher scores reflecting higher Social-emotional competence. In this study, the Cronbach’s α for this scale was 0.96. In the present study, the scale demonstrated acceptable construct validity: χ²/df = 3.663, TLI = 0.919, CFI = 0.927, RMSEA = 0.047, SRMR = 0.046.

Psychological Capital

The Psychological capital questionnaire (PCQ), developed by Luthans, Avey, Avolio, Norman and Combs²⁴ and adapted by Zhang Kuo, consists of 26 items.²⁵ It measures an individual’s level of psychological capital across four dimensions: self-efficacy, psychological resilience, hope, and optimism. Participants rated each item on a 7-point scale (1=strongly disagree, 7=strongly agree), with higher scores reflecting higher psychological capital. In this study, the Cronbach’s α for this scale was 0.90. In the present study, the scale demonstrated acceptable construct validity: χ²/df = 5.496, TLI = 0.916, CFI = 0.931, RMSEA = 0.061, SRMR = 0.050.

Data Processing

Following questionnaire collection, data‑cleaning procedures were performed to exclude responses with patterned answering, yielding a final valid sample of 1217 participants. Simulation work on regularized partial-correlation network analysis suggests that for networks with around eight nodes, a sample size of 250 can yield stable parameter estimates.¹⁵ With 1217 valid observations, the present study far exceeds this minimum sample‑size threshold and provides adequate statistical power for estimating edge weights and all centrality indices. Descriptive statistics and Pearson correlation analysis were conducted using SPSS 26.0 software to calculate the mean, standard deviation, and correlation coefficients for each dimension. Based on the correlation matrix, a preliminary screening for multicollinearity was performed. After completing the basic correlation analysis, R 4.4.1 statistical software was used, along with the bootnet and qgraph packages, to construct a regularized partial correlation network model based on the EBICglasso algorithm. Following the standards of previous studies, the tuning parameter γ was set to 0.5 to balance the sensitivity and specificity of network edge identification.

Statistically significant differences in scores across dimensions were observed between groups for several demographic variables. Therefore, four demographic confounding covariates were included to control for these factors. We employed the residual regression method, in which the four covariates were treated as predictors in separate linear regressions for the eight psychological dimensions. We extracted the regression residuals for each dimension and constructed a partial correlation network based on the residual matrix to isolate the influence of demographic variables on the association structure.^15,26

This study divided the network modules according to the theoretical dimensions of the original scale, categorizing all variables into two independent modules. Module 1 represents the psychological capital dimension, comprising self-efficacy, resilience, hope, and optimism. Module 2 represents the social-emotional competence dimension, comprising self-relationships, interpersonal relationships, group relationships, and responsible decision-making. Subsequent analyses of network strength are all conducted based on this theoretical module classification. After the model was constructed, four core network indicators were calculated simultaneously. The statistical implications of each indicator are as follows. First, strength centrality refers to the sum of the absolute values of the weights of all edges connected to a given node, reflecting the node’s global connectivity breadth within the entire network. Second is betweenness centrality, which counts the total number of times a node lies on the shortest path between any two nodes and is used to measure the node’s role in mediating information flow. Third, closeness centrality is the reciprocal of the average shortest distance from one node to all other nodes within the network, which reflects the node’s efficiency of information transmission. Fourth, bridge strength, which counts only the sum of edge weights spanning two major modules and is specifically used to identify key connecting nodes linking different modules. After calculating all metrics, global core nodes in the network were selected based on strength centrality, and key bridge nodes spanning modules were identified using bridge strength, enabling a systematic analysis of the interaction patterns and associative characteristics among the various dimensions of medical students’ psychological capital and social-emotional competencies.¹⁵

Results

Common Method Bias Test

This study employed Harman’s one-factor test to assess common method bias. An unrotated principal component analysis was conducted on all scale items. The results showed that a total of 8 common factors with eigenvalues greater than 1 were extracted. The initial eigenvalue of the first common factor was 18.091, accounting for 34.79% of the variance, which did not exceed the critical threshold of 40%.²⁷ The results indicate that there is no serious common method bias in this study.

Descriptive Statistics and Correlation Analysis

Descriptive statistics and Pearson correlation analysis were conducted on the study variables (Table 1). All dimensions of psychological capital and social-emotional competence showed significant positive correlations with their corresponding total scores, indicating a strong internal relationship within each of the two variable systems. The total score for psychological capital showed a significant moderate positive correlation with the total score for social-emotional competence (r=0.58, p<0.01). Cross-system correlation results showed that, among the dimensions of psychological capital, self-efficacy (r=0.53, p<0.01), hope (r=0.57, p <0.01), and optimism (r=0.50, p<0.01) showed moderate positive correlations with the total score for social-emotional competence, while resilience (r=0.18, p<0.01) showed a relatively weaker correlation. Among the dimensions of social-emotional competence, responsible decision-making (r=0.48, p<0.01) and group relationships (r=0.40, p<0.01) showed strong correlations with the total psychological capital score, while interpersonal relationships (r=0.37, p<0.01) showed a relatively weaker correlation.

Table 1.

Descriptive Statistics and Pearson Correlation Analysis of the Study Variables

Variables	1	2	3	4	5	6	7	8	9	10
Psychological capital	1.00
Self-efficacy	0.89**	1.00
Resilience	0.50**	0.30**	1.00
Hope	0.87**	0.72**	0.24**	1.00
Optimism	0.84**	0.68**	0.14**	0.75**	1.00
Social-emotional competence	0.58**	0.53**	0.18**	0.57**	0.50**	1.00
Self-relationships	0.69**	0.69**	0.17**	0.67**	0.59**	0.75**	1.00
Interpersonal relationships	0.37**	0.30**	0.19**	0.37**	0.30**	0.80**	0.45**	1.00
Group relationships	0.40**	0.33**	0.13**	0.40**	0.36**	0.89**	0.48**	0.68**	1.00
Responsible decision-making	0.48**	0.46**	0.13**	0.46**	0.42**	0.90**	0.59**	0.64**	0.74**	1.00
	4.65	4.61	4.32	4.87	4.87	4.04	3.79	4.23	4.11	4.05
S	0.61	0.86	0.61	0.72	0.90	0.41	0.51	0.49	0.50	0.48

Notes: n=1217; **p<0.01.

Network Analysis of Social-Emotional Competencies and Psychological Capital

Network Structure

A one-way analysis of variance (ANOVA) revealed that grade level was a statistically significant factor in the differences in scores for self-efficacy (F=3.585, P=0.013), hope (F=4.044, P=0.007), and interpersonal relationships (F=3.674, P=0.012). Pearson correlations indicated that age was weakly negatively correlated with hope (r=−0.057, P=0.045) and interpersonal relationships (r=−0.064, P=0.026). An independent samples t-test revealed that gender significantly influenced only resilience (P=0.038) and group relationships (P=0.010) scores. Regarding whether participants were only children, group relationships were the only dimension showing between-group differences (P=0.029). To eliminate confounding biases introduced by the four demographic indicators, grade level, age, gender, and whether participants were only children were treated as covariates. Linear regressions were performed separately on the eight psychological dimensions, and residuals were extracted. An EBIC-regularized partial correlation network was constructed based on the residual matrix. The overall network structure after adjustment is shown in Figure 1.

Figure 1.

Variable network structure diagram.

Notes: The shorter and thicker the line, the stronger the connection.

Abbreviations: SE, Self-Efficacy; RE, Resilience; HO, Hope; OP, Optimism; SR, Self-Relationships; IR, Interpersonal Relationships; GR, Group Relationships; RD, Responsible Decision-Making.

Among the four dimensions of psychological capital, the connection between hope and optimism had the highest weight (r=0.459). This was followed by the connections between self-efficacy and hope (r=0.273), self-efficacy and optimism (r=0.265), and self-efficacy and resilience (r=0.210). All of these were positive correlations. Resilience and optimism showed a weak negative correlation (r=−0.096), while resilience and hope showed a weak positive correlation (r=0.081).

Among the four dimensions of social-emotional competence, the relationship between group relationships and responsible decision-making was the strongest (r=0.491). This was followed by the relationships between interpersonal relationships and group relationships (r=0.372), interpersonal relationships and responsible decision-making (r=0.238), and self-relationships and responsible decision-making (r=0.221), all of which were positively correlated. The relationship between self-relationship and other-relationship showed a weak positive correlation (r=0.078).

Among cross-module correlations, the correlation between self-efficacy and self-relationship had the highest weight (r=0.328), followed by the correlation between hope and self-relationship (r=0.227). Both of these correlations were positive. There were also several weak positive correlations, namely between hope and interpersonal relationships (r=0.029), hope and group relationships (r=0.033), optimism and group relationships (r=0.045), and self-relationship and group relationships (r=0.028). There were also several weak negative correlations, namely between self-efficacy and interpersonal relationships (r=−0.064), self-efficacy and group relationships (r=−0.061), resilience and self-relationships (r=−0.053), and resilience and responsible decision-making (r=−0.041).

Node Centrality

Strength centrality reflects the breadth of a variable’s connections within a network. A higher value indicates that the node is more closely linked to other variables. Based on the results of the network model adjusted for four covariates, self-efficacy (Strength=1.292) exhibited the highest strength centrality. This was followed by hope (Strength=1.107), responsible decision-making (Strength = 1.088), and group relationships (Strength=1.030). This indicates that these four variables are core nodes in the network. Resilience (Strength=0.601) had the lowest strength centrality, indicating relatively weak connections with the other variables in the network.

Betweenness centrality is used to measure a node’s bridging role in the network. Self-relationships (Betweenness=20) had the highest betweenness centrality, followed by responsible decision-making (Betweenness=16). These two are the key mediating nodes for the transmission of relationships between variables in the network. Self-efficacy, hope, and interpersonal relationships possessed a certain degree of mediating function. Resilience, optimism, and group relationship all had a betweenness centrality value of 0 and do not play a role in transmitting paths between variables.

Closeness centrality reflects the tightness of connections between a node and all other dimensions in the network as well as information transmission efficiency. Larger values mean closer connections with other nodes in the whole network and higher efficiency of information transmission. The variables sorted by closeness centrality from highest to lowest are Self-relationship (Closeness=0.024), self-efficacy (Closeness=0.023), responsible decision-making (Closeness=0.020), hope, group relationships, optimism, resilience, interpersonal relationships (Closeness=0.0156). Self-relationship has shorter connection paths to all other dimensions and presents the optimal information transmission efficiency. See Figure 2.

Figure 2.

Node Centrality.

Notes: The values for closeness centrality were amplified by a factor of 1,000 to improve the visibility of the graph.

Abbreviations: SE, Self-Efficacy; RE, Resilience; HO, Hope; OP, Optimism; SR, Self-Relationships; IR, Interpersonal Relationships; GR, Group Relationships; RD, Responsible Decision-Making.

Bridge Strength

The bridge strength indicator is used to measure the degree of association between nodes across two modules. A higher value indicates that the variable is more likely to serve as a bridge transmitting effects between the psychological capital and social-emotional competence modules. The network model results show that self-relationship had the highest bridge strength (Bridge Strength=0.657), making it the core cross-module bridge node connecting the psychological capital module and the social-emotional competence module. Self-efficacy (Bridge Strength= 0.545) and hope (Bridge Strength=0.294) followed, playing secondary bridging roles in cross-module transmission. Resilience (Bridge Strength=0.214) and interpersonal relationships (Bridge Strength = 0.053) ranked in the middle. Group relationships (Bridge Strength=0.139), responsible decision-making (Bridge Strength = 0.138), and optimism (Bridge Strength=0.094) exhibited progressively lower bridge strengths. Among these, optimism had the lowest bridge strength value, the weakest association with other modules, and the strongest module independence. See Figure 3.

Figure 3.

Centrality indicators for bridge strength.

Abbreviations: SE, Self-Efficacy; RE, Resilience; HO, Hope; OP, Optimism; SR, Self-Relationships; IR, Interpersonal Relationships; GR, Group Relationships; RD, Responsible Decision-Making.

Network Accuracy and Stability

This study employed 2,000 nonparametric bootstrap iterations to test the estimation accuracy of network edge weights. The overall confidence intervals for edge weights were narrow, indicating good accuracy in estimating the strength of partial correlations among variables. To test the robustness of node centrality measures, 2,000 bootstrap stability analyses were conducted. The correlation stability coefficient (CS) was used to assess the reliability of the results. The generally accepted threshold in the academic community is CS > 0.5,¹⁵ indicating that an indicator is stable The results show that the CS coefficients for node strength, betweenness centrality, and closeness centrality were 0.872, 0.678, and 0.717, respectively. All three indicators exceeded the 0.5 threshold. Among them, the CS coefficients for node strength and closeness centrality were higher than 0.7, indicating excellent stability. The CS coefficient for betweenness centrality fell between 0.25 and 0.7, indicating acceptable stability. Overall, the centrality indicators for each node in this network are robust and reliable, and the interpretation of node functions based on centrality is supported by credible data.

Discussion

This study employed regularized partial correlation network analysis to investigate the association structure, node characteristics, and connectivity patterns of the two positive psychological systems—social-emotional competence and psychological capital—among medical students, thereby clarifying the dynamic interdependencies among the various dimensions. The overall results indicate that the dimensions of these two psychological systems are extensively interconnected, with a generally stable network structure, though a small number of weak negative correlations were also observed. This confirms that the positive psychological traits of medical students do not exist in isolation but rather form an organic whole characterized by mutual influence and synergistic effects. These findings align with the core tenets of resource conservation theory,²³ which posits that an individual’s positive psychological resources possess characteristics of accumulation, depletion, and complementarity. In the context of long-term, high-intensity academic and developmental environments, medical students continuously mobilize and integrate various psychological resources to cope with stress and challenges, thereby forming a stable network of psychological connections.²⁸

Network Structure Analysis

The network results adjusted for demographic confounders showed that dimensions of social‑emotional competence and psychological capital among medical students were dominated by positive connections, with only a small number of weak negative edges. The overall magnitude of associations was moderate, and no severe multicollinearity was detected. Previous studies have largely relied on traditional correlation and regression analyses to explore linear relationships among variables,^10,12 which struggle to intuitively illustrate the complex interrelationships among multidimensional indicators. In contrast, network analysis can effectively capture the nonlinear characteristics of psychological systems and has been widely applied in research on medical students’ stress, emotions, and interpersonal relationships.^16–18 The findings of this study further demonstrate that network analysis is equally applicable to exploring the mechanisms underlying positive psychological traits. This approach can provide new analytical insights for research in the field of positive psychology.^14,15

Pearson correlation analysis demonstrated a moderate positive correlation between the total scores of psychological capital and social‑emotional competence, suggesting close intrinsic associations between the two constructs. Riolli, Savicki and Richards²² noted that psychological capital serves as an important protective resource that buffers students against stress. Meanwhile, social-emotional competencies—such as effective emotional regulation and interpersonal skills—can also help individuals maintain a positive mindset and cultivate positive psychological qualities. The two form a mutually reinforcing relationship. Based on an analysis of resource conservation theory, the small number of weak negative correlations can be explained by the limited nature of individual psychological resources and the slight diversion of resource allocation among certain dimensions.²³ For medical students, sufficient psychological capital—such as self-efficacy, hope, and optimism—enables them to cope calmly with academic pressures and clinical setbacks. Meanwhile, strong emotional management, interpersonal collaboration, and responsible decision-making skills can reduce the accumulation of negative emotions and ensure that psychological capital remains at a stable level.^6,8 Nevertheless, given the cross-sectional design, we can only observe concurrent covariation among variables, and the data cannot confirm a bidirectional causal process of mutual promotion.

Correlations and Node Characteristics Among the Internal Dimensions of Psychological Capital

In terms of network edge weights, among the four dimensions of psychological capital, hope exhibited the strongest association with optimism. Self-efficacy also showed strong positive associations with hope and optimism. Resilience had relatively weaker connections with the remaining dimensions, and a weak negative association was observed between resilience and optimism. Hope and optimism are typical positive emotional traits. They complement each other and help medical students withstand feelings of frustration and maintain their enthusiasm for learning and professional aspirations throughout their long medical education.²⁹

Results regarding strength centrality indicate that self-efficacy is the central node of the entire network, which is consistent with the findings of relevant studies both domestically and internationally. In their review, Klassen and Klassen³⁰ proposed that self-efficacy is a key psychological variable influencing medical students’ learning status, clinical performance, and professional mindset. Ferriby and Schaefer³¹ also confirmed that insufficient self-efficacy significantly increases burnout levels among medical students. As the core of the network, self-efficacy permeates various contexts of medical students’ learning, clinical practice, and interpersonal interactions. It serves not only as the core driving force of the psychological capital system but also as a vital link connecting the two major psychological modules. By contrast, resilience showed low strength centrality. Combined with sample characteristics, participants in this study were predominantly junior medical students. This group had limited exposure to clinical practice and little experience coping with major setbacks; thus, resilience may not be fully activated, resulting in its weaker associative role within the network.

Correlations and Node Characteristics Among the Internal Dimensions of Social-Emotional Competence

Within the social-emotional competence module, the edge weight between group relationship and responsible decision-making was the highest (r = 0.491), indicating strong covariation between the two dimensions. This finding aligns with the context of medical education, which has long emphasized the cultivation of teamwork and professional ethics.³² Meanwhile, group relationship and responsible decision-making exhibited high strength centrality, serving as important nodes inside the social-emotional competence module.

Self-related competence had the highest betweenness centrality, closeness centrality and bridge strength across the entire network. It showed extensive associations with multiple nodes from both modules and acted as a critical cross-module bridge node connecting psychological capital and social-emotional competence. This finding provides direct empirical support for the study aim of identifying key cross-module variables. Medical students, who are constantly exposed to a high-pressure, demanding training environment, are prone to issues such as self-doubt and emotional suppression.^1,2 Self-relationship, which focuses on self-awareness, emotional acceptance, and self-regulation, serves as the foundation for individuals to integrate internal and external psychological resources. On the one hand, a robust self-relationship can amplify the effects of positive psychological capital, such as self-efficacy and hope. On the other hand, a stable and positive psychological state also promotes the development of healthy self-perception. This explains why self-relationship can serve as a cross-module resource conduit. An intervention study by Versel, Plezia, Jennings, Sontag-Milobsky, Adams and Shahid⁹ confirmed that training focused on emotions and self-regulation can effectively enhance medical students’ stress management abilities and overall well-being, further substantiating the significant value of self-relationship competencies.

Practical Implications of Cross-Module Connections

The cross-module network results showed that the edge between self-efficacy and self-related competence had the highest weight among all cross-module edges linking the two modules. Combined with the analysis of bridge nodes, it is evident that self-relationship competence serves as the core link running through the entire positive psychology network. Based on these findings, medical schools need not implement indiscriminate interventions across all psychological dimensions when cultivating students’ psychological well-being; instead, they can adopt targeted intervention strategies. First, prioritize self-efficacy as the primary intervention target. Enhance medical students’ academic and clinical self-efficacy through skill training, experiences of success, and positive reinforcement to leverage its role as a core driver. Second, prioritize the cultivation of self-relationship. Offer courses on emotion management, self-acceptance, and stress regulation, leveraging this bridge node to drive the coordinated development of the two psychological systems. Third, address key nodes such as hope, responsible decision-making, and group relationships. Incorporate team collaboration, professional ethics, and career aspiration education tailored to the medical profession to comprehensively strengthen medical students’ foundation in positive psychology.

Current medical education tends to emphasize the teaching of professional knowledge and procedural skills, while lacking systematic cultivation of social-emotional competencies and positive psychological capital.⁶ The findings of this study suggest that medical schools should integrate positive psychological development into daily teaching and student management, establish a routine mental health support system, and reduce the risk of psychological issues such as burnout and anxiety at their source.³

Limitations and Implications

This study inevitably has several limitations. First, as a cross-sectional study, it can only reveal static associations among variables and cannot infer causal relationships or longitudinal patterns. Future longitudinal studies could be conducted to analyze the dynamic evolution of medical students’ psychological networks across different stages. Second, the sample consisted solely of undergraduate medical students from a single region, limiting its representativeness. Future research could involve multicenter, large-sample surveys covering medical students from different institutions, academic programs, and clinical rotation stages. Third, this study examined only the network structures of two variables, social-emotional competence and psychological capital. Future research could incorporate various variables to construct a more complex positive psychology network and comprehensively analyze the interactive mechanisms of medical students’ positive psychological traits. Overall, this study enriches the body of research in the field of positive psychology among medical students and provides empirical references for mental health interventions and the optimization of talent development models in medical education.

Conclusions

Using network analysis methods to explore the interactive structure between medical students’ social-emotional competencies and psychological capital, the following conclusions were drawn. First, there were widespread associations between medical students’ social-emotional competencies and the various dimensions of psychological capital. The overall network structure was stable, with variables primarily connected by positive links. Second, self-efficacy served as the central node of the network. Hope, responsible decision-making, and group relationships also played important roles. Third, self-relationship exhibited both the highest betweenness centrality and bridge strength. It served as a key node connecting the two psychological modules and facilitating the transmission of psychological resources. Fourth, stable interrelationships existed within each dimension of psychological capital and social-emotional competencies. Self-efficacy and self-relationship constituted the most central connection pathways between the two modules. This study adopted network analysis to explore the association structure between social-emotional competence and psychological capital among medical students, and the main conclusions are as follows. First, extensive concurrent associations were observed across all dimensions of the two positive psychological traits. The overall network structure was stable and predominantly characterized by positive connections. Second, self-efficacy served as the core node of the network, while hope, responsible decision-making, and group relationship also functioned as crucial internal nodes. Third, self-related competence exhibited the highest betweenness centrality and bridge strength, acting as the key tie that connects and maintains the association between the two psychological modules. Fourth, stable covariation patterns existed within both the psychological capital and social-emotional competence modules, and the edge between self-efficacy and self-related competence represented the primary cross-module connection.

As a cross-sectional study, the present research only captured concurrent covariation patterns among variables and cannot infer causal or bidirectional effects. The identified core and bridge nodes provide targeted references for psychological intervention among medical students. Future studies could adopt longitudinal network designs, expand multi-center samples, and incorporate negative psychological indicators such as burnout and stress to further deepen relevant research findings.

Funding Statement

This research was supported by the Health Humanities Research Center, Clinical College of Anhui Medical University (Project No. 2025PT002), the study on factors influencing medical students’ social-emotional competence competency (2025XJKY021), the research on the course of medical psychology (2025XJ029), the study on the relationship between career calling among medical students and medical humanities competence (2025XJKY026), Anhui, China, and the Research on Social Equity and Common Prosperity in the Health Sector Based on the Spirit of the 20th National Congress (2024AH052625).

Data Sharing Statement

All scales used in this study were mature scales. And all of them are labeled with the literature referenced in the Materials section of the article. The data presented in this study are available on request. The data are not publicly available due to concerns about privacy and ethics in personal decision-making.

Ethics Approval and Consent to Participate

The studies involving humans were approved by the Clinical College of Anhui Medical University, Hefei, Anhui, China. This study has been conducted in accordance with the Declaration of Helsinki. The participants provided their written informed consent to participate in this study.

Consent for Publication

The article does not contain any information or images that could lead to the identification of participants.

Disclosure

The authors report no conflicts of interest in this work.