Next Article in Journal
Trends in Swimming Competence Among Youth: A Repeated Cross-Sectional Study (2020–2024)
Previous Article in Journal
Youth Addiction and Well-Being: Analysis of Social, Behavioral, and Economic Factors
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Youth
Volume 5
Issue 4
10.3390/youth5040116
youth-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Rethinking Mental Health Assessment: A Network-Based Approach to Understanding University Students’ Well-Being with Exploratory Graph Analysis

by
Laura García-Pérez
1,*,
Mar Cepero-González
1 and
Jorge Mota
2
1
Department of Didactics of Corporal Expression, Faculty of Education, University of Granada, 18071 Granada, Spain
2
Research Centre in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sport, University of Porto (FADE-UP), 4050-313 Porto, Portugal
*
Author to whom correspondence should be addressed.
Youth 2025, 5(4), 116; https://doi.org/10.3390/youth5040116
Submission received: 26 June 2025 / Revised: 24 October 2025 / Accepted: 29 October 2025 / Published: 3 November 2025
Browse Figures
Versions Notes

Abstract

Mental health (MH) in university students is often studied through isolated variables. However, a dynamic systems perspective suggests that psychological well-being results from interactions among multiple dimensions such as personality, mood, resilience, self-esteem, and psychological distress. A total of 928 university students (M = 21.01 ± 1.95) completed validated questionnaires: Big Five Inventory (BFI-44) for personality, Profile of Mood States (POMS), Connor-Davidson Resilience Scale (CD-RISC 25), Rosenberg Self-Esteem Scale, and Depression Anxiety Stress Scale (DASS-21). Exploratory Graph Analysis (EGA) using the EGAnet package in RStudio (v. 2025.09.01) was employed to identify latent dimensions and their interconnections. EGA revealed five stable and interconnected dimensions with good fit indices (TEFI = −9.00; ≥0.70): (a) Personality as socio-emotional regulation, (b) Mood as a generalized affective continuum, (c) Resilience as a unified coping process, (d) Self-esteem based on competence and self-worth, and (e) Psychological distress integrating depression, anxiety, and stress. MH appears as a complex and dynamic network of interrelated psychological components. This network-based approach provides a more integrative understanding of well-being in students and supports the development of interventions that target multiple dimensions simultaneously, enhancing effectiveness in academic settings.

1. Introduction

The concept of mental health (MH) has undergone a significant transformation, shifting from a disease-centered perspective to an approach that prioritizes psychological well-being and life satisfaction (Gautam et al., 2024; Lombardo et al., 2018). Initially, the term “mental hygiene,” coined by Adolf Meyer and popularized after the establishment of the National Committee for Mental Hygiene in 1909, focused on the treatment of psychiatric disorders and later expanded to include the prevention and management of mild forms of mental disability (Shorter, 2008; Bertolote, 2008). This approach evolved when the World Health Organization (WHO) introduced the term “MH”, adopting a more holistic perspective. In its Constitution, the WHO defined MH as “a state of complete physical, mental, and social well-being, and not merely the absence of disease or infirmity” (Larsen, 2021), surpassing traditional mind–body dichotomies and incorporating a social dimension inspired by 19th-century European developments (Bertolote, 2008). Today, this perspective remains relevant, as MH is understood not only as the absence of disorders such as anxiety, stress, or depression but as a positive state that integrates elements such as psychological well-being (Ruggeri et al., 2020), self-esteem (Henriksen et al., 2017), and coping abilities (Fradelos et al., 2023).
University settings represent a uniquely demanding psychosocial environment in which academic workload, evaluation cycles, social transition, and identity formation co-occur at high pace. These features make students a salient population for MH monitoring and early support. Evidence from university cohorts shows a high prevalence of anxiety, depression and stress as well as fluctuating mood profiles and resilience demands, with direct links to academic functioning and help-seeking patterns (Auerbach et al., 2018; Griffin et al., 2025; Li et al., 2022; Lipson et al., 2022; Song & Hu, 2024). Accordingly, focusing on students is not merely convenient but conceptually warranted for a dynamic, systems-oriented assessment that can inform campus-based prevention and service delivery.
Currently, MH is conceptualized as an individual’s capacity to maintain optimal functioning in complex situations, encompassing key aspects such as the absence of psychopathology, emotional balance, adaptability, emotional regulation, and resilience (WHO, 2022). This multidimensional approach highlights MH as a dynamic and positive construct, emphasizing not only treatment but also prevention and the promotion of overall well-being. In recent decades, numerous validated psychometric instruments have been developed to assess and understand MH and its underlying mechanisms, evaluating specific psychological domains such as depression, anxiety, and general MH (Breedvelt et al., 2020). While traditional scales such as the Center for Epidemiologic Studies Depression Scale (CES-D), the Beck Depression Inventory-II (BDI-II), and the Patient Health Questionnaire for Depression (PHQ-9) assess depression, and tools such as the Generalized Anxiety Disorder Scale (GAD-7) and the State-Trait Anxiety Inventory (STAI) measure anxiety, more recent approaches incorporate broader dimensions of well-being. Instruments such as the Depression, Anxiety, and Stress Scale (DASS-21) and the General Health Questionnaire (GHQ-12) evaluate general MH, expanding the scope of assessment to include key protective factors such as resilience, self-esteem, and mood states, which play a crucial role in psychological well-being and the ability to adapt to academic and social demands (Abulfaraj et al., 2024; Garces et al., 2024; Huguenin et al., 2024; Osborn et al., 2022). Within this protective set, resilience is especially salient in students, where it relates to reduced distress and better adaptive functioning, thereby justifying its inclusion as a core dimension in university contexts (García-Pérez et al., 2025).
Resilience, understood as the ability to adapt and recover from adversity through mental, emotional, and behavioral flexibility, is a fundamental component of psychological well-being, particularly for university students. It allows them to manage academic, social, and personal challenges while reducing the effects of stress (American Psychological Association, 2018; Abulfaraj et al., 2024). To measure resilience, instruments such as the Connor-Davidson Resilience Scale (CD-RISC) and the Brief Resilience Scale (BRS) provide key metrics on how individuals cope with everyday challenges, highlighting its role in promoting adaptive functioning, academic success, and stress reduction (Suhaimi et al., 2024; Linden et al., 2022). Research has shown that resilience contributes significantly to psychological well-being and academic performance, helping students recover more effectively from stressors (Howe et al., 2012; Pidgeon et al., 2014). Alongside resilience, self-esteem plays a central role in MH, influencing emotional stability, motivation, and interpersonal relationships (de Prada et al., 2024; Savitri et al., 2023). Commonly assessed using the Rosenberg Self-Esteem Scale (RSE), self-esteem is linked to a positive self-perception, reducing the risk of depressive symptoms and fostering greater academic and professional achievements. Strengthening self-esteem during university years facilitates a smoother transition into the workforce and provides a solid foundation for facing future challenges (Pierce et al., 2017; Savoji & Ganji, 2013).
Mood states are also essential for emotional regulation and psychological balance, as they influence individuals’ ability to cope with daily stressors. Tools such as the Trait Meta-Mood Scale (TMMS) and the Profile of Mood States (POMS) allow for the assessment of different dimensions of mood, including tension, anger, confusion, depression, fatigue, and vigor (Moeller et al., 2020; Shichiri et al., 2016). While negative emotions increase stress vulnerability and hinder adaptive responses, vigor emerges as a key protective factor that enhances motivation, strengthens resilience, and promotes better functioning in academic and social environments. Consistent with self-determination theory, higher negative affect and stress are linked to reduced self-regulation and poorer academic functioning, whereas positive affect supports autonomous motivation and persistence (Ryan & Deci, 2000, 2017). Similarly, personality traits play a significant role in how students cope with university life demands. The Big Five Personality Traits—neuroticism, extraversion, openness, agreeableness, and conscientiousness—are strong predictors of MH outcomes (Sadeghi et al., 2015). Neuroticism is associated with higher levels of anxiety and stress, whereas conscientiousness and agreeableness contribute to greater resilience and emotional regulation (Mestre et al., 2017). The Big Five Inventory (BFI-44) is one of the most widely used tools for assessing the relationship between personality traits and MH.
Among university students, mood fluctuates with assessment peaks; personality shapes socio-emotional regulation under coursework pressure; self-esteem tracks competence and self-efficacy beliefs; resilience buffers stress exposure; and distress captures the convergence of depression–anxiety–stress under intensive academic calendars.
Recent work frames MH as an interconnected system of psychological constructs rather than a collection of isolated symptoms (Borsboom, 2017; Borsboom & Cramer, 2013; Gautam et al., 2024). Within this network perspective, Exploratory Graph Analysis (EGA)—a complexity-informed psychometric approach—identifies data-driven communities and connectivity patterns by estimating Gaussian Graphical Models with the graphical lasso (GLASSO; Friedman et al., 2008), thereby clustering variables based on their partial-correlation structure (Golino & Epskamp, 2017; Golino et al., 2020). This yields dimensions that are psychometrically stable yet sensitive to context, facilitating visualization of MH as a dynamic, interdependent network rather than discrete categories (Bandeira et al., 2024). Conceptually, such a dynamic view acknowledges reciprocal influences among distress, mood, self-esteem, resilience, and personality, and practically, it supports assessment models that integrate both risk factors and protective mechanisms in university populations (Epskamp & Fried, 2018; Golino & Epskamp, 2017; Golino et al., 2020). Practically, network information can guide brief multi-domain screening (reducing burden while preserving signal), transdiagnostic, skills-based groups (e.g., emotion regulation, cognitive restructuring, behavioral activation) that target highly connected nodes, and stepped-care monitoring aligned with exam periods to detect emerging network “hot spots.”
We selected these dimensions based on their documented relevance in student populations, their complementary coverage of risk (distress) and protection (resilience, self-esteem, mood), and the availability of widely validated instruments that support comparability and replication.
The present study aims to analyze MH in university students using the EGA approach, proposing that MH is composed of a network of interrelated constructs with direct implications for psychological well-being. By integrating variables commonly included in MH assessment protocols, this study seeks to offer a holistic and multidimensional understanding of the construct. The findings will contribute to optimizing intervention strategies, improving preventive and therapeutic approaches, and enhancing the assessment of psychological well-being in academic settings. By adopting a network-based approach, this study emphasizes the importance of viewing MH as a dynamic, interconnected system, allowing for more precise and effective interventions that support university students in maintaining psychological well-being.

2. Materials and Methods

2.1. Study Description

This study was conducted as part of the European Erasmus+ project “RESUPERES” (2021-1-ES01-KA220-HED-000031173), with the primary objective of optimizing and reducing five widely used questionnaires designed to assess psychological constructs in university students. The aim was to develop a unified and integrative model that would capture the most relevant elements from each questionnaire, minimizing redundancy, improving assessment efficiency, and consolidating a more practical and effective tool for application.
The study focused on identifying the strongest connections between items from different MH questionnaires and determining which items contributed most significantly to an integrated model. To achieve this, the replicability and statistical loadings of each item were analyzed, prioritizing key associations between MH indicators such as depression, anxiety, and stress and other relevant constructs, including self-esteem, resilience, personality traits, and mood states, within the university context.

2.2. Participants and Sampling

The sample comprised university students aged 18–26 enrolled in the Faculty of Educational Sciences at the University of Granada (Spain). The sampling frame covered four undergraduate programs; for logistical accessibility and higher student density, we focused on the Early Childhood Education and Primary Education degrees. Inclusion criteria were (a) enrollment in one of these two degrees; (b) age between 18 and 26 years; and (c) signed informed consent with voluntary participation. Exclusion criteria were (a) incomplete or inconsistent questionnaire responses and (b) current use of medication related to diagnosed psychological disorders. Of the 1008 students who initially participated, 80 were excluded for these reasons, yielding a final analytic sample of 928 students (625 women, 303 men; M_age = 21.01, SD = 1.95). This corresponds to 27.6% of the total student population in the two targeted programs. Participation was voluntary, and students could opt into a raffle for a gift card to purchase technological or academic materials. Although recruitment was non-probabilistic, the large sample size and the coverage across academic years strengthen the contextual validity of the findings for these degrees; nevertheless, generalization beyond this setting should be made cautiously.

2.3. Instruments

To assess the MH of participants, five widely validated instruments were used in university settings.
Psychological distress was measured using the Depression, Anxiety, and Stress Scale (DASS-21) (Lovibond & Lovibond, 1995), which evaluates three key dimensions of MH: depression, anxiety, and stress. In this study, the validated Spanish version by Daza et al. (2002) was used, which has demonstrated high reliability and validity, with strong internal consistency (α = 0.94; 95% CI: 0.94–0.95) and has been previously applied in university populations (Kavvadas et al., 2023). The questionnaire consists of 21 items distributed across three subscales, with responses recorded on a four-point Likert scale (0 = not applicable at all to 3 = highly applicable). The final score for each subscale was obtained by summing the corresponding items.
Resilience was assessed using the Connor-Davidson Resilience Scale (CD-RISC 25) (Connor & Davidson, 2003), specifically the Spanish version validated by Manzano-García and Ayala (2013). This scale has shown high validity and reliability in university studies (Gras et al., 2019) and consists of 25 items rated on a five-point Likert scale (0 = never true to 4 = almost always true), yielding a total score ranging from 0 to 100, where higher values indicate greater resilience. In this study, the scale demonstrated excellent internal consistency (α = 0.84; 95% CI: 0.82–0.85).
Self-esteem was measured using the Rosenberg Self-Esteem Scale (RSE) (Rosenberg, 1965), utilizing the Spanish version translated and validated by Atienza et al. (2000), which has been applied in similar studies (Acosta-Gonzaga, 2023). The questionnaire consists of 10 items, with the first five framed positively and the last five negatively to minimize response bias. A letter-based response scale (A to D) was used, where positively worded items were scored from 4 to 1 and negatively worded items from 1 to 4. The total score ranges from 10 to 40, with higher values indicating greater self-esteem. In this study, the scale showed adequate reliability (α = 0.76; 95% CI: 0.74–0.78).
Mood states were assessed using the short version of the Profile of Mood States (POMS), which includes 29 items and was validated in Spanish by Andrade et al. (2010). Participants responded on a five-point Likert scale (0 = not at all to 4 = extremely), measuring five dimensions: one positive (Vigor) and four negative (Tension, Depression, Anger, and Fatigue). The score for each dimension was obtained by summing the corresponding item responses. In this study, the questionnaire exhibited high internal consistency (α = 0.92; 95% CI: 0.91–0.92) and has been applied in previous university studies (Barney et al., 2022).
Personality traits were evaluated using the Big Five Inventory (BFI-44) (Benet-Martínez & John, 1998), a self-report questionnaire developed to measure the five major personality traits. The 44-item questionnaire follows a five-point Likert scale (1 = strongly disagree to 5 = strongly agree) and is organized into five dimensions: Extraversion (8 items), Agreeableness (9 items), Conscientiousness (9 items), Neuroticism (8 items), and Openness to Experience (10 items). Among these, 16 items require reverse scoring. A high score in each dimension indicates a greater level of that specific trait. In this study, the questionnaire demonstrated good reliability and internal consistency (α = 0.73; 95% CI: 0.71–0.75) and has been extensively validated in university contexts (Yang et al., 2022).

2.4. Procedure

The study was approved by the institutional ethics committee [3678/CEIH/2023] and conducted in accordance with the principles established in the Declaration of Helsinki. Data collection took place between October 2023 and May 2024, with research methods and assessment procedures approved by research group [HUM727]. Participants were assured that their data would be treated anonymously and used exclusively for research purposes, ensuring confidentiality and respect for their participation.
The final version of the questionnaire, integrating the previously described instruments, was administered digitally to facilitate accessibility and encourage participation. Clear instructions were provided for proper completion, and the questionnaire was distributed to the university community via institutional mailing lists. Additionally, faculty members from the selected degree programs played a key role in promoting the study by encouraging student participation through direct communication. All participants, as legal adults, provided informed consent by completing the form, confirming their willingness to participate. This data collection method, recognized for its effectiveness, has been successfully employed in previous studies, reinforcing the validity of the chosen approach (García-Pérez et al., 2023).

2.5. Analysis

The statistical procedure was carried out in four main stages, following methodologies previously validated in similar studies (Bandeira et al., 2024).
First, EGA was conducted to examine whether the MH construct could be represented as a network of interrelated constructs. To achieve this, a network estimation method was initially applied, followed by the Walktrap algorithm, which was used to detect communities within weighted networks (Fortunato, 2010). Then, the GLASSO was used to estimate the Gaussian Graphical Model (GGM) (Friedman et al., 2008; Lauritzen, 1996). This process also incorporated the Least Absolute Shrinkage and Selection Operator (LASSO) (Chen & Chen, 2008), with model selection performed using the Extended Bayesian Information Criterion (EBIC) with a gamma (γ) value of 0.5 (Tibshirani, 1996).
Second, to detect communities within the networks, the Walktrap algorithm was employed again, given its capacity to identify clusters within complex structures (Golino et al., 2020; Pons & Latapy, 2005).
Third, network loadings were calculated to determine the contribution of each item to the dimensions identified in the EGA. These loadings were derived from a factor model, allowing for the quantification of each item’s impact within the detected communities (Christensen & Golino, 2021).
Finally, Exploratory Graph Analysis with Bootstrap (bootEGA) was conducted to estimate and analyze the dimensional structure of the model using resampling. A total of 1000 bootstrap samples were generated, allowing for the creation of a sampling distribution of the EGA results and providing greater robustness in data interpretation. This analysis was performed iteratively using a nonparametric approach, replicating the EGA across each bootstrap sample.

2.5.1. Item Selection and Replicability

Item selection was based on replicability, a fundamental criterion for identifying elements that consistently and robustly contributed to the detected communities. To ensure model stability and reduce biases associated with low-reliability items, a replicability threshold of ≥0.70 was established. Only items surpassing this value were retained, ensuring a solid and precise representation of each evaluated construct. This threshold, supported by previous literature (Christensen & Golino, 2021), is considered appropriate for determining node stability in complex structures such as psychological questionnaires.

2.5.2. Model Fit Evaluation

Model fit was assessed using the Total Entropy Fit Index (TEFI), a method particularly suitable for exploratory analyses as it does not require the specification of a prior confirmatory model. Statistical and graphical analyses were performed using the EGAnet package (version 0.9.9) in RStudio, while results were visualized with GGally (version 2.1.2) and ggplot2 (version 3.3.5). These tools facilitated a clear and precise representation of the network structure and its subcomponents. Methodological note on validity after reduction: The reduced item sets resulting from EGA should be interpreted as sample-optimized indicators for network modeling rather than substitutes for the full, validated instruments. Content and criterion validity for the original questionnaires remain anchored in their manuals.

3. Results

Table 1 provides a descriptive overview of the analyzed variables.
The EGA identified the main dimensions of the evaluated constructs within the five applied questionnaires, facilitating the reduction in items and the development of a more parsimonious and robust model.

3.1. Exploratory Graph Analysis—Results of the Applied Questionnaires

The EGA applied to the Depression, Anxiety, and Stress Scale (DASS-21) confirmed its original theoretical structure, identifying three main communities corresponding to the dimensions of depression, anxiety, and stress (Figure S1). The replicability analysis allowed for the optimization of the scale by selecting items with the highest stability. Thirteen items (1, 3, 4, 7, 8, 10, 11, 12, 13, 15, 16, 17, and 21) demonstrated high replicability (>0.90), ensuring their consistency and relevance in measuring the construct. Six items (2, 5, 8, 9, 19, and 20) showed moderate replicability (0.70–0.90) and were retained due to their acceptable stability within the model. Three items (6, 14, and 18) exhibited low replicability (<0.70) and were removed due to their instability and potential impact on the measurement’s reliability. As a result, the DASS-21 scale was reduced to 18 items, ensuring a more parsimonious, efficient, and reliable assessment of depression, anxiety, and stress symptoms in the university population (Figure S2).
The EGA of the Connor-Davidson Resilience Scale (CD-RISC 25) initially identified five communities aligned with the questionnaire’s theoretical dimensions. However, some items showed low replicability, affecting the model’s stability (Figure S3). To optimize the questionnaire structure, items with the highest stability were selected: Seven items (6, 8, 11, 12, 14, 16, and 18) exhibited high replicability (>0.90), confirming their essential contribution to the construction of resilience. Three items (9, 10, and 23) demonstrated moderate replicability (0.70–0.90) and were retained in the final version. Items with low replicability (<0.70) were excluded due to their inconsistency and potential negative impact on the model’s reliability. As a result, the scale was reduced to 10 items, reorganized into two main dimensions, preserving its validity and robustness (Figure S4).
The EGA applied to the Rosenberg Self-Esteem Scale identified three communities within the model (Figure S5). Most items demonstrated high replicability (>0.97), ensuring their stability and coherence within the construct. Item 8 exhibited moderate replicability (0.84), justifying its inclusion in the final version. However, items 6 and 7 showed low replicability (<0.50) and were removed due to their inconsistent contribution, which could affect the model’s reliability. As a result, the final Self-Esteem Scale was reduced to 8 items, ensuring a more parsimonious and robust measurement (Figure S6).
The Exploratory Graph Analysis applied to the Profile of Mood States (POMS-29) identified three communities, differing from the five original communities in the initial version of the questionnaire (Figure S7). The replicability results allowed for the optimization of the scale: Seven items (2, 7, 12, 17, 19, 24, and 27) exhibited high replicability (>0.90), confirming their stable contribution to the model. Twelve items (4, 5, 6, 9, 11, 14, 16, 20, 21, 23, 25, and 29) displayed moderate replicability (0.70–0.90) and were retained due to their acceptable stability. However, ten items (1, 3, 8, 10, 13, 15, 18, 22, 26, and 28) were eliminated due to low replicability (<0.70) and their instability within the model. As a result, the final version of POMS-29 consisted of 19 items, providing a more efficient and reliable measurement of mood states (Figure S8).
The EGA applied to the Big Five Inventory (BFI-44) identified six communities, expanding the five original dimensions of the questionnaire (Figure S9). The replicability analysis allowed for the scale’s optimization: Twenty-four items (1, 3, 4, 5, 6, 9, 14, 15, 16, 17, 19, 20, 21, 23, 26, 27, 29, 31, 35, 36, 38, 39, 43, and 44) exhibited high replicability (>0.90), ensuring their consistency and stability within the model. Seven items (7, 8, 18, 25, 30, 40, and 42) showed moderate replicability (0.70–0.90), justifying their retention in the final version. Thirteen items (2, 10, 11, 12, 13, 22, 24, 28, 32, 33, 34, 37, and 41) exhibited low replicability (<0.70) and were removed due to their limited and inconsistent contribution. As a result, the final version of the BFI-44 was reduced to 31 items, preserving the instrument’s validity while improving its efficiency (Figure S10).

3.2. Exploratory Graph Analysis-Final Model

A new EGA was conducted to generate consistent and parsimonious dimensions. To achieve this, nodes with low replicability in the previous models were excluded, ensuring a more stable and representative structure.
In the final model, five dimensions were identified, aligned with the five initial questionnaires: (1) Personality dimension; (2) Generalized mood state; (3) Psychological resilience; (4) Self-esteem; (5) Psychological distress. Across instruments, the initial pool comprised 129 items, and the final network retained 86 items, indicating that 43 items were removed in total.
The BootEGA analysis confirmed the stability of the structure, indicating a five-dimension solution in 75.8% of the iterations. Additionally, the item stability analysis demonstrated that all nodes in the final model were stable and replicable (≥0.70) (Figure 1).
In terms of model fit, the results showed a TEFI of −8.18 for the previous model and a TEFI of −9.00 for the optimized model, demonstrating a significant improvement in the network structure after the optimization process. Since a more negative TEFI indicates a better fit, these findings confirm that the optimized model reduces complexity without compromising accuracy, providing a more robust, stable, and efficient representation of the evaluated constructs.

3.3. Final Model Stability Assessment

To analyze the robustness and replicability of the final model, stability tests were conducted using the BootEGA procedure, with bootstrap resampling to evaluate the consistency of connections and centrality metrics within the network. This procedure assesses whether the estimated structure and node importance are preserved across repeated samples drawn from the data-generating process.

Connection Stability

Figure 2 presents the distribution of mean connection values within the network (Bootstrap mean) compared to the values from the original sample (Sample).
The S-shaped pattern indicates that most connections within the network are highly stable, with consistent values across resampling iterations. The overlap between the bootstrapped points (black) and the original sample points (red) confirms that the network structure remains intact even under data fluctuations.
Figure 3 presents the stability coefficients for the betweenness, closeness, expected influence, and strength metrics.
The personality community aggregates socio-emotional regulatory tendencies; the mood community behaves as a generalized affective continuum; resilience items converge on a unified coping core; self-esteem focuses on competence/self-worth; and distress integrates depression–anxiety–stress symptoms. Stability profiles show that expected influence and strength are highly reproducible (coefficients > 0.75), indicating that the same items tend to occupy structurally important positions across resamples. By contrast, betweenness displays the lowest stability and declines as resampled case size decreases, so interpretations based solely on bridging should be made with caution. Closeness exhibits moderate stability, maintaining acceptable correspondence with the original estimates.
For instance, an item with high expected influence within the distress community suggests that small changes in that node may propagate broadly through the network; such items are priority targets for screening or skills-based interventions. Conversely, items identified only by betweenness should be corroborated with additional evidence before drawing strong conclusions.
These findings confirm a robust and replicable network structure, with stable connections and centrality metrics under different sampling conditions. To complement this analysis, Table S1 reports node-level betweenness, closeness, expected influence, and strength, providing a detailed view of each variable’s structural role and further supporting the stability of the optimized solution.

4. Discussion

This study provides evidence that MH in university students cannot be considered as a set of isolated dimensions but rather as an interconnected system of variables that mutually influence one another. Through Exploratory Graph Analysis (EGA), five interrelated dimensions were identified: (1) personality, which acts as an emotional regulator; (2) mood state, reorganized as a generalized affective continuum; (3) resilience, structured as an integrated coping process; (4) self-esteem, centered on the perception of competence and self-worth; and (5) psychological distress, which merged depression, anxiety, and stress into a single global phenomenon of psychological impact. These findings suggest that in the university context, traditional constructs tend to integrate based on their interactions rather than functioning as entirely independent entities. While item reduction improved parsimony and clarified the network structure, it does not replace the full content domains of the original instruments. Accordingly, we interpret the reduced indicators as context-optimized proxies and recommend confirmatory validation and external benchmarking before any clinical or administrative use beyond research settings.
Regarding personality, the network analysis revealed that the five traits of the BFI-44 (extraversion, conscientiousness, agreeableness, openness to experience, and emotional stability) did not remain distinct but rather clustered into a single structure. Previous studies have demonstrated that certain personality traits show significant correlations, suggesting that they do not operate completely independently (Chico, 2006). In particular, agreeableness and conscientiousness have consistently been linked across different populations, both reflecting forms of prosocial regulation and their impact on life satisfaction (Mayungbo, 2016). Similarly, extraversion tends to correlate with neuroticism and openness to experience, reinforcing the idea that personality is an interconnected continuum. In academic settings, self-discipline and organization favor cooperation and adherence to rules, promoting social and academic adaptation (Shih & Chuang, 2013; Guay et al., 2013). This conceptual overlap may have contributed to the fusion of traits into a single dimension within the network analysis, reflecting a functional core centered on socio-emotional regulation (Hughes et al., 2020).
The POMS analysis showed that mood states, rather than clustering into distinct categories, converged into a single generalized affective state. The elimination of key items such as “nervous,” “relaxed,” “exhausted,” or “vigorous” suggests that in this university context, emotional states do not form separate structures but instead converge into a global affective continuum (Paoloni, 2014). Theoretically, mood states are dynamic and transient constructs, characterized by fluctuations in emotional valence (positive or negative) and physiological activation. Although POMS traditionally distinguishes different dimensions, the fusion observed in the network analysis suggests that in academic settings, where psychological and social demands are constantly changing (Brennan et al., 2018), mood may not manifest in separate categories but rather as a single dynamic spectrum reflecting overall psychological well-being (Febrilia et al., 2011). The reduction of items seems to have facilitated this integration, removing unnecessary distinctions between positive and negative effects and favoring the emergence of a global mood state. This approach has already been implemented in various contexts, where a single composite score has been used to assess general changes in participants’ well-being (Yoshihara et al., 2011).
Network analysis revealed a significant reduction in the number of resilience-related items, retaining only those that represent its core essence. Key elements related to the ability to recover from difficult experiences, perseverance, self-confidence, and regulation under pressure were preserved. From a theoretical perspective, resilience is often divided into multiple components, such as social support (Kay, 2016), cognitive flexibility, emotional regulation (Nakhostin-Khayyat et al., 2024), and sense of purpose (Sharma & Yukhymenko-Lescroart, 2024). However, the results suggest that in the academic context, resilience is better understood as a unified coping process, where the ability to stay focused under pressure and seek support during critical moments act as central regulatory mechanisms (García-Jiménez et al., 2023). Another relevant aspect is the inclusion of items emphasizing a positive and adaptive outlook on life, indicating that optimistic thinking (Salgado-Lévano, 2012) and acceptance of uncertainty are key factors in how university students experience and develop resilience (Castro-Méndez & Suárez-Cretton, 2024).
Network analysis identified self-esteem as a dimension focused on perceived competence and self-worth (Rama & Sarada, 2017), removing items related to general life satisfaction. The retained items assessed personal worth, self-efficacy, and negative self-criticism, ensuring a balance between positive and negative indicators. The removal of items such as “Overall, I feel satisfied with myself” suggests that these may be more closely linked to life satisfaction than to a stable perception of self-worth. This restructuring indicates that in university students, self-esteem is primarily related to personal competence (Acosta-Gonzaga, 2023; Ümmet, 2015) rather than an overall emotional evaluation of the self. Its final structure in the network analysis maintains a balance between self-confidence and self-deprecating thoughts, allowing for a more precise assessment of this construct.
Network analysis also revealed that depression, anxiety, and stress converged into a single dimension of generalized distress, reflecting their high interdependence and shared manifestations in MH. Although theoretically distinct, these states tend to coexist, suggesting that in this context, they function as a global phenomenon of psychological affectation rather than separate entities. Previous studies have shown that these disorders share neurobiological and cognitive foundations (Daviu et al., 2019), supporting their integration within network analysis. Common mechanisms, such as cognitive rumination (Wong et al., 2023), emotional hypervigilance (Meng et al., 2020), and difficulties in emotion regulation (Menefee et al., 2022), contribute to their interconnection. Additionally, symptoms such as fatigue, emotional tension, and concentration difficulties are common in all three conditions (Kraft et al., 2023; Van de Leur, 2024; Lv et al., 2024), reinforcing the idea that they form part of a continuum of psychological distress rather than independent categories. In the university context, academic stress serves as a central trigger, fostering the interaction between anxiety and depression. The fluctuation and feedback among these negative states explain their interdependence, justifying their grouping into a single dimension.
Most systematic reviews and meta-analyses have examined the variables included in this study individually, focusing on their isolated impact on MH. Generally, these studies highlight that the main MH issues in university students are associated with depression, anxiety, stress, and suicidal thoughts (Ahmed et al., 2023; Al-Garni et al., 2025; Sheldon et al., 2021; Vidović et al., 2024). On the other hand, various studies have independently analyzed the influence of resilience, self-esteem, mood states, and personality in university settings, recognizing that all these variables contribute to the global MH construct (Cui et al., 2024; Yamanaka et al., 2021; Yong et al., 2022). However, this fragmented approach limits the ability to understand how these variables interact and organize within a complex psychological system.
However, the lack of integration in these studies limits the ability to interpret MH as a whole, as they focus exclusively on the isolated impact of each variable. In contrast, this study introduces a network-based perspective, allowing MH to be analyzed as a dynamic system in which multiple factors interconnect and influence one another. From this standpoint, highly connected and co-activated variables are integrated into a model that more accurately captures the complexity of psychological distress and its impact on well-being. This approach avoids reductionist interpretations and facilitates the development of more effective intervention strategies tailored to the reality of university students.
EGA emerges as an innovative methodology for examining the underlying structure of MH, enabling the identification of dimensions without the need for prior assumptions (Peralta et al., 2020). This approach conceptualizes MH as a dynamic network of relationships, allowing for a more flexible and adaptive interpretation. Additionally, EGA has proven to be a reliable technique for detecting latent dimensions, even in the presence of variations in sample size (Golino et al., 2020). Network-based approaches challenge the traditional view of MH as a set of static categories, proposing instead a flexible structure in which connections between variables shift depending on the context. Constructs such as self-esteem can play a transversal role across different areas of well-being, emphasizing the need to reconsider conventional theoretical models.
A key challenge in MH research is the precise identification of its underlying dimensions. While this study does not fully resolve this issue, the findings indicate that EGA facilitates the detection of patterns without imposing predefined structures (Epskamp et al., 2018), thereby enhancing data interpretation (Tamplain et al., 2020). To achieve a rigorous assessment of MH, it is essential to use multidimensional tools. Measurement imprecision can lead to misinterpretations, affecting both research findings and clinical practice. In university settings, where psychological factors interact dynamically, traditional methods may fail to accurately capture these relationships.
The findings highlight the overlap of assessment tools used in university populations, underscoring the need to improve their validity. Furthermore, adopting theoretical frameworks such as complex systems theory and network science represents a paradigm shift in MH understanding, offering new possibilities for its evaluation and academic application.

4.1. Practical Implications for University Services

First, network-informed screening can pragmatically target a compact set of indicators spanning distress, mood, resilience, self-esteem, and socio-emotional regulation—reducing burden while preserving information. Second, because distress nodes couple strongly with mood and self-esteem, transdiagnostic, skills-based groups (e.g., emotion regulation, cognitive restructuring, behavioral activation) may yield broader gains than siloed, diagnosis-specific pathways. Third, resilience and competence/self-worth nodes suggest adding skill-building components (problem-solving, values-based goal setting, self-efficacy training) to prevention workshops. Fourth, periodic short-form monitoring mapped onto the five communities can inform stepped care (escalating intensity only when multi-domain risk accumulates) and allow services to rapidly identify shifts in network hotspots during high-demand academic periods. Finally, partnership with academic units can align calendars and stressor windows (e.g., exam weeks) with preventive programming.

4.2. Limitations and Future Research Directions

EGA identifies communities and connectivity from regularized partial correlations, indicating functional proximity among variables. It does not infer hierarchical or causal direction; such claims require longitudinal or experimental designs. Although this study offers an innovative perspective by conceptualizing MH as a dynamic system of interrelated dimensions, several limitations should be acknowledged. First, the sample comprised university students from a single geographical region and, more specifically, from two degree programs within one faculty. This constrains external validity and limits representativeness to similar institutional contexts. Future work should conduct multi-site replications across faculties, universities, and countries to enhance generalizability.
Second, the cross-sectional design prevents causal inference and precludes testing the temporal stability of the identified dimensions. Longitudinal designs—ideally with multiple waves—are needed to examine how these psychological variables evolve and interact in response to contextual or life changes, and to determine their long-term impact on well-being.
Third, although we employed validated and reliable instruments, further breadth could be achieved by incorporating complementary measures (e.g., SF-36, GHQ-12, TMMS) and adding behavioral or physiological indicators (e.g., sleep quality, physical activity levels). Future studies integrating multimethod data would enable more comprehensive and ecologically valid network models.
Fourth, the sex imbalance (approximately two-thirds women) may have influenced item retention and network topology. Planned replications should examine measurement invariance and network comparability across sex—and, where possible, across academic year and program—to assess robustness.
Finally, advancing toward complex systems and network science frameworks remains a promising avenue to design more integrated, context-sensitive interventions for university populations, linking network “hot spots” to targeted prevention, stepped care, and service planning on campus.

5. Conclusions

This study shows that university students’ MH is best understood as a dynamic, interconnected system rather than a set of isolated constructs. Within this network, personality functions as socio-emotional regulation, mood forms a generalized affective continuum, resilience consolidates into a unified coping mechanism, self-esteem centers on perceived competence and self-worth, and psychological distress reflects a global experience of emotional suffering. These dimensions co-evolve and influence one another, capturing the complexity of students’ psychological functioning.
Adopting a network-based perspective yields a more nuanced account of MH by emphasizing relational structure and moving beyond reductionist models. In practical terms, network information can guide brief, multi-domain screening that reduces burden while preserving signal; transdiagnostic, skills-focused interventions (e.g., emotion regulation, behavioral activation) that target highly connected nodes; and stepped-care monitoring aligned with academic stress periods to detect emerging “hot spots.” Taken together, our findings support multidimensional, systems-oriented frameworks for MH research and practice in higher education—frameworks that better reflect lived experience and enable more precise, adaptive support for students. Future longitudinal and multi-site work can test the temporal stability and generalizability of these network structures, strengthening their value for prevention and service design.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/youth5040116/s1, Figure S1: Initial model of DASS-21; Figure S2: Final model of DASS-21; Figure S3: Initial model of CD-RISC 25; Figure S4: Final model of CD-RISC 25; Figure S5: Initial model of Rosenberg scale; Figure S6: Final model of Rosenberg scale; Figure S7: Initial model of POMS; Figure S8: Final model of POMS; Figure S9: Initial model of BFI-44; Figure S10: Final model of BFI-44; Table S1: Node Centrality.

Author Contributions

Conceptualization, L.G.-P. and J.M.; methodology, L.G.-P.; software, L.G.-P.; validation, L.G.-P., J.M. and M.C.-G.; formal analysis, L.G.-P.; investigation, J.M.; resources, L.G.-P.; data curation, J.M.; writing—original draft preparation, L.G.-P.; writing—review and editing, L.G.-P.; visualization, L.G.-P.; supervision, J.M.; project administration, J.M.; funding acquisition, L.G.-P. and M.C.-G. All authors have read and agreed to the published version of the manuscript.

Funding

This study is part of the results obtained within the framework of the Erasmus+ RESUPERES project (2021-1-ES01-KA220-HED-000031173), the only European international project awarded to the University of Granada in 2021, funded by the European Union. Additionally, this research was supported by funding from the Spanish Ministry of Universities through the “Formación de Profesorado Universitario (FPU)” predoctoral grant awarded to L.G.-P. (FPU20/01373).

Institutional Review Board Statement

The authors confirm that the research presented in this article met the ethical guidelines, including adherence to the legal requirements of Spain. The study was conducted in accordance with the Declaration of Helsinki (1964), and the protocol was approved by the Ethics Committee of the University of Granada (3678/CEIH/2023, 26 September 2023). All data were handled in compliance with European data protection legislation, in accordance with Regulation (EU) 2016/679 (GDPR) of the European Parliament and of the Council of 27 April 2016.

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors upon request.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
BDI-IIBeck Depression Inventory-II
BFI-44Big Five Inventory (44 ítems)
BRSBrief Resilience Scale
CD-RISCConnor-Davidson Scale
CES-DCenter for Epidemiologic Studies Depression
DASS-21Depression, Anxiety, and Stress Scale (21 ítems)
EGAExploratory Graph Analysis
GAD-7Generalized Anxiety Disorder Scale (7 ítems)
GHQ-12General Health Questionnaire (12 ítems)
GLASSOGraphical Least Absolute Shrinkage and Selection Operator
MHMental Health
PHQ-9Patient Health Questionnaire for Depression
POMSProfile of Mood States
RSERosenberg Self-Esteem Scale
STAIState-Trait Anxiety Inventory
TMMSTrait Meta-Mood Scale
WHOWorld Health Organization

References

  1. Abulfaraj, G. G., Upsher, R., Zavos, H. M. S., & Dommett, E. J. (2024). The impact of resilience interventions on university students’ mental health and well-being: A systematic review. Education Sciences, 14(5), 510. [Google Scholar] [CrossRef]
  2. Acosta-Gonzaga, E. (2023). The effects of self-esteem and academic engagement on university students’ performance. Behavioral Sciences, 13(4), 348. [Google Scholar] [CrossRef]
  3. Ahmed, I., Hazell, C. M., Edwards, B., Glazebrook, C., & Davies, E. B. (2023). A systematic review and meta-analysis of studies exploring prevalence of non-specific anxiety in undergraduate university students. BMC Psychiatry, 23(1), 240. [Google Scholar] [CrossRef] [PubMed]
  4. Al-Garni, A. M., Shati, A. A., Almonawar, N. A., Alamri, G. M., Alasmre, L. A., Saad, T. N., Alshehri, F. M., Hammouda, E. A., & Ghazy, R. M. (2025). Prevalence of depression, anxiety, and stress among students enrolled at King Khalid University: A cross-sectional study. BMC Public Health, 25(1), 354. [Google Scholar] [CrossRef] [PubMed]
  5. American Psychological Association. (2018). Resilience. Dictionary of Psychology. Available online: https://dictionary.apa.org/resilience (accessed on 5 May 2025).
  6. Andrade, E., Arce, C., Torrado, J., Garrido, J., & De Francisco y Arce, I. (2010). Factor structure and invariance of the POMS mood state questionnaire in Spanish. Spanish Journal of Psychology, 13(1), 444–452. [Google Scholar] [CrossRef] [PubMed]
  7. Atienza, F. L., Moreno, Y., & Balaguer, I. (2000). An analysis of the dimensionality of the rosenberg self-esteem scale in a sample of valencian adolescents. Revista de Psicología Universitas Tarraconensis, 22, 29–42. [Google Scholar]
  8. Auerbach, R. P., Mortier, P., Bruffaerts, R., Alonso, J., Benjet, C., Cuijpers, P., Demyttenaere, K., Ebert, D. D., Green, J. G., Hasking, P., Murray, E., Nock, M. K., Pinder-Amaker, S., Sampson, N. A., Stein, D. J., Vilagut, G., Zaslavsky, A. M., Kessler, R. C., & WHO WMH-ICS Collaborators. (2018). WHO World Mental Health Surveys International College Student Project: Prevalence and distribution of mental disorders. Journal of Abnormal Psychology, 127(7), 623–638. [Google Scholar] [CrossRef]
  9. Bandeira, P. F. R., Lemos, L. F., Estevan, I., Webster, E. K., Clark, C. T., Duncan, M. J., Mota, J. A., & Martins, C. L. (2024). Are we assessing motor competence? Evidence-informed constructs for motor competence in preschoolers through an Exploratory Graph Analysis. Journal of Sports Sciences, 43, 109–116. [Google Scholar] [CrossRef]
  10. Barney, D., Pleban, F. T., & Gishe, J. (2022). The measurement of mood states in college students induced by physical activity. American Journal of Health Studies, 36(3), 1–14. Available online: https://www.amjhealthstudies.com/index.php/ajhs/article/view/679 (accessed on 5 May 2025).
  11. Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across cultures and ethnic groups: Multitrait-multimethod analyses of the Big Five in Spanish and English. Journal of Personality and Social Psychology, 75(3), 729–750. [Google Scholar] [CrossRef]
  12. Bertolote, J. (2008). The roots of the concept of mental health. World Psychiatry, 7(2), 113–116. [Google Scholar] [CrossRef] [PubMed]
  13. Borsboom, D. (2017). A network theory of mental disorders. World Psychiatry, 16(1), 5–13. [Google Scholar] [CrossRef] [PubMed]
  14. Borsboom, D., & Cramer, A. O. J. (2013). Network analysis: An integrative approach to the structure of psychopathology. Annual Review of Clinical Psychology, 9, 91–121. [Google Scholar] [CrossRef] [PubMed]
  15. Breedvelt, J. J. F., Zamperoni, V., South, E., Uphoff, E. P., Gilbody, S., Bockting, C. L. H., Churchill, R., & Kousoulis, A. A. (2020). A systematic review of mental health measurement scales for evaluating the effects of mental health prevention interventions. European Journal of Public Health, 30(3), 539–545. [Google Scholar] [CrossRef]
  16. Brennan, J., Cochrane, A., Lebeau, Y., & Williams, R. (2018). Universities, social change and transformation: Global perspectives. In The university in its place (pp. 13–31). Springer. [Google Scholar] [CrossRef]
  17. Castro-Méndez, N. P., & Suárez-Cretton, X. A. (2024). Resilience in non-traditional university students. Revista Colombiana de Educacion, 91(91), 33–55. [Google Scholar] [CrossRef]
  18. Chen, J., & Chen, Z. (2008). Extended Bayesian information criteria for model selection with large model spaces. Biometrika, 95(3), 759–771. [Google Scholar] [CrossRef]
  19. Chico, E. (2006). Personality dimensions and subjective well-being. The Spanish Journal of Psychology, 9(1), 38–44. [Google Scholar] [CrossRef]
  20. Christensen, A. P., & Golino, H. (2021). Estimating the stability of psychological dimensions via bootstrap exploratory graph analysis: A monte carlo simulation and tutorial. Psych, 3(3), 479–500. [Google Scholar] [CrossRef]
  21. Connor, K. M., & Davidson, J. R. T. (2003). Development of a new resilience scale: The Connor-Davidson Resilience Scale (CD-RISC). Depression and Anxiety, 18(2), 76–82. [Google Scholar] [CrossRef]
  22. Cui, M., Ma, X., Tian, L., Xu, W., & Dai, H. (2024). The chain mediating role of stress and resilience in the relationship between anxiety sensitivity and depressive symptoms among Chinese college students. Journal of Affective Disorders Reports, 17, 100821. [Google Scholar] [CrossRef]
  23. Daviu, N., Bruchas, M. R., Moghaddam, B., Sandi, C., & Beyeler, A. (2019). Neurobiological links between stress and anxiety. Neurobiology of Stress, 11, 100191. [Google Scholar] [CrossRef]
  24. Daza, P., Novy, D. M., Stanley, M. A., & Averill, P. (2002). The depression anxiety stress scale-21: Spanish translation and validation with a Hispanic sample. Journal of Psychopathology and Behavioral Assessment, 24(3), 195–205. [Google Scholar] [CrossRef]
  25. de Prada, E., Mareque, M., & Pino-Juste, M. (2024). Self-esteem among university students: How it can be improved through teamwork skills. Education Sciences, 14(1), 108. [Google Scholar] [CrossRef]
  26. Epskamp, S., Borsboom, D., & Fried, E. I. (2018). Estimating psychological networks and their accuracy: A tutorial paper. Behavior Research Methods, 50(1), 195–212. [Google Scholar] [CrossRef] [PubMed]
  27. Epskamp, S., & Fried, E. I. (2018). A tutorial on regularized partial correlation networks. Psychological Methods, 23(4), 617–634. [Google Scholar] [CrossRef] [PubMed]
  28. Febrilia, I., Warokka, A., & Abdullah, H. (2011). University students’ emotion state and academic performance: New insights of managing complex cognitive. Journal of E-Learning & Higher Education, 2011, 879553. [Google Scholar] [CrossRef]
  29. Fortunato, S. (2010). Community detection in graphs. Physics Reports, 486(5), 75–174. [Google Scholar] [CrossRef]
  30. Fradelos, E. C., Papathanasiou, I. V., Dafogianni, C., Misouridou, E., Koutelekos, I., Dousis, E., Vlachou, E., Evangelou, E., Alikari, V., Gerogianni, G., Polikandrioti, M., & Zartaloudi, A. (2023). The effect of psychological resilience and coping strategies on mental health of nurses. In Advances in experimental medicine and biology (pp. 23–30). Springer International Publishing. [Google Scholar] [CrossRef]
  31. Friedman, J., Hastie, T., & Tibshirani, R. (2008). Sparse inverse covariance estimation with the graphical lasso. Biostatistics, 9(3), 432–441. [Google Scholar] [CrossRef]
  32. Garces, N. N., Esteves, Z. I., Santander, M. L., Mejía, D. R., & Quito, A. C. (2024). Relationships between mental well-being and academic performance in university students: A systematic review. Salud, Ciencia y Tecnologia—Serie de Conferencias, 3, 972. [Google Scholar] [CrossRef]
  33. García-Jiménez, M., Trigo, M. E., Varo, M., Aires-González, M. M., & Cano-García, F. J. (2023). Confirmatory factor analysis and gender invariance of the coping strategies inventory in academic university stress. Journal of Empirical Research in Psychology, 35(1), 13–19. [Google Scholar] [CrossRef]
  34. García-Pérez, L., Ubago-Jiménez, J. L., Padial-Ruz, R., & Melguizo-Ibáñez, E. (2025). Profiling pre-service teachers’ resilience, self-esteem, and healthy habits: Associations with psychological well-being and readiness to teach. Acta Psychologica, 261, 105799. [Google Scholar] [CrossRef] [PubMed]
  35. García-Pérez, L., Villodres, G. C., & Muros, J. J. (2023). Differences in healthy lifestyle habits in university students as a function of academic area. Journal of Public Health, 45(2), 513–522. [Google Scholar] [CrossRef] [PubMed]
  36. Gautam, S., Jain, A., Chaudhary, J., Gautam, M., Gaur, M., & Grover, S. (2024). Concept of mental health and mental well-being, it’s determinants and coping strategies. Indian Journal of Psychiatry, 66(Suppl. 2), S231–S244. [Google Scholar] [CrossRef] [PubMed]
  37. Golino, H. F., Christensen, A., & Moulder, R. (2020). EGAnet: Exploratory graph analysis: A framework for estimating the number of dimensions in multivariate data using network psychometrics (R package version 0.9). R Project. [Google Scholar]
  38. Golino, H. F., & Epskamp, S. (2017). Exploratory graph analysis: A new approach for estimating the number of dimensions in psychological research. PLoS ONE, 12(6), e0174035. [Google Scholar] [CrossRef]
  39. Gras, M.-E., Font-Mayolas, S., Baltasar, A., Patiño, J., Sullman, M. J. M., & Planes, M. (2019). The Connor-Davidson Resilience Scale (CD-RISC) amongst Young Spanish Adults. Clínica y Salud, 30(2), 73–79. [Google Scholar] [CrossRef]
  40. Griffin, S. M., Rennie, K., Turner, R., & Treadway, M. T. (2025). Identity change and the transition to university: Implications for cortisol awakening response, psychological well-being and academic performance. Applied Psychology: Health and Well-Being, 17(1), e12608. [Google Scholar] [CrossRef]
  41. Guay, R. P., Oh, I., Choi, D., Mitchell, M. S., Mount, M. K., & Shin, K. (2013). The interactive effect of conscientiousness and agreeableness on job performance dimensions in South Korea. International Journal of Selection and Assessment, 21(2), 233–238. [Google Scholar] [CrossRef]
  42. Henriksen, I. O., Ranøyen, I., Indredavik, M. S., & Stenseng, F. (2017). The role of self-esteem in the development of psychiatric problems: A three-year prospective study in a clinical sample of adolescents. Child and Adolescent Psychiatry and Mental Health, 11(1), 68. [Google Scholar] [CrossRef]
  43. Howe, A., Smajdor, A., & Stöckl, A. (2012). Towards an understanding of resilience and its relevance to medical training. Medical Education, 46(4), 349–356. [Google Scholar] [CrossRef]
  44. Hughes, D. J., Kratsiotis, I. K., Niven, K., & Holman, D. (2020). Personality traits and emotion regulation: A targeted review and recommendations. Emotion, 20(1), 63–67. [Google Scholar] [CrossRef]
  45. Huguenin, F. M., de Almeida, V. A. R., Rodrigues, M. V. F., Ferreira, M. E. C., & Morgado, F. F. d. R. (2024). Body image of university students: A systematic review of the characteristics of interventions. Psicologia: Reflexao e Critica, 37, 1. [Google Scholar] [CrossRef] [PubMed]
  46. Kavvadas, D., Kavvada, A., Karachrysafi, S., Papaliagkas, V., Chatzidimitriou, M., & Papamitsou, T. (2023). Stress, anxiety, and depression levels among university students: Three years from the beginning of the pandemic. Clinics and Practice, 13(3), 596–609. [Google Scholar] [CrossRef] [PubMed]
  47. Kay, S. A. (2016). Emotion regulation and resilience: Overlooked connections. Industrial and Organizational Psychology, 9(2), 411–415. [Google Scholar] [CrossRef]
  48. Kraft, B., Bø, R., Jonassen, R., Heeren, A., Ulset, V. S., Stiles, T. C., & Landrø, N. I. (2023). The association between depression symptoms and reduced executive functioning is primarily linked by fatigue. Psychiatry Research Communications, 3(2), 100120. [Google Scholar] [CrossRef]
  49. Larsen, L. T. (2021). Nor merely the absence of disease: A genealogy of the WHO’s positive health definition. History of the Human Sciences, 35(1), 111–131. [Google Scholar] [CrossRef]
  50. Lauritzen, S. L. (1996). Graphical models. Oxford Statistical Science Series. Oxford University Press. [Google Scholar]
  51. Li, W., Zhao, Z., Chen, D., Peng, Y., & Lu, Z. (2022). Prevalence and associated factors of depression and anxiety symptoms among college students: A systematic review and meta-analysis. Journal of Child Psychology and Psychiatry, 63(11), 1222–1230. [Google Scholar] [CrossRef]
  52. Linden, B., Ecclestone, A., & Stuart, H. (2022). A scoping review and evaluation of instruments used to measure resilience among post-secondary students. SSM—Population Health, 19, 101227. [Google Scholar] [CrossRef]
  53. Lipson, S. K., Zhou, S., Abelson, S., Heinze, J., Jirsa, M., Morigney, J., Patterson, A., Singh, M., & Eisenberg, D. (2022). Trends in college student mental health and help-seeking by race/ethnicity: Findings from the national Healthy Minds Study, 2013–2021. Journal of Affective Disorders, 306, 138–147. [Google Scholar] [CrossRef]
  54. Lombardo, P., Jones, W., Wang, L., Shen, X., & Goldner, E. M. (2018). The fundamental association between mental health and life satisfaction: Results from successive waves of a Canadian national survey. BMC Public Health, 18(1), 342. [Google Scholar] [CrossRef]
  55. Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335–343. [Google Scholar] [CrossRef]
  56. Lv, W., Qiu, H., Lu, H., Yajuan, Z., Yongjie, M., Xing, C., & Zhu, X. (2024). Moderating effect of negative emotion differentiation in chronic stress and fatigue among Chinese employees. Frontiers in Psychology, 15, 1358097. [Google Scholar] [CrossRef] [PubMed]
  57. Manzano-García, G., & Ayala, J. C. C. (2013). Psychometric properties of Connor-Davidson Resilience Scale in a Spanish sample of entrepreneurs. Psicothema, 25(2), 245–251. [Google Scholar] [CrossRef] [PubMed]
  58. Mayungbo, O. (2016). Agreeableness, conscientiousness and subjective wellbeing. People: International Journal of Social Sciences, 2(3), 68–87. [Google Scholar] [CrossRef]
  59. Menefee, D. S., Ledoux, T., & Johnston, C. A. (2022). The importance of emotional regulation in mental health. American Journal of Lifestyle Medicine, 16(1), 28–31. [Google Scholar] [CrossRef]
  60. Meng, J., Wang, X., Wei, D., & Qiu, J. (2020). State loneliness is associated with emotional hypervigilance in daily life: A network analysis. Personality and Individual Differences, 165, 110154. [Google Scholar] [CrossRef]
  61. Mestre, J. M., Núñez-Lozano, J. M., Gómez-Molinero, R., Zayas, A., & Guil, R. (2017). Emotion regulation ability and resilience in a sample of adolescents from a suburban area. Frontiers in Psychology, 8, 1980. [Google Scholar] [CrossRef]
  62. Moeller, R. W., Seehuus, M., & Peisch, V. (2020). Emotional intelligence, belongingness, and mental health in college students. Frontiers in Psychology, 11, 499794. [Google Scholar] [CrossRef]
  63. Nakhostin-Khayyat, M., Borjali, M., Zeinali, M., Fardi, D., & Montazeri, A. (2024). The relationship between self-regulation, cognitive flexibility, and resilience among students: A structural equation modeling. BMC Psychology, 12(1), 337. [Google Scholar] [CrossRef]
  64. Osborn, T. G., Li, S., Saunders, R., & Fonagy, P. (2022). University students’ use of mental health services: A systematic review and meta-analysis. International Journal of Mental Health Systems, 16, 57. [Google Scholar] [CrossRef]
  65. Paoloni, P. V. R. (2014). Emotions in academic contexts. Theoretical perspectives and implications for educational practice in college. Electronic Journal of Research in Educational Psychology, 12(3), 567–596. [Google Scholar] [CrossRef]
  66. Peralta, V., Gil-Berrozpe, G. J., Sánchez-Torres, A., & Cuesta, M. J. (2020). The network and dimensionality structure of affective psychoses: An exploratory graph analysis approach. Journal of Affective Disorders, 277, 182–191. [Google Scholar] [CrossRef]
  67. Pidgeon, A. M., Rowe, N. F., Stapleton, P., Magyar, H. B., & Lo, B. C. Y. (2014). Examining characteristics of resilience among university students: An international study. Open Journal of Social Sciences, 2(11), 14–22. [Google Scholar] [CrossRef]
  68. Pierce, S., Gould, D., & Camiré, M. (2017). Definition and model of life skills transfer. International Review of Sport and Exercise Psychology, 10(1), 186–211. [Google Scholar] [CrossRef]
  69. Pons, P., & Latapy, M. (2005). Computing communities in large networks using random walks (long version). arXiv. [Google Scholar] [CrossRef]
  70. Rama, L., & Sarada, S. (2017). Role of self-esteem and self-efficacy on competence—A conceptual framework. IOSR Journal of Humanities and Social Science, 22(2), 33–39. [Google Scholar] [CrossRef]
  71. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton University Press. [Google Scholar] [CrossRef]
  72. Ruggeri, K., Garcia-Garzon, E., Maguire, Á., Matz, S., & Huppert, F. A. (2020). Well-being is more than happiness and life satisfaction: A multidimensional analysis of 21 countries. Health and Quality of Life Outcomes, 18(1), 192. [Google Scholar] [CrossRef]
  73. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78. [Google Scholar] [CrossRef]
  74. Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness. The Guilford Press. [Google Scholar] [CrossRef]
  75. Sadeghi, A., Ofoghi, N., & Azizi, S. (2015). Relationship between students’ personality and mental health at university of guilan (faculty of humanities). Health, 7(07), 896–901. [Google Scholar] [CrossRef]
  76. Salgado-Lévano, C. (2012). Medición de la resiliencia en adolescentes y adultos [Resilience Measure in adolescents and adults]. Temáticas Psicológicas, 8(1), 21–27. [Google Scholar] [CrossRef]
  77. Savitri, J., Kiswantomo, H., & Tambun, G. N. (2023). The role of self-esteem and life satisfaction on university students’ engagement. Journal An-Nafs: Kajian Penelitian Psikologi, 8(2), 249–263. [Google Scholar] [CrossRef]
  78. Savoji, A. P., & Ganji, K. (2013). Increasing Mental Health of University Students through Life Skills Training (LST). Procedia—Social and Behavioral Sciences, 84, 1255–1259. [Google Scholar] [CrossRef]
  79. Sharma, G., & Yukhymenko-Lescroart, M. A. (2024). Life purpose as a predictor of resilience and persistence in college students during the COVID-19 pandemic. Journal of College Student Retention: Research, Theory and Practice, 26(2), 334–354. [Google Scholar] [CrossRef]
  80. Sheldon, E., Simmonds-Buckley, M., Bone, C., Mascarenhas, T., Chan, N., Wincott, M., Gleeson, H., Sow, K., Hind, D., & Barkham, M. (2021). Prevalence and risk factors for mental health problems in university undergraduate students: A systematic review with meta-analysis. Journal of Affective Disorders, 287, 282–292. [Google Scholar] [CrossRef] [PubMed]
  81. Shichiri, K., Shibuya, M., Watanabe, M., Tahashi, M., Kaminushi, K., Uenoyama, T., Mashima, I., Murayama, K., Kuroda, T., & Suzuki, Y. (2016). Correlations between the Profile of Mood States (POMS) and the WHOQOL-26 among Japanese University Students. Health, 8(5), 416–420. [Google Scholar] [CrossRef]
  82. Shih, C. T., & Chuang, C. H. (2013). Individual differences, psychological contract breach, and organizational citizenship behavior: A moderated mediation study. Asia Pacific Journal of Management, 30(1), 191–210. [Google Scholar] [CrossRef]
  83. Shorter, E. (2008). History of psychiatry. Current Opinion in Psychiatry, 21(6), 593–597. [Google Scholar] [CrossRef]
  84. Song, X., & Hu, Q. (2024). The relationship between Freshman students’ mental health and academic achievement: Chain mediating effect of learning adaptation and academic self-efficacy. BMC Public Health, 24(1), 3207. [Google Scholar] [CrossRef]
  85. Suhaimi, A. F., Ahmad, N., & Kamaruzaman, H. (2024). Examining the resilience of university students: A comparative mental health study. Cureus, 16(9), e69293. [Google Scholar] [CrossRef]
  86. Tamplain, P., Webster, E. K., Brian, A., & Valentini, N. C. (2020). Assessment of motor development in childhood: Contemporary issues, considerations, and future directions. Journal of Motor Learning and Development, 8(2), 391–409. [Google Scholar] [CrossRef]
  87. Tibshirani, R. (1996). Regression Shrinkage and Selection via the Lasso. Journal of the Royal Statistical Society Series B: Statistical Methodology, 58(1), 267–288. [Google Scholar] [CrossRef]
  88. Ümmet, D. (2015). Self esteem among college students: A study of satisfaction of basic psychological needs and some variables. Procedia—Social and Behavioral Sciences, 174, 1623–1629. [Google Scholar] [CrossRef]
  89. Van de Leur, J. C. (2024). Psychological treatment of stress-induced exhaustion disorder. In Digital comprehensive summaries of uppsala dissertations from the faculty of social sciences (Vol. 223, pp. 1–144). Universitatis Upsliensis. [Google Scholar]
  90. Vidović, S., Kotromanović, S., & Pogorelić, Z. (2024). Depression, anxiety, and stress symptoms among students in Croatia during the COVID-19 pandemic: A systematic review. Journal of Clinical Medicine, 13(20), 6240. [Google Scholar] [CrossRef] [PubMed]
  91. Wong, S. M. Y., Chen, E. Y. H., Lee, M. C. Y., Suen, Y. N., & Hui, C. L. M. (2023). Rumination as a transdiagnostic phenomenon in the 21st century: The flow model of rumination. Brain Sciences, 13(7), 1041. [Google Scholar] [CrossRef] [PubMed]
  92. World Health Organization. (2022). Mental health. Health Topics. Available online: https://www.who.int/health-topics/mental-health#tab=tab_1 (accessed on 5 May 2025).
  93. Yamanaka, T., Yamagishi, N., Nawa, N. E., & Anderson, S. J. (2021). Assessing changes in mood state in university students following short-term study abroad. PLoS ONE, 16(12), e0261762. [Google Scholar] [CrossRef]
  94. Yang, T., Guo, Y., Cheng, Y., & Zhang, Y. (2022). Effects of traditional Chinese fitness exercises on negative emotions and sleep disorders in college students: A systematic review and meta-analysis. Frontiers in Psychology, 13, 908041. [Google Scholar] [CrossRef]
  95. Yong, C., Nor Zainudin, Z., Mohd Anuar, M. A., & Wan Othman, W. N. (2022). Personality traits and their effects among university students in Malaysia: A systematic review. International Journal of Academic Research in Business and Social Sciences, 12(10), 1147–1162. [Google Scholar] [CrossRef]
  96. Yoshihara, K., Hiramoto, T., Sudo, N., & Kubo, C. (2011). Profile of mood states and stress-related biochemical indices in long-term yoga practitioners. BioPsychoSocial Medicine, 5, 6. [Google Scholar] [CrossRef]
Youth 05 00116 g001
Figure 1. Dimensionality structure of MH using Exploratory Graph Analysis for Final Model. Note: Nodes represent retained items; edges are regularized partial correlations (thicker = stronger). Community colors denote five dimensions (personality, mood, resilience, self-esteem, distress). Denser connections between distress and mood indicate shared activation under academic demands; self-esteem clusters with competence-related items.
Figure 1. Dimensionality structure of MH using Exploratory Graph Analysis for Final Model. Note: Nodes represent retained items; edges are regularized partial correlations (thicker = stronger). Community colors denote five dimensions (personality, mood, resilience, self-esteem, distress). Denser connections between distress and mood indicate shared activation under academic demands; self-esteem clusters with competence-related items.
Youth 05 00116 g001
Youth 05 00116 g002
Figure 2. Stability of Network Connections. Note: Edge stability under BootEGA. Black points = bootstrapped mean edge weights; red points = original sample edge weights. The pronounced overlap and S-shaped pattern indicate high edge stability.
Figure 2. Stability of Network Connections. Note: Edge stability under BootEGA. Black points = bootstrapped mean edge weights; red points = original sample edge weights. The pronounced overlap and S-shaped pattern indicate high edge stability.
Youth 05 00116 g002
Youth 05 00116 g003
Figure 3. Stability of Model Metrics. Betweenness indexes a node’s bridging potential between communities; closeness reflects its average proximity to all nodes; strength is the sum of absolute edge weights attached to a node; and expected influence extends strength by considering the signed impact a node exerts on its neighbors.
Figure 3. Stability of Model Metrics. Betweenness indexes a node’s bridging potential between communities; closeness reflects its average proximity to all nodes; strength is the sum of absolute edge weights attached to a node; and expected influence extends strength by considering the signed impact a node exerts on its neighbors.
Youth 05 00116 g003
Table 1. Means, Standard Deviations, and Confidence Intervals of the Sample Variables.
Table 1. Means, Standard Deviations, and Confidence Intervals of the Sample Variables.
VariablesMeanSDCI 95%
Psychological distress
Depression13.8510.0813.20–14.50
Anxiety13.4910.1712.83–14.14
Stress17.769.1017.10–18.34
Self-esteem30.185.2529.84–30.52
Resilience
Persistence21.142.8620.96–21.32
Control17.332.6617.76–17.50
Adaptability13.421.8913.30–13.54
Future Purpose7.721.547.65–7.82
Spirituality4.401.204.32–4.48
Dimensions of personality
Extraversion13.305.3912.96–13.63
Agreeableness29.325.3929.67–28.97
Conscientiousness16.605.6416.24–16.97
Neuroticism13.807.2013.33–14.26
Openness33.343.9933.08–33.60
Mood States
Anger5.921.995.61–5.86
Tension9.821.779.14–9.36
Depression5.661.856.26–6.48
Fatigue6.031.696.09–6.31
Vigor12.191.8313.40–13.63
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

García-Pérez, L.; Cepero-González, M.; Mota, J. Rethinking Mental Health Assessment: A Network-Based Approach to Understanding University Students’ Well-Being with Exploratory Graph Analysis. Youth 2025, 5, 116. https://doi.org/10.3390/youth5040116

AMA Style

García-Pérez L, Cepero-González M, Mota J. Rethinking Mental Health Assessment: A Network-Based Approach to Understanding University Students’ Well-Being with Exploratory Graph Analysis. Youth. 2025; 5(4):116. https://doi.org/10.3390/youth5040116

Chicago/Turabian Style

García-Pérez, Laura, Mar Cepero-González, and Jorge Mota. 2025. "Rethinking Mental Health Assessment: A Network-Based Approach to Understanding University Students’ Well-Being with Exploratory Graph Analysis" Youth 5, no. 4: 116. https://doi.org/10.3390/youth5040116

APA Style

García-Pérez, L., Cepero-González, M., & Mota, J. (2025). Rethinking Mental Health Assessment: A Network-Based Approach to Understanding University Students’ Well-Being with Exploratory Graph Analysis. Youth, 5(4), 116. https://doi.org/10.3390/youth5040116

Article Metrics

Back to TopTop