Mental well-being and depression among university students in Palestine under the war

Abstract

Background

Since October 2023, the West Bank has experienced an “invisible war” characterized by intensified military raids, movement restrictions, and economic instability. While global attention focuses on the destruction in Gaza, the impact of these “attritional” war conditions on the mental health of West Bank university students remains under-researched. This study aims to assess the prevalence of depressive symptoms and mental well-being among this population and to model the bio-psychosocial determinants of these outcomes.

Methods

A quantitative cross-sectional study was conducted in November 2024 among university students (N = 497) across the North, Middle, and South West Bank. Data were collected using a digital survey comprising the Beck Depression Inventory-Short Form (BDI-SF), the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS), and a customized bio-psychosocial stressor inventory. Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to analyze the relationships between environmental, familial, individual, social, and university factors and mental health outcomes.

Results

The results revealed a profound “Flourishing Gap”: while 36% of students exhibited moderate-to-severe depressive symptoms, a staggering 93% failed to reach the threshold for high mental well-being, with only 7% classified as flourishing. Structural modeling explained 63% of the variance in depression and 40% in well-being. Individual factors were the strongest risk factor for depression (beta = 0.45), while Social factors served as the most significant protective buffer, strongly predicting higher well-being (beta = 0.73) and lower depression. University factors also positively influenced well-being (beta = 0.47).

Conclusions

Palestinian university students are navigating a mental health crisis characterized not only by high depression but by a mass extinction of positive psychological functioning. However, the university environment and social support systems act as critical sanctuaries of resilience. Interventions must move beyond pathology to strengthen these collective social scaffolds and protect higher education institutions as essential psychosocial infrastructure.

Introduction

The field of Conflict Epidemiology has increasingly identified higher education institutions not merely as collateral casualties of war, but as active theaters of systemic disruption. In conflict zones globally, university students represent a uniquely vulnerable demographic, forced to navigate a “dual burden”: the rigorous cognitive demands of academic performance and the chronic, physiological imperatives of survival [1]. Unlike general populations, students must maintain high-level executive functioning, attention, memory, and complex reasoning while inhabiting environments characterized by extreme unpredictability and threat [2]. This intersection creates a distinct epidemiological profile where educational persistence becomes inextricably linked to psychological endurance. However, the existing literature on student mental health often presupposes a stable baseline environment, failing to account for contexts where the very infrastructure of learning is targeted.

This phenomenon, increasingly defined in academic discourse as scholasticide, refers to the systematic destruction of educational centers and the erasure of indigenous knowledge systems [3, 4]. It is within this global framework of educational fragility that the current status of Palestinian higher education must be examined [5].

Regionally, nevertheless, some research found that major barriers to mental health include pervasive social stigma, insufficient training for healthcare professionals, and limited access to evidence-based interventions [6].

While global attention since October 2023 has predominantly focused on the kinetic destruction in the Gaza Strip, where all twelve universities have been damaged or destroyed [1, 7]. The West Bank has experienced a parallel, albeit distinct, escalation of violence that constitutes an “invisible war.” This environment differs from the direct aerial bombardment of Gaza, presenting instead as a matrix of “environmental deprivation” and pervasive control [8]. Since the onset of the war, West Bank students have faced intensified military raids (including incursions into campuses like Birzeit and An-Najah), severe movement restrictions via an expanded checkpoint apparatus, and economic strangulation resulting from the withholding of public sector salaries ([8–9]). For example, Nursing students in the seventh of October war in Palestine undergo elevated levels of violence and deprived quality of life during the war, particularly in environmental and psychological dimensions [10]. These structural realities act as potent environmental stressors that disrupt daily academic functioning. Furthermore, the psychological landscape is complicated by “secondary trauma” and “vicarious trauma,” as students relentlessly consume digital media depicting the devastation in Gaza [11, 12]. This exposure generates a state of vicarious distress, where the boundaries between the direct victims of war and the observers blur, exacerbating the psychological load on students who are physically distant but emotionally proximate to the violence [13].

To analyze the impact of these conditions, this study adopts the Bio-Psychosocial Model as its primary theoretical lens [14, 15]. This framework rejects the reductionist view of mental health as solely a biological or intra-psychic phenomenon. Instead, it posits that psychological outcomes in conflict zones are the product of complex interactions between macro-level political determinants (occupation, war conditions), meso-level institutional structures (university closures, family economic instability), and micro-level individual responses (resilience, anxiety) [16]. In the current West Bank context, macro-level stressors filter down to deregulate the biological systems of students manifesting in sleep disruption and hyperarousal which in turn impairs psychological well-being [11]. By utilizing this multi-layered approach, the study asserts that observed mental health outcomes are often rational responses to abnormal environmental determinants rather than indicators of inherent biological vulnerability [17].

Despite the severity of the situation, a critical tripartite gap exists in the current literature. First, the temporal gap: the vast majority of studies regarding Palestinian mental health predate October 2023. These earlier datasets fail to capture the intensity of the current “war conditions,” rendering previous baselines obsolete. Second, a methodological gap compromises the validity of standard assessment tools in this population [17]. Traditional depression inventories, such as the full BDI-II, frequently conflate “somatic symptoms” such as fatigue, weight loss, and sleep disturbance with psychopathology [18, 19]. In the West Bank, these symptoms are often consequences of environmental deprivation (e.g., food insecurity, noise pollution from raids, micronutrient deficiencies) rather than cognitive depression [20]. To address this, this study employs a “metric transformation,” utilizing the BDI-SF (Short Form) to strictly isolate cognitive-affective distress from somatic confounding. Third, the “functioning” gap: prevailing research tends to focus exclusively on pathology, such as PTSD. This study argues for the necessity of measuring eudaimonic well-being the ability to function and find meaning despite adversity [21]. By integrating the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS), this research shifts the focus from mere symptom counting to an assessment of functional resilience [22].

Consequently, the primary objective of this study is to assess the prevalence of depressive symptoms and levels of mental well-being among university students in the West Bank of Palestine during the active war conditions of November 2024. The study employs a Quantitative Cross-Sectional design, analyzing data via Partial Least Squares Structural Equation Modeling (PLS-SEM) to infer predictive pathways between environmental stressors and mental health outcomes. Furthermore, data were collected across the North, Middle, and South distinct governorates to capture the “heterogeneity” of the war experience, acknowledging that the intensity of military incursions and economic hardship varies significantly across the fragmented geography of the West Bank [1, 23]. Through this rigorous analytical approach, the study aims to provide a clinically precise evidence base for urgent psychosocial and academic interventions.

Methods

This study adopts a quantitative, cross-sectional research design to investigate the prevalence of depression and the status of mental well-being among university students in the West Bank, Palestine.

The selection of a cross-sectional approach is predicated on several critical methodological and practical imperatives inherent to conflict epidemiology. First, cross-sectional designs provide a rapid, simultaneous assessment of both exposure variables (e.g., war-related stressors, economic instability) and outcome variables (depression, well-being) at a single point in time. In the context of the West Bank, where the political and security situation is subject to rapid and unpredictable fluctuation ranging from movement restrictions to active military incursions longitudinal designs face prohibitive risks regarding participant attrition and the safety of data collection personnel.

Furthermore, this design facilitates the immediate modeling of structural relationships between environmental determinants and psychological outcomes. By capturing a “snapshot” of the student population during the active war conditions in November 2024, the study aims to provide actionable data that can inform immediate psychosocial interventions, rather than waiting for the extended timeline of longitudinal observation. The design is descriptive in its capacity to quantify the burden of mental health disorders and analytical in its use of Structural Equation Modeling (SEM) to infer complex associative pathways between the independent and dependent variables.

Theoretical framework: the bio-psychosocial model in conflict

The methodology is underpinned by a bio-psychosocial framework that posits mental health outcomes as the product of complex interactions between biological vulnerabilities, psychological resilience, and, crucially, social and environmental determinants. In the context of the occupied Palestinian territories, the “environmental” component is heavily weighted by political violence, “siege” mentalities, and structural restrictions on movement and opportunity.

To operationalize this framework, the study’s independent variables are structured into a multi-dimensional model comprising five distinct categories: Individual Factors, Familial Factors, Social Factors, University Factors, and Environmental Factors. This categorization allows for a granular analysis of how macro-level stressors (war, occupation) filter down through meso-level structures (university, family) to impact micro-level psychological functioning (individual well-being).

Study population and setting

The target population for this research consists of the aggregate body of university students currently enrolled in higher education institutions within the West Bank, Palestine. According to the most recent data from the Ministry of Education and Higher Education (MEHE) for the academic year 2022/2023, this population comprises approximately 94,372 students attending traditional, regular universities [24]. This demographic is of particular strategic interest for mental health research.

The study was conducted across eight major universities located in the West Bank. The selection of institutions was designed to ensure geographic representativeness, covering the three primary administrative and social regions of the territory:

• The North West Bank: Including institutions in Nablus and Jenin, areas frequently subject to intense military raids and closures.

• The Middle West Bank: Including Ramallah and Jerusalem peripheries (e.g., Al-Quds University), characterizing the administrative center but also the flashpoints of the separation wall.

• The South West Bank: Including Hebron and Bethlehem (e.g., Hebron University, Palestine Ahliya University, Bethlehem University), areas with unique challenges related to settler proximity and movement restrictions.

This geographic dispersion is methodologically vital. The experience of “war conditions” in the West Bank is not uniform; it varies significantly by locality depending on the density of checkpoints, the proximity of settlements, and the frequency of military operations. By sampling from all three regions, the study aims to capture the heterogeneity of the Palestinian student experience.

Sampling strategy and sample size determination

Given the logistical constraints imposed by the security situation where physical access to campuses is often interrupted by strikes, checkpoints, or incursions a convenience sampling technique was employed. The questionnaire was distributed digitally via university official channels (emails from administration) and student-led social media groups.

While probability sampling is theoretically superior for generalization, convenience sampling is the standard pragmatic choice in conflict zones where complete, accessible sampling frames are often compromised, and physical safety is a priority.

To mitigate the selection bias inherent in non-probability sampling, the distribution strategy was aggressive and multi-channel, aiming to reach students from diverse faculties, years of study, and socio-economic backgrounds across the identified universities. As university students, the target population possesses high digital literacy and reliable access to official university communication channels, ensuring broad reach. Duplications were prohibited by allowing the electronic form to be filled only once, and by reviewing and clearing the data post collection.

To ensure the statistical validity of the findings and the robustness of the Structural Equation Modeling (SEM) analysis, a rigorous sample size calculation was conducted. Using the Raosoft sample size calculator [25], the following parameters were applied based on the total population (N) of 94,372 students:

• Confidence level: 95% (Z = 1.96), ensuring that the results represent the true population parameters with high reliability.

• Margin of error: 5% (E = 0.05), providing a standard level of precision.

• Response distribution: 50% (p = 0.5), a conservative estimate used to maximize the required sample size when the true prevalence of the outcome (depression/well-being levels) is unknown.

Based on these inputs, the minimum required sample size was calculated to be 383 participants.

The data collection phase yielded a total of 497 valid responses. This final sample size significantly exceeds the minimum requirement of 383. The inclusion of these additional respondents (an oversampling of approximately 30%) serves to increase the statistical power of the study (1- beta), reducing the likelihood of Type II errors (false negatives).

Furthermore, larger sample sizes are particularly beneficial for PLS-SEM analysis, which relies on bootstrapping techniques to determine significance. A larger sample helps stabilize the parameter estimates and allows for more reliable testing of complex structural paths between multiple latent variables. The sample was confirmed to be representative of the regional distribution (North, Middle, South) of the West Bank universities, enhancing the external validity of the findings.

Even though sampling bias in such a technique is still present, it may tend toward a self-selection bias. Specifically, it could lead to an underestimation of psychological distress if students experiencing the highest levels of depression were less motivated to engage with the digital survey, or conversely, an overestimation if those most affected by the current security situation felt a greater incentive to report their experiences.

Instrumentation and measures

The study employed a structured questionnaire divided into two primary segments. The first segment assessed independent variables (stressors), while the second utilized two internationally validated psychometric scales to measure dependent variables: The Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS) [26] and the Beck Depression Inventory-Short Form (BDI-SF) [27].

To operationalize the bio-psychosocial framework, a specific measurement tool was developed or adapted to assess the determinants of mental health in this specific context. This section comprised 25 questions, categorized into five distinct groups (5 questions per group):

1. Individual Factors (IF): Items assessing personal psychological assets, such as resilience, self-efficacy, and internal locus of control. These factors mediate how external stressors are processed.

2. Familial Factors (FF): Items evaluating the home environment, including family support, parental pressure, and household economic stability a critical factor given the economic strangulation often accompanying war conditions.

3. Social Factors (SF): Items measuring peer support, social integration, and conversely, social isolation caused by movement restrictions or societal fragmentation.

4. University Factors (UF): Items assessing the academic environment, including workload, availability of psychosocial support facilities on campus, and the flexibility of the institution in responding to crisis-related interruptions.

5. Environmental Factors (EF): Items specifically targeting the war context. These questions likely probe exposure to political violence, difficulties at checkpoints, fear of arrest, and general feelings of insecurity and unpredictability regarding the future.

While traditional psychology often focuses solely on pathology (e.g., depression), this study incorporates the SWEMWBS to measure positive psychological functioning, aligning with the World Health Organization’s definition of health as more than merely the absence of disease. The SWEMWBS is a 7-item shortened version of the original 14-item WEMWBS. It was selected over the longer version because the 7 items relate more specifically to functioning (eudaimonic well-being e.g., thinking clearly, dealing with problems) rather than just feeling (hedonic well-being e.g., feeling interested or cheerful). In a war zone, where “happiness” may be scarce or transient, the capacity to function and maintain psychological competence is a more resilient and relevant indicator of mental health.

This scale consists of seven positively worded statements. Respondents rate their experience over the past two weeks on a 5-point Likert scale, where the Scoring involves raw score calculation for parametric statistical analyses like regression and SEM.

It has demonstrated high internal consistency (Cronbach’s alpha > 0.80) in various populations, including adolescents and adults in the UK, Norway, and Denmark [28–30]. The scale has been translated and validated for Arabic-speaking populations.

Studies involving undergraduate students in Gulf countries (Oman, Qatar, Saudi Arabia) reported a Cronbach’s alpha of 0.77, indicating good reliability, and confirmed its construct validity through confirmatory factor analysis [31]. This supports its applicability to the Palestinian student population. Also see Table 3 under data analysis section for more details on the tool reliability.

Table 3.

Assessment of the measurement model construct reliability and validity

	Cronbach’s alpha (α)	Composite reliability (rho / ρ)	Average variance extracted (AVE)
BDI	0.891	0.909	0.436
MWB	0.767	0.834	0.419
EF	0.680	0.551	0.258
FF	0.174	0.398	0.297
IF	0.330	0.504	0.342
SF	0.041	0.416	0.409
UF	0.306	0.151	0.281

Thresholds typically recommended are alpha > 0.7, rho > 0.7, and AVE > 0.5

Depressive symptoms were evaluated using the Beck Depression Inventory-Short Form (BDI-SF), a 13-item self-report scale developed by Beck and Beck (1972), each rated on a scale of 0 to 3 based on symptom severity. The BDI-SF focuses on the cognitive and affective components of depression: pessimism, sense of failure, self-dislike, guilt, and indecisiveness. This ensures that the scores reflect genuine psychological distress rather than physical hardship. The Arabic version of the Beck Depression Inventory has been extensively validated by Alansari (2006) reporting high internal consistency (Cronbach’s alpha = 0.85 to 0.91) and confirming its cross-cultural validity. Also, the instrument has been validated for use in the Palestinian territories by researchers such as R. Bseiso, A. Thabet [32] who reported a Cronbach’s alpha of 0.86.

Data collection and analysis

Data collection was executed over a four-week period in November 2024. This timing is significant, as it corresponds to a period of heightened conflict, ensuring the data reflects the immediate impact of the “war conditions” referenced in the research title. But due to the mobility restrictions and safety hazards inherent to the war context, traditional face-to-face data collection was deemed unsafe and logistically unfeasible. Instead, a digital ethnography approach was adopted using online Google Form. The survey link was distributed through two primary verified channels to ensure it reached the target population; university officials disseminated the link via internal student email systems, lending institutional legitimacy to the request, and sharing the link via official university social media groups and student union pages, utilizing the high connectivity of Palestinian youth to maximize reach across the fragmented territories. To ensure data integrity, the Google Form settings were configured to mandate responses and technical constraints were employed to prevent multiple submissions from the same user.

Inclusion criteria included any university student in any of the Palestinian universities participated in the study. Exclusion criteria meant to avoid any student who is not an active student, which was ensured via targeting students through universities emails and students’ groups.

The analytical framework for this study is bifurcated into descriptive analysis using IBM SPSS Statistics (Version 25), including data cleaning, demographic profiling and prevalence calculation, and inferential structural modeling using SmartPLS 4 to test the study’s hypotheses and model the complex relationships between the independent and dependent variables, by means of Partial Least Squares Structural Equation Modeling (PLS-SEM).

Measurement model assessment (reliability and validity) was done before testing relationships to verify the quality of the measurement instruments. This included indicator reliability by examining outer loadings. Internal consistency reliability assessed using Cronbach’s Alpha and composite reliability (rho_c). Convergent validity evaluated using the Average Variance Extracted (AVE). Discriminant validity assessed to ensure constructs are distinct. The primary metric used is the Heterotrait-Monotrait Ratio (HTMT) of correlations. The Fornell-Larcker criterion is also checked as a secondary measure.

Structural model assessment (hypothesis testing) done once the measures are validated, the causal paths are analyzed. Collinearity assessment by checking The Variance Inflation Factor (VIF) for all predictor constructs. Path coefficients and significance to test the strength of the relationships between variables is measured by the path coefficient. Statistical significance is determined using bootstrapping to generate t-statistics and p-values. Coefficient of determination (R2) done to measure the model’s predictive accuracy, indicating the proportion of variance in the dependent variables (depression, well-being) explained by the combined independent variables. Effect size (f2) to measures the impact of a specific independent variable on a dependent variable also done.

The proposed structural model was specified to evaluate the simultaneous influence of the five exogenous latent variables Environmental (EF), Familial (FF), Individual (IF), Social (SF), and University (UF) factors on two endogenous outcomes: Mental Well-being (MWB) and Depression (BDI). As illustrated in the path diagram, each latent construct was measured by multiple observed indicators, such as the seven items assessing MWB and the thirteen items assessing BDI. The structural model hypothesized direct paths from all five determinants to both mental health outcomes, allowing for a comparative analysis of how these factors differentially predict positive well-being versus depressive symptoms (Fig. 1).

Fig. 1.

Structural model of the study variables. (key: EF: Environmental Factors; FF: Familial Factors; IF: Individual Factors; SF: Social Factors; UF: University Factors; BDI: Beck Depression Inventory (Depression); MWB: Mental Well-being)

The results of the structural model robustness are illustrated in Table 1. Cramér-von Mises test indicated significant deviations from normality for the majority of the constructs (p < 0.001), including BDI, EF, FF, IF, SF, and UF. Specifically, BDI exhibited a notable positive skewness (0.985) and kurtosis (0.910), reflecting a distribution where the majority of students’ cluster at lower to moderate levels of symptom severity, with a significant tail extending toward severe psychopathology.

Table 1.

Descriptive statistics for the standardized latent variables, skewness, kurtosis, and normality assessments (n = 497)

	Mean	Median	Observed min	Observed max	SD	Excess kurtosis	Skewness	Cramér-von Mises p-value
BDI	-0.00	-0.13	-1.33	3.86	1.00	0.910	0.985	0.000
MWB	0.00	0.03	-2.87	2.34	1.00	-0.283	-0.223	0.155
EF	0.54	0.67	-0.17	1.49	0.43	-0.857	0.054	0.000
FF	-0.09	-0.41	-0.43	1.16	0.40	0.266	1.018	0.000
IF	0.26	0.00	-0.15	1.26	0.44	-0.543	0.961	0.000
SF	0.52	0.61	-0.45	1.01	0.46	-0.970	-0.618	0.000
UF	-0.01	-0.00	-0.61	1.13	0.45	-0.849	0.135	0.000

All values represent standardized latent variable scores

EF Environmental Factors, FF Familial Factors, IF Individual Factors, SF Social Factors, UF University Factors, BDI Beck Depression Inventory, MWB Mental Well-being

Similarly, other Factors displayed high positive skewness (e.g. FF = 1.018 and IF = 0.961), suggesting that high levels of these stressors are present in a subset of the population rather than uniformly distributed. In contrast, MWB was the unique exception, demonstrating a relatively normal distribution (Skewness= -0.223; Kurtosis= -0.283) with a non-significant Cramér-von Mises test (p = 0.16). The pervasive lack of normality across the independent variables and the primary outcome of depression empirically validates the study’s methodological choice to employ Partial Least Squares Structural Equation Modeling (PLS-SEM), given the algorithm’s robustness in handling non-normal data distributions without the need for transformation.

The predictive power of the structural model was evaluated using the coefficient of determination (R²), which measures the proportion of variance in the dependent variables explained by the exogenous bio-psychosocial constructs. The results in Table 2 indicate that the model possesses a robust explanatory capacity, particularly regarding psychopathology. The model explained 62% of the variance in BDI.

Table 2.

Predictive power of the structural model using the coefficient of determination R-square

	R2 (adjusted)
BDI	0.622
MWB	0.391

BDI Beck Depression Inventory MWB Mental Well-being

R² values of 0.75, 0.50, and 0.25 are considered substantial, moderate, and weak, respectively

According to standard PLS-SEM criteria, this represents a moderate-to-substantial predictive accuracy, suggesting that the selected stressors (Environmental, Familial, Individual, etc.) are highly effective predictors of depressive symptoms among the Palestinian students in the current war context. Conversely, the model explained 39% of the variance in MWB, within the moderate range, indicating that while the bio-psychosocial framework significantly accounts for positive functioning, other unmeasured factors (maybe spiritual or community-specific resilience mechanisms) may also play a role in sustaining well-being.

The evaluation of the measurement model in Table 3 revealed distinct psychometric profiles between the established outcome measures and the context-specific stressor determinants. The internationally validated instruments demonstrated robust reliability; the Beck Depression Inventory-Short Form (BDI-SF) achieved high internal consistency (α = 0.89; ρ = 0.91), and the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS) similarly exhibited satisfactory reliability (α = 0.77; ρ = 0.83). However, the convergent validity for these constructs, as measured by the Average Variance Extracted (AVE), was slightly below the strict 0.5 threshold (0.44 and 0.42, respectively), though often deemed acceptable given their high composite reliability. Conversely, the five stressor constructs (EF, FF, IF, SF, UF) displayed lower reliability metrics, with Cronbach’s alpha values below 0.7 and AVE values below 0.5 across all five independent variables. This variability suggests that while the “Stressor Scale” captures the breadth of the students’ war experience, the individual items within these factors (e.g., SF, UF) are highly heterogeneous and may represent formative indices of stress rather than tightly correlated reflective constructs.

To ensure that the bio-psychosocial constructs represented empirically distinct phenomena, discriminant validity was assessed using the Heterotrait-Monotrait Ratio (HTMT) of correlations Table 4. The analysis confirmed that all HTMT values remained below the conservative threshold of 0.85. While these values indicate a high degree of correlation reflecting the conceptual relatedness of mental health outcomes and the overlapping nature of personal and domestic stressors, they do not exceed the critical limit. Consequently, the results confirm that the discriminant validity of the measurement model is established; specifically, the scale successfully differentiates between the presence of depressive symptoms and the absence of positive well-being, and distinctly categorizes the various sources of stress (environmental, social, institutional) within the student population (Table 4).

Table 4.

Assessment of the measurement model discriminant validity

	BDI	EF	FF	IF	MWB	SF
BDI
EF	0.263
FF	0.548	0.271
IF	0.748	0.340	0.809
MWB	0.821	0.237	0.542	0.606
SF	0.743	0.244	0.613	0.733	0.708
UF	0.749	0.479	0.482	0.749	0.701	0.750

Values < 0.85 (strict) or < 0.90 (liberal) indicate that discriminant validity is established (i.e., the constructs are statistically distinct)

HTMT Heterotrait-Monotrait Ratio

A regression analysis was conducted to examine the influence of BDI scores on MWB. The Analysis of Variance confirmed that the regression model significantly predicts Mental Well-being scores, indicating a good fit for the data (p < 0.001). Regression assumptions were met, with a Durbin-Watson statistic of 2.05 indicating independent residuals and Variance Inflation Factor (VIF) values of 1.0 indicating no multicollinearity. As well, a multiple linear regression analysis was conducted to predict BDI scores based on EF, FF, IF, SF, and UF. Preliminary analyses confirmed that the assumptions of the regression model were met; the Durbin-Watson statistic was 2.000, indicating independence of residuals, and VIF values ranged from 1.12 to 1.48, indicating no multicollinearity among the predictors. The overall model was statistically significant (F = 110, p < 0.001, and explained 53% of the variance in depression scores (R²= 0.529).

Ethical considerations and safety protocol

Conducting mental health research in a conflict zone imposes a “duty of care” that exceeds standard ethical requirements. The researchers implemented a rigorous ethical protocol to protect the safety and well-being of the participants. Ethical clearance was formally granted by the Institutional Review Board (IRB) at Palestine Ahliya University. Additionally, administrative permissions were secured from the participating universities to allow access to student communication channels. Upon accessing the Google Form, participants were presented with a comprehensive informed consent statement. For the use of the study measurement tools, the authors obtained the needed permissions. The WEMWBS is free to use but is copyrighted to NHS Health Scotland and the universities of Warwick and Edinburgh. The BDI is publicly available for use.

A critical ethical dilemma in depression research is the identification of participants at immediate risk (e.g., those scoring in the “Severe” range). To address this, the study implemented a specific intervention protocol including contacting those with severe depression scores, via emails ( that are kept in a password-protected file) by the principal investigator directly with a message informing them (sensitively) that their scores indicated a high level of distress and strongly recommended they seek support. The email provided specific referral pathways, advising them to contact their university’s psychosocial counseling unit or a private psychotherapist of their preference. This protocol ensures that the research adheres to the ethical principle of beneficence, actively serving the well-being of the most vulnerable participants rather than merely extracting data from them.

Results

Descriptive analyses were conducted to determine the prevalence of different levels of Mental Well-being (MWB) within the study sample (N = 497). The results indicated that the vast majority of participants reported either moderate or low levels of well-being. Specifically, the largest proportion of the sample fell into the ‘Moderate’ category (n = 236, 47%), closely followed by the ‘Low’ well-being category (n = 228, 46%). Only a small minority of participants demonstrated ‘High’ levels of mental well-being, representing just 7% of the total sample (n = 33), as seen in Fig. 2.

Fig. 2.

Pie chart of the prevalence of mental well-being levels within the study sample (N = 497)

Descriptive analyses were performed using the Beck Depression Inventory-Short Form to assess the prevalence of depressive symptoms among participants (N = 497). The results indicated that the majority of the sample (n = 316, 64%) reported ‘Mild/Normal’ levels of depressive symptoms. However, a substantial portion of the participants (36%) exhibited clinically relevant symptoms. specifically, 23% of the sample (n = 115) fell into the ‘Moderate’ depression category, while 13% (n = 66) were classified as having ‘Severe’ depressive symptoms (Fig. 3).

Fig. 3.

Pie chart of the prevalence of depressive symptoms among the study sample (N = 497)

The structural model was assessed to examine the simultaneous relationships between the five exogenous latent factors, Depression (BDI), and Mental Well-being (MWB).

The model demonstrated strong explanatory power, accounting for 63% of the variance in BDI scores (R² = 0.626) and 40% of the variance in MWB scores (R² = 0.397). The analysis identified IF as the strongest risk factor for BDI (β = 0.45), while SF emerged as the strongest protective factor (β = -0.44). Regarding MWB, SF exerted the strongest positive direct effect on MWB (β = 0.73), followed by UF (β = 0.47). Conversely, FF (β = -0.48), EF (β = -0.16) and IF (β = -0.14) displayed negative direct paths to MWB in the adjusted model. Furthermore, the model revealed a direct negative influence between BDI and MWB (β = -0.39) as seen in Fig. 4.

Fig. 4.

Explanatory powers and effects of latent variables correlations in the study structural model paths. (key: EF: Environmental Factors; FF: Familial Factors; IF: Individual Factors; SF: Social Factors; UF: University Factors; BDI: Beck Depression Inventory (Depression); MWB: Mental Well-being)

To visualize the relative magnitude of the structural relationships, a bar chart of the path coefficients was generated (Fig. 5). The chart highlights the significant positive predictors within the model. SF exhibited the most substantial positive influence on MWB, demonstrating a dominant effect compared to other variables. IF displayed the strongest positive path coefficient toward BDI, identifying them as the primary risk factor in this analysis.

Fig. 5.

Visualization of the relative magnitude of the structural relationships. (key: EF: Environmental Factors; FF: Familial Factors; IF: Individual Factors; SF: Social Factors; UF: University Factors; BDI: Beck Depression Inventory (Depression); MWB: Mental Well-being)

The assessment of latent variables correlations revealed significant associations between the bio-psychosocial determinants and mental health outcomes, elucidating the complex dynamics of Palestinian university students’ well-being under war conditions (Table 5). A strong inverse relationship was observed between BDI and MWB (r= -0.69), confirming that while these constructs are distinct, high levels of depressive symptomatology strongly erode positive psychological functioning. Among the independent variables, SF emerged as the most potent protective buffer, demonstrating the strongest negative correlation with BDI (r= -0.61) and the highest positive association with MWB (r = 0.55), suggesting that social cohesion is the primary resilience resource for the Palestinian students. Other factors also displayed a substantial protective role, significantly correlating with lower depression and higher well-being such as UF (r = -0.53 and 0.45, respectively). Conversely, IF were identified as the strongest risk factor for psychopathology, exhibiting the highest positive correlation with Depression (r = 0.53), followed by FF (r = 0.43).

Table 5.

Latent variables correlations in the study structural model

	EF	FF	IF	SF	UF
BDI	0.332	0.425	0.534	-0.609	-0.528
MWB	-0.252	-0.405	-0.382	0.545	0.454
EF	1	0.164	0.231	-0.222	-0.294
FF	0.164	1	0.408	-0.371	-0.250
IF	0.231	0.408	1	-0.438	-0.374
SF	-0.222	-0.371	-0.438	1	0.457
UF	-0.294	-0.250	-0.374	0.457	1

EF Environmental Factors FF Familial Factors, IF Individual Factors, SF Social Factors, UF University Factors, BDI Beck Depression Inventory (Depression), MWB Mental Well-being (SWEMWBS)

Interestingly, while EF representing the direct context of war and occupation were positively associated with Depression (r = 0.33) and negatively with Well-being (r= -0.25), their direct correlational impact was weaker than that of individual and familial stressors, implying that the macro-political situation may impact mental health largely by exacerbating micro-level personal and domestic vulnerabilities (Table 5).

The results indicated a strong negative relationship between the two constructs, with BDI significantly predicting lower levels of MWB (β= -0.65, SE = 0.034, p < 0.001). This model accounted for a substantial proportion of the variance in the outcome, explaining 43% of the variation in mental well-being scores (R² = 0.43) as seen in Fig. 6.

Fig. 6.

Regression model for the relationship between depression and mental well-being

A multiple regression analysis was also conducted to examine the influence of the independent factors on BDI scores. The overall model explained 53% of the variance in depression scores (R² = 0.529). All five predictors were statistically significant (p < 0.001). The analysis revealed distinct patterns of influence among the predictors. SF emerged as the strongest negative predictor (β= -0.33), followed by UF (β= -0.23).

Conversely, IF demonstrated the strongest positive association with depression (β = 0.21), followed by FF (β = 0.13) and EF (β = 0.12). Moreover, A multiple regression analysis was conducted to examine the influence of five independent factors on MWB scores. The model accounted for 36% of the variance in MWB (R²= 0.356). SF emerged as the strongest positive predictor (β = 0.31), followed by UF (β = 0.22), indicating that higher satisfaction in these areas is associated with better mental well-being. Conversely, FF showed the strongest negative association with MWB (β= -0.17) suggesting that higher stress or issues in these domains may detract from mental well-being. However, neither EF (β= -0.07, p = 0.08) nor IF (β= -0.05, p = 0.24) reached statistical significance in this model.

Discussion

The pivotal finding of this investigation extends beyond the mere prevalence of psychopathology to reveal a profound asymmetry in the psychological landscape of Palestinian university students. The data indicates that while 36% of the student population in the West Bank meets the clinical criteria for moderate-to-severe depression during the active war conditions of late 2024, a staggering 93% fail to reach the threshold for “high mental well-being” as measured by the Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS). Only a negligible minority 7% of the sample can be classified as “flourishing.” This disparity defines what we term the “Flourishing Gap”; a chasm between the absence of acute clinical depression and the presence of functional psychological health.

In the context of the West Bank’s “invisible war”, this finding suggests that the absence of severe pathology does not equate to health. Instead, the vast majority of the student population inhabits a state of “languishing”; a condition characterized by the stagnation of cognitive and emotional potential, where individuals are suspended in a survival mode that precludes the deep engagement, optimism, and executive functioning required for higher education. The clinical weight of this finding is immense; it implies that while one-third of the student body requires immediate psychiatric intervention for depression, nearly the entire generation requires broad-spectrum psychosocial support to restore functional capacity.

To understand the severity of this current profile, it is imperative to contrast these findings with pre-war baselines. Prior to the escalation of hostilities in October 2023, epidemiological studies utilizing the Beck Depression Inventory and similar instruments among Palestinian university students typically reported moderate-to-severe depression prevalence rates ranging between 23% and 40% [33]. For instance, a study of medical students in the Arab region prior to 2023 identified a pooled prevalence of approximately 33.1%, with Palestinian cohorts often scoring at the higher end of this spectrum due to the chronic baseline of occupation [34].

However, the current data reveals a qualitative shift. While the prevalence of depression (36%) remains within the upper bounds of historical highs, the erosion of positive well-being appears to have accelerated. The “War Effect” is thus not solely characterized by a spike in new depressive cases but by a reduced resilience capacity. The flourishing rate of 7% stands in stark contrast to global benchmarks; for example, normative data from university populations in the United Kingdom typically categorizes 15% to 24% of students as having “high mental well-being,” even amidst the stressors of academic life [35]. The West Bank cohort is therefore navigating a distinct epidemiological reality where the “Dual Burden” of academic rigor and existential threat has systematically decimated the psychological reserves necessary for thriving.

Comparing these results with other active conflict zones elucidates the specific impact of the West Bank’s “chronic structural violence”. In Ukraine, following the full-scale invasion in 2022, studies reported depression rates among university students ranging from 33.6% to as high as 85.8%, depending on proximity to the frontline and the measurement tools used (e.g., PHQ-9 vs. DASS-21) [36]. The Ukrainian context, characterized by high-intensity kinetic warfare, aerial bombardment, and massive internal displacement, generates an acute trauma response often manifesting as immediate, severe symptomatology.

In contrast, the West Bank experience in late 2024 is defined by “environmental deprivation” and “attritional violence” a matrix of military raids, movement restrictions, and economic strangulation rather than carpet bombing [37]. This chronic, grinding stress produces a distinct profile where depressive symptoms are characterized less by the panic of acute shock and more by a pervasive hopelessness, cognitive exhaustion, and elevated distress burden. This structural violence is insidious; it does not always produce the dramatic spikes in PTSD seen in kinetic zones like Gaza or eastern Ukraine, but it steadily erodes the bio-psychosocial substrates of well-being, resulting in the alarmingly low flourishing rates observed in this study. The “Flourishing Gap” is thus the signature wound of the invisible war.

While global attention has focused on the visible destruction in Gaza where the systematic obliteration of the education system has been termed “scholasticide” the West Bank has been subjected to a parallel, “invisible” campaign of systemic disruption. Our structural model’s Environmental Factors (EF) construct captures this matrix, which operates through specific mechanisms of control that target the continuity of daily life and education.

The term “scholasticide” in the West Bank context refers not to the total demolition of buildings, as in Gaza, but to the processual dismantling of the educational environment. This includes intensified military incursions into university campuses. For instance, in November 2024, Israeli occupation forces raided Birzeit University, arresting students and damaging university property, a continuation of a policy that has seen over 150 students detained from that institution alone [38]. Similarly, raids on An-Najah National University and the surrounding areas of Nablus and Jenin have turned the commute to campus into a life-threatening navigation of checkpoints and military operations [39].

These incursions are not merely security operations; they are “environmental stressors” that disrupt the circadian and psychological rhythms of students. The unpredictability of these events creates a state of hypervigilance. Students cannot plan their study schedules, sleep is fragmented by night raids, and the sanctuary of the campus is violated. This constant disruption of routine is a potent driver of executive dysfunction, impairing the cognitive processes attention, memory, and planning essential for academic success [40].

A critical component of the “Environmental Factor” in this study is the severe economic crisis gripping the West Bank, driven by the withholding of Palestinian Authority (PA) tax revenues by Israel. Since October 2023, the PA has been forced to pay civil servants a demographic that includes many of the students’ parents only 50% to 70% of their salaries [41]. This “salary crisis” is not an abstract macroeconomic statistic; it is a direct determinant of student distress.

The economic contraction, estimated at nearly 20–30% of the West Bank’s GDP in 2024 [23], filters down to the individual student as food insecurity, inability to pay tuition, and the psychological burden of witnessing parental helplessness. In our structural model, the negative path from Familial Factors (FF) to Mental Well-being (β= -0.48) and the positive correlation with Depression (r = 0.43) strongly reflect this economic reality. The “Environmental” stressor of salary withholding mediates through the “Familial” domain, creating a domestic atmosphere of anxiety that suffocates the student’s capacity to flourish.

The most powerful finding in our structural model is the dominant protective role of Social Factors (SF). Social Factors emerged as the strongest positive predictor of Mental Well-being (β = 0.73) and the most potent negative predictor of Depression (β= -0.33). This finding fundamentally challenges Western, individualistic models of mental health that locate resilience in personal traits like “grit” or “self-efficacy”.

In the Palestinian context, resilience is structural and collective. This aligns with Social Cure Theory, which posits that shared social identities such as identifying with a university cohort, a political cause, or a tight-knit family unit provide a “psychological buffer” that absorbs the shock of external stressors [42]. When state institutions collapse and the economy fails, the “social scaffolding” of the community remains the last line of defense. The strong negative correlation between Social Factors and Depression (r= -0.61) indicates that social isolation is the most dangerous risk factor for these students. Conversely, the maintenance of social bonds acts as collective coping mechanisms, protecting the individual from the full toxic impact of the war.

To interpret this statistical finding effectively, we must integrate the indigenous Palestinian concept of Sumud (steadfastness) [43]. Sumud is not passive endurance; it is a collective, active psycho-political stance of “rootedness”. It reframes the act of daily survival going to class despite checkpoints, rebuilding after raids as an act of political resistance.

Our data effectively quantifies Sumud. The high path coefficient from Social Factors to Well-being suggests that students derive deep psychological meaning from their shared struggle. The “we” protects the “I”. This “collective effervescence” transforms potential trauma into a shared narrative of endurance [42]. This has critical implications for intervention: therapeutic approaches that focus solely on the individual (e.g., individual CBT) may be less effective than group-based interventions that validate and strengthen this collective identity. Resilience, in this context, is a community resource, not an individual trait.

A superficial reading of our results might highlight a paradox. Environmental Factors (EF) representing the war itself showed a weaker direct path coefficient to depression (β = 0.12) and well-being (β= -0.16) compared to Individual (IF) and Social (SF) factors. One might erroneously conclude that the war is a secondary factor. However, this interpretation ignores the complex causal architecture of conflict epidemiology. We argue that Environmental Factors function as Distal Stressors that exert their primary toxic effects by mediating through Proximal Stressors (Familial and Individual factors)³⁴. The war is not a discrete event; it is the “background radiation” that destabilizes the entire ecosystem of the student’s life.

Our structural model supports a mediation hypothesis consistent with the Minority Stress Model and Bio-Ecological Systems Theory:

1. Distal stressor (Macro-Level): The occupation policies, salary withholdings, and military raids constitute the Environmental Factor.

2. Mediating mechanism (Meso-Level): These environmental shocks degrade the Familial Factor (FF). The raid on a campus disrupts the University Factor (UF).

3. Proximal outcome (Micro-Level): The collapse of these meso-level support systems depletes the student’s individual coping reserves (IF), leading to depression. Individual Factors were the strongest direct risk for depression (β = 0.45), but these individual vulnerabilities are largely the downstream product of environmental toxicity. Thus, the environment is the “cause of the causes”. The statistical model captures the direct hit of the war, but its true magnitude is revealed in how it dismantles the family and the self.

Despite the systemic attacks on education, University Factors (UF) emerged as a significant positive predictor of Mental Well-being (β = 0.47) and a protective factor against depression. This suggests that the university functions as a “Psychosocial Sanctuary”. In a landscape of chaos, the university offers structure, routine, and a future-oriented identity. To be a student is to claim a future; to study is to defy the erasure of one’s potential. The campus provides a physical and social space where Sumud is enacted. Even when learning is disrupted, the institution remains a symbol of continuity. This finding aligns with research on “Schools as Zones of Peace” and the protective role of education in emergencies [44].

The concept of Scholasticide is critical here. The military raids on Birzeit and An-Najah universities [38], the arrests of student leaders, and the physical damage to campuses are not collateral damage; they are strategic attempts to dismantle this sanctuary [45]. By attacking the university, the occupation targets the primary generator of Palestinian resilience and future leadership.

Strengths and limitations

The primary strength of this study is its methodological alignment with the conflict context. The use of PLS-SEM was empirically justified by the non-normal distribution of the data (e.g., the high skewness of depression scores). Unlike standard regression, PLS-SEM is robust to non-normality and allows for the modeling of complex, mediated pathways between latent variables, providing a more accurate map of the “bio-psychosocial” terrain. Additionally, the use of the BDI-SF ensures that our findings are not inflated by the physical hardships of war. By excluding somatic items, we have isolated the cognitive core of depression, providing a rigorous evidence base that distinguishes between “feeling sick” (somatic) and “feeling hopeless” (cognitive) [46].

The study is limited by its cross-sectional design. The “War Effect” is dynamic; a snapshot in November 2024 cannot capture the fluctuations in distress that track with the intensity of military operations. Convenience sampling also introduces potential bias, although the large sample size (n = 497) and geographic stratification across the West Bank mitigate this. Finally, while we model mediation, we cannot strictly prove causality without longitudinal data.

Finally, the use of digital data collection may introduce self-selection bias; students experiencing the most acute distress or those in areas with severed internet connectivity due to military operations may be underrepresented.

Conclusion

This study provides empirical evidence of the mental health crisis in West Bank of Palestine universities during the 2024 war conditions. It reveals a generation inhabiting a “Flourishing Gap” surviving the trauma but denied the psychological conditions necessary to thrive. The “invisible war” of structural violence manifested in checkpoints, raids, and economic strangulation has permeated the bio-psychosocial fabric of student life.

However, the data also highlights the profound resilience of this population. The protective power of Social Factors and the University demonstrates that Palestinian students are not passive victims; they are actively engaged in a collective project of Sumud. They draw strength from each other and from the institution of education itself.

The implications are clear: protecting Palestinian higher education is not merely an academic concern; it is a mental health imperative. The international community must recognize universities as essential psychosocial infrastructure and act to halt the “scholasticide” that threatens them. Internally, universities must adopt policies of leniency and integrate evidence-based psychosocial support to ensure that this generation does not merely survive the war, but retains the capacity to build the future.

Authors' contributions

M.K designed the study and performed the statistical analysis, and reviewed the manuscript. A.K, M.A, A.O and I.S assisted with data collection and contributed to the writing of the manuscript. All authors read and approved the final manuscript.

Funding

The authors received no specific funding for this work.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to the sensitive nature of the data and the need to protect participant privacy in a conflict setting, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board (IRB) at Palestine Ahliya University. All participants provided informed consent prior to accessing the online questionnaire. The study adhered to the ethical standards of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.