The Sustainability of E-Learning in UAE Higher Education: Digital Transformation, Inequality, and Student Well-Being in the Time Crisis

Abstract

The COVID-19 pandemic accelerated the global shift to e-learning, raising concerns about its sustainability, student well-being, and educational inequality. This study evaluates e-learning in higher education during crises by examining its psychological, behavioral, social, and academic impacts on university students in the United Arab Emirates over two academic years of remote learning. Using a mixed-methods approach, data were collected from two cohorts (n = 228): Group 1 (G1, n = 76; 2020–2021) and Group 2 (G2, n = 152; 2021–2022). Analysis included descriptive statistics, independent sample t-tests, and thematic analysis. The results revealed significant differences between groups across most domains (p < 0.001). G2 reported higher psychological distress, including increased depression, stress, and reduced focus, while G1 demonstrated stronger behavioral and social adaptation, such as better self-care, family communication, and engagement in hobbies and sports. Regression analysis showed a strong linear relationship between online and campus grade distributions (R² = 0.7862), indicating academic consistency across learning modes. However, the findings highlight a sustainability paradox: although e-learning enhances flexibility and access and reduces environmental impact, prolonged reliance is linked to psychological strain, behavioral risks, and widening social inequality. The study underscores the need for a resilient and sustainable education model that supports students academically, psychologically, and socially to ensure the well-being and public health of all. These insights are particularly relevant amid ongoing regional crises and the continued expansion of e-learning and generative AI in education. Aligning with sustainable education goals, such approaches contribute to SDGs 3, 4, and 10, and support broader progress toward the 2030 Agenda.

1. Introduction

A global health crisis declared in March 2020 caused significant socio-economic disruptions worldwide. As of 1 February 2026, overall totals reached 779 million confirmed cases and 7.11 million deaths [1,2,3]. Lockdown measures reduced transmission through restricting social interactions, including educational activities, affecting 1.6 billion students across 192 countries [4,5]. This crisis posed serious challenges to achieving the UN’s Sustainable Development Goals (SDGs), with universities playing a key role in equipping future generations to advance the UN Agenda 2030 [6]. The universities shifted rapidly to online education, often unprepared for such a digital transition [7,8]. This shift exposed gaps in teaching strategies, creating uncertainties for higher education institutions (HEIs) [9,10,11]. Early studies highlighted negative effects on students’ well-being, learning, and psychological health during lockdowns [12,13,14]. Most research has examined the impact of e-learning in technical disciplines such as engineering, medicine, pharmacy, and computer science; investigations in the social sciences remain relatively limited [15,16,17,18,19]. Nevertheless, the United Nations Sustainable Development Goals (SDGs) have highlighted the principle of “Leaving No One Behind” as central to Sustainable Development [20]. From a developmental perspective, “people are the real wealth of a nation”, in which “education” and “health” are crucial for human development [21]. Therefore, education should be the key to improving public health and reducing social inequality outcomes through disease prevention knowledge [22].

The pandemic has revealed how health crises can disrupt education, threatening SDG 3 “healthy lives and well-being” SDG 4, “inclusive quality education for all,” and SDG 10, “reduce inequality” [20]. Therefore, education is considered crucial for achieving the SDGs by 2030 [23]. While the pandemic could threaten the achievement of those goals, it particularly impacts health and education, revealing a significant research gap regarding its dual effect on these vital sectors [24]. Research on the comprehensive impacts of e-learning on the psychological, behavioral, social, and academic performance of university social sciences students in the UAE during the pandemic remains limited, often addressing these factors in isolation [25,26,27].

From a sustainability perspective, this disparity is especially noticeable when connecting “health”, “education”, and “inequality,” posing obstacles to reaching the SDGs by 2030. By measuring the psychological, behavioral, social, and academic impacts of e-learning and digital transformation on university students during the pandemic and analyzing how educational strategies align with SDGs (3, 4, and 10), this study attempts to assess the sustainability of e-learning based on a multidisciplinary approach [28].

Specifically, the study seeks to investigate how the e-learning environment influenced students’ well-being, equality, and academic performance; shaped their preferences for post-pandemic education; and how future strategies can contribute to advancing SDGs (3, 4, 10) to assure the “well-being” and good public health for all and support the broader SDGs by 2030. Based on this context, the study is founded on a systems-oriented conceptualization of sustainability, in which outcomes are not assessed via direct SDG metrics but through evidence-based indicators widely recognized in scientific research [4]. Specifically, sustainability is operationalized via student well-being (SDG 3), learning conditions (SDG 4), and perceived inequality (SDG 10), reflecting the understanding that complex social systems are best examined through intermediate determinants. Within this framework, the study does not measure SDG attainment per se; rather, it analyses how e-learning environments mediate key mechanisms theoretically linked to these goals. Therefore, the main question for this study is: Do e-learning models adopted during the pandemic contribute to the long-term sustainability and resilience of education systems, with specific reference to SDGs 3, 4, and 10?

To respond to this key question, this study seeks to answer the following:

Q1—To what extent did e-learning impact students’ psychological, behavioral, and social well-being during the pandemic?

Q2—How did the transition to e-learning during the COVID-19 crisis impact academic performance, and were there statistically significant differences in grades between students in e-learning and on-campus learning environments?

Q3—What are the students’ perceptions and preferences for post-pandemic in-person, e-learning, or hybrid modalities?

By addressing these questions, the research aims to provide actionable insights to generate holistic, actionable recommendations for policymakers and educators in fostering resilient and sustainable education systems that can withstand future crises and foster sustainable, high-quality education to meet the UN SDGs by 2030 to enhance students’ academic performance, reduce inequalities, and ensure well-being and sustainability.

1.1. E-Learning Impacts During the Pandemic

The term “impact” indicates positive or negative effects on individuals or organizations [29]. The concept of “impact” is based on the long-term primary and secondary outcomes of activities [30]. It is influencing knowledge, skills, behavior, health, or living conditions with lasting effects across economic, socio-cultural, and environmental domains [31]. This study explores the “impact” through the lens of both expected and unexpected changes at personal, community, and global levels, focusing on how the pandemic and the e-learning environment influenced students’ well-being, academic performance, and inequalities. Meanwhile, “learning” represents a crucial adaptation process, including “e-learning,” which uses digital tools and media [32].

However, e-learning is defined as the use of multimedia and internet technologies in education to enhance access to resources and facilitate remote collaboration [33,34]. Although its rapid implementation during public health crises has yielded mixed outcomes, some studies have underscored its potential to improve educational quality and flexibility, complementing and enriching traditional learning in HEIs. Before the pandemic, studies highlighted the advantages and disadvantages of e-learning, indicating that its effectiveness varies throughout global crises [35]. Flexibility, cost-effectiveness, broader access, and inclusivity are effectively emphasized [36]. In contrast, e-learning is considered comparable to traditional education and prefers synchronous hybrid models for engagement and versatility [37]. Yet, e-learning and the long-term effects were often neglected, with some viewing the pandemic as an opportunity to “reinvent pedagogy” [38]. The social isolation measures were crucial for public health, causing stress and emotional issues [39,40]. Some studies have found that lifestyle modifications are associated with greater depression symptoms [41,42]. However, students and academics experienced stress from extended screen time and the risk of technology addiction, which affected both mental health and public health [43,44,45,46]. The overuse of technology produced physical health problems such as insomnia and obesity [47]. Therefore, quantitative research is recommended to evaluate social impacts and adaptive educational practices [48].

1.2. The Interactions Between Psychological, Behavioral, and Social Impacts

Studying the psychological impacts focused on stress and life satisfaction, Selye outlined the “General Adaptation Syndrome” by dividing stress physiology into three stages of “alarm, resistance, and exhaustion,” suggesting that prolonged stress may increase illness risk and differentiating between motivating eustress and detrimental distress [49,50]. Stress is described as a threat to well-being that triggers adaptive behaviors, leading to dysfunction if adaptations fail [51]. While emotions are categorized as influencing behavior, with a mix of negative and positive impacts, chronic stress is commonly linked to health issues and diminished well-being [52]. Prolonged stress could enhance physical and emotional disorders [53]. Additionally, the Social Impact Theory (SIT) explains how others’ presence or actions affect an individual’s emotions and behaviors, focusing on the factors of “strength,” “proximity,” and “number of people involved” [54]. Social impact assessments affect various domains, such as economics, the environment, and psychology [55]. These theories highlight the need for in-depth investigation. As a result, this study focused on students’ perceptions of e-learning and its psychological, behavioral, social, and academic performance impacts. Research from Saudi, Egyptian, and Jordanian universities linked e-learning with mental health and academic challenges, including anxiety, sleep disturbances, and stress [14,56,57,58]. Similar impacts were seen in Europe and the United States [59,60]. Recent studies have shown limited quantitative data, while the United Arab Emirates (UAE) lacks extensive impact assessments and overlooks broader social and academic challenges and pedagogical adjustments during crises [61,62]. The importance of addressing physical and psychosocial well-being in educational environment settings is clearly recommended [63].

1.3. Crisis Management Strategies and E-Learning Experience in the UAE

The UAE reported its first cases on 23 January 2020. The lockdown from 20 March to 31 July 2020 saw 889,797 cases and 2302 deaths, with a 9.5% infection rate and a low 0.3% death rate, exceeding many developed countries [4,64,65,66]. Measures included suspending activities, halting flights, e-learning, and remote work. In response to the pandemic, the UAE education system rapidly shifted to online learning on 22 March 2020, to ensure continuity [67]. Nonetheless, the UAE recorded visible progress toward SDG 4, reaching 99.7%, 87.5% for SDG 3, and 99% for SDG 10 [68]. In 2023, the University of Sharjah ranked first nationally and placed in the global top 251–300 [69]. The UOS’s “Prioritizing Interim Strategy” (2020–2021) emphasized sustainable teaching and safety, integrating “Blackboard Ultra” to increase annual enrollment from 571 to 1105 students, ultimately reaching 19,240 by the 2023–2024 academic year [70]. However, ongoing debates about the effectiveness of e-learning have highlighted challenges in content delivery and student engagement [71,72].

2. Materials and Methods

2.1. Study Design and Sustainability Framework

This study employs a comparative observational approach, enabling students to acquire scientific inquiry, problem-solving, decision-making, and creative skills [32]. Within this method, students actively engage with the psychological, behavioral, and social components of e-learning, therefore contributing to the generation of practical knowledge [73,74]. In line with the concept of sustainability, the study further assesses the sustainability of e-learning systems throughout the pandemic and lockdown circumstances by analyzing their efficacy, flexibility, and long-term implications for inclusive and resilient education. Emphasis is placed on achieving Sustainable Development Goals (SDGs) 3 (good health and well-being), 4 (quality education), and 10 (reduced inequalities). Mixed methods (quantitative and qualitative) were employed to capture multidimensional impacts, behavioral dynamics, and future educational perceptions and preferences [75,76,77]. The analytical framework blends inductive and deductive reasoning to achieve both exploratory depth and theoretical rigor [78,79,80].

2.2. Study Context and Period

This study was conducted over the two academic years of full implementation of remote education, from September 2020 to June 2022, in the United Arab Emirates (UAE), during times of disruption, uncertainty, and reduced social interaction. Such contextual conditions thus constitute potential confounders that could exacerbate negative perceptions and, therefore, cannot be entirely separated from the e-learning experience. For this reason, the analysis adopts a contextualized framework in which a comparative blueprint between G1 (early adaptation) and G2 (prolonged exposure) captures temporal dynamics rather than causal relationships. This period is a key transformational phase in higher education, characterized by rapid digitalization and pedagogical adaptation. While temporally specific, the results are of analytical significance, providing longitudinal insight into student adaptation and informing sustainability and resilience in post-transition and new scenarios, such as those following the most recent changes in March–April 2026 related to continued war in the region.

2.3. Participants and Sampling Strategy

The study sample includes 228 students from the College of Arts, Humanities, and Social Sciences at the University of Sharjah (UOS) in the UAE, all engaged in fully online education via the “Blackboard platform”. To reduce potential bias, random independent sampling was used within each group to the greatest extent practical. There are structural variations between Group 1 and Group 2 (for example, academic level, course setting, sample size, duration of exposure to online education, and kind of evaluation). Importantly, the study does not aim to establish exact causal relationships; rather, it seeks to capture and understand context-dependent differences in students’ experiences at various stages of e-learning throughout the pandemic. Participants were presented in two groups to enable temporal and comparative analyses [45,77].

-

Group 1 (G1; n = 76): Social and Cultural Change class (academic year 2020–2021).

-

Group 2 (G2; n = 152): Two Social Psychology classes (academic year 2021–2022).

This model of group-based design enables examination of changes in students’ experiences, attitudes, and behavioral responses across different stages of e-learning implementation, thereby contributing to the evaluation of long-term sustainability.

2.4. Data Collection Procedures

Data collected using standardized online techniques were integrated into the “Blackboard learning system,” which is a Software with Platform (a Learning Management System - LMS) employed in the UOS, and hosted by an e-learning technology company based in Florida, USA. Students’ participation was optional, and they were given standard instructions describing the study objectives and answer criteria. Quantitative and qualitative data were gathered using the “free association” corpus [74,75,76] based on frequency-based lexical analysis. This method was selected for its capacity to elicit spontaneous cognitive and emotional responses. This approach enables a more nuanced examination of the psychological, behavioral, and social dimensions of the e-learning experience, extending beyond the constraints of structured survey instruments. Participants responded to four open-ended questions based on their “lived experiences”, addressing the following:

(a)

What are the psychological, behavioral, and social impacts of the e-learning environment (both positive and negative)?

(b)

What is the impact of e-learning during the COVID-19 crisis on academic performance?

(c)

What are your perceptions of the current educational process, and what are your expectations for the post-COVID-19 phase?

(d)

Do you prefer online, on-campus, or hybrid learning?

The data were collected from G1 and G2 for Q1, Q2, and Q3. However, data for the fourth question (post-pandemic preferences) were only gathered from G2 (n = 152), which represented students in the second year of e-learning under more stable circumstances. Some of them were in their first year in the university, and others were in their second, third, or fourth year. G1 was excluded from this stage because the question was inappropriate during the early stages of e-learning (first year), when the pandemic trajectory was uncertain.

2.5. Data Analysis Tools

A thematic analysis (TA) is used to facilitate the extraction of potential underlying patterns and chain effects across psychological, behavior and social domains, as well as secondary effects on academic performance and long-term well-being perceived by students [79,80,81]. All entries were coded by a coder who traced each response and organized them into 40 classified substantial themes. Frequency counts were then generated to ensure the robustness and transparency of the analysis. To reduce errors and improve quality, all data input was double-checked. This thorough verification method increases the reliability, consistency, and validity of the qualitative dataset’s conclusions. In the first stage, descriptive quantitative statistics (mean, median, standard deviation, and frequency distributions) were carried out using “Statistical Package for the Social Sciences-SPSS” (version 29). In the second stage, a visual method was developed to facilitate a holistic assessment of the relative intensity and the multidimensional effects of e-learning over time. A radar (spider) and polar mapping tools, supplied with the “Microsoft Excel Software 365 (the 2023 version)” are employed. Radar charts were created to compare aggregated scores across psychological, behavioral, and social variables over two time periods (G1/G2), facilitating the detection of spreading patterns and imbalances between positive and negative impacts. Together, these visualization methods aided comparative and longitudinal research by highlighting structural variations in impact profiles and allowing the detection of cumulative negative effects associated with extended e-learning exposure.

In this context, the “independent sample t-test” was used to investigate differences between G1 and G2 based on key psychological, behavioral, social, and academic variables retrieved from the first stage of the study to evaluate the sustainability of e-learning. Recognizing that statistical significance is insufficient to reflect the whole impact, quantitative results were combined with qualitative findings to create a holistic, sustainability-oriented assessment of the data [78]. In the second stage, the study is constructed on institutional information obtained through the “Blackboard learning system”, and students’ final course marks to measure change in students’ performance. The grade distribution analyses for the online and campus exams were conducted using both exponential and linear regression models created with the Microsoft Excel Software 365 (the 2023 version). In the third stage, for analyzing students’ perspectives, a qualitative approach is used for G1 and G2 and a quantitative method for G2, respectively, to investigate students’ preferences for the post-pandemic, ignoring G1, because this subject was not under discussion in the pandemic’s first year.

2.6. Ethical Considerations

At the time of data collection (2020–2022), The formal ethical approval was not necessary for social science research at the institutional level. Ethical clearance requirements were restricted to medical research involving human beings’ tissues. The ethical approval for social sciences was initiated institutionally in 2023. However, the study was conducted in accordance with the Declaration of Helsinki protocol. Participants were anonymized using coded identities (RE1–RE76 for G1 and RE77–RE228 for G2) to maintain confidentiality. Verbal informed consent was obtained from the participants. All participants were informed that their participation was optional, and they were guaranteed confidentiality and anonymity. The data from these surveys are intended to help improve educational procedures and strategies.

3. Results

The findings are organized into four key parts: (a) demographic characteristics, (b) descriptive positives and negatives psychological, behavioral, and social impacts of e-learning environment, (c) independent sample t-test for G1 and G2 to measure the different impacts on students during the time according to different variables, (d) the correlation between e-learning and academic performance, and (e) students’ preference and perceptions for the post-pandemic era.

3.1. Demographic Characteristics

Table 1. The demographic characteristics of the participants in G1 and G2.

	Variables	G1	G2	Total	G1 (Mean)	G2 (Mean)	G1 Std. Deviation	G2 Std. Deviation	Combined (Mean)	Std. Deviation
Gender	Female	56	92	148
	Male	20	60	80	1.26	1.39	0.443	0.490	1.35	0.478
Specialization	Communication	30	63	93
	Sociology	46	89	135	1.61	1.59	0.492	0.494	1.59	0.493
Branch *	Sharjah	44	76	120
	Al Dhaid	11	35	46
	Khorfakkan	20	36	56
	Kalba	1	5	6	1.71	1.80	0.907	0.914	1.77	0.910
Academic year	First	40	60	100
	Second	19	48	67
	Third	14	27	41
	Fourth	3	17	20	1.74	2.01	0.900	1.013	1.92	0.983
	Total %	76 33.3	152 66.7	228 100

* Beyond the university’s main campus in Sharjah, it includes three additional branches located in the cities of Khorfakkan (approximately 137 km from UOS), Kalba, and Al Dhaid (approximately 100 km from UOS). After the pandemic, Kalba, Al Dhaid, and Khorfakkan became independent universities in 2024. Source: Author.

summarizes the demographics of the 228 participants. As demonstrated, the descriptive analysis reveals significant differences in demographic and academic factors between G1 and G2. Overall, the sample is composed of (66.7%) participants from G2 and (33.3%) participants from G1, indicating a stronger representation of the second group in the dataset. Females make up most of the sample (148 out of 228). Females have somewhat higher average scores in G2 (M = 1.39, SD = 0.490) than in G1 (M = 1.26, SD = 0.443), for a total mean of 1.35 (SD = 0.478). The comparatively low standard deviations in both groups indicate that female participants responded consistently across situations. In terms of specialization, sociology students (n = 135) exceed communication students (n = 93). However, the mean values for communication students are practically equal across G1 (M = 1.61, SD = 0.492) and G2 (M = 1.59, SD = 0.494), yielding a combined mean of 1.59 (SD = 0.493).

Table 1 shows that the slight modification related to specialization has no significant impact on the observed outcomes. There are similar results for branch localization; the Sharjah campus has the biggest subgroup (n = 120), followed by Khorfakkan, Al Dhaid, and Kalba. G2 had higher mean scores (M = 1.80, SD = 0.914) than G1 (M = 1.71, SD = 0.907), for a total mean of 1.77 (SD = 0.910). However, the significantly higher standard deviations observed in this measure show greater dispersion, implying variation in experiences or performance across campuses. During the academic year, first-year students make up the biggest proportion (n = 100), followed by second-, third-, and fourth-year students. Notably, G2 has a higher mean (M = 2.01, SD = 1.013) than G1 (M = 1.74, SD = 0.900), for a total mean of 1.92 (SD = 0.983). As a result, the findings demonstrate a greater variety in G2, implying more diverse response patterns. This higher dispersion might indicate varying levels of adaptation or involvement throughout academic phases.

3.2. Descriptive Analysis: Positive and Negative Impacts of E-Learning

Table 2. The positive and negative impacts of e-learning for G1 and G2.

Variables		G1 (76)		G2 (152)		G1 & 2 (228)
		n	%	n	%	n	%
PPIs	1. Self-reliance and responsibility	17	22.4	30	19.7	47	20.6
	2. Self-confidence and the ability to confront and compete	3	3.9	9	5.9	12	11.4
	3. Getting to know the self and taking care of nature	18	23.7	8	5.3	26	5.3
	4. Feeling the value of life and family	21	27.6	2	1.3	23	10.1
	5. Health awareness and self-care	40	52.6	7	4.6	47	20.6
Total PPIs:		99	63.87	56	36.12	155	18.4
PBIs	6. Increase communication between family members	54	71.1	38	25	92	40.4
	7. Practicing sports activities	15	19.7	4	2.6	19	8.3
	8. Practicing hobbies and gaining new skills	34	44.7	14	9.2	48	21.1
	9. Our lives become more organized, calm, focused, and accomplished	20	26.3	15	9.9	35	15.4
	10. Adaptation, changing consumer behaviors, and avoiding luxuries	20	26.3	2	1.3	22	9.6
Total PBIs:		143	66.20	73	33.79	216	25.65
PSIs	11. Enhance family interaction and cohesion	49	64.5	36	23.7	85	37.3
	12. Save money, effort, and time	15	19.7	51	33.6	66	28.9
	13. Changing some customs and traditions, speeding up marriage to save costs	19	25	1	0.7	20	8.8
	14. Promoting a culture of social solidarity and adaptability	12	15.8	11	7.2	23	10.1
	15. Flexibility and facility of learning and obtaining the scientific materials	19	25	69	45.4	88	38.6
	16. Enhance relationships with colleagues, cooperate, and experience exchange	14	18.4	17	11.2	31	13.6
	17. Facility of electronic communication with professors at any time	8	10.5	26	17.1	34	14.9
	18. Getting to know a lot of colleagues through workgroups	4	5.3	21	13.8	25	11
	19. Improving technical skills and employing technology in the community	20	26.3	55	36.2	75	32.9
	20. Reducing pollution, traffic congestion, and improving the environment	17	22.4	7	4.6	24	10.5
Total PSIs:		177	37.57	294	62.42	471	55.93
Total PPIs, PBIs, PSIs		419	49.76	423	50.23	842	35.17
NPIs	21. Isolation and social distancing	35	46.1	112	73.7	147	64.5
	22. Fear of illness, failure, the future, and loss of self-confidence	34	44.7	63	41.4	97	57.5
	23. Depression and negative feelings	8	10.5	70	46.1	78	34.2
	24. Worry	17	22.4	75	49.3	92	40.4
	25. Addiction to electronic devices	3	3.9	53	34.9	56	24.6
	26. Pessimism	7	9.2	0	0	7	3.1
	27. Stress	7	9.2	50	32.9	57	25
	28. Sadness because of the death of relatives and friends	15	19.7	20	13.2	35	15.4
	29. Deprivation, unequal opportunities, restricted access to public spaces	34	44.7	38	25	72	31.6
	30. Psychological pressure, family (domestic violence), increases the burden on the student; exams have become more difficult	9	11.8	99	65.1	108	47.4
	31. Mental health, distraction, and lack of focus	2	2.6	87	57.2	89	39
Total NPIs:		171	20.40	667	79.59	838	53.99
NBIs	32. Difficulty of managing time and daily routine, conduct disorder	23	30.3	75	49.3	98	43
	33. Poor ability to organize behavioral disturbances (study, sleep, food)	4	5.3	45	29.6	49	21.5
	34. Cheating and violating the rules of conduct in the exam	0	0	39	25.7	39	17.1
	35. Laziness, boredom, indifference, and dependence on others, less (cooperation, creativity, participation, and educational motivation)	7	9.2	96	63.2	103	45.2
	36. Change in the educational environment, weak interactive relationship with professors, and the ability to accurately assess students	15	19.7	59	38.8	74	32.5
Total NBIs:		49	13.49	314	86.50	363	23.38
NSIs	37. Poor communication, social isolation, and shyness	44	57.9	101	66.4	145	63.6
	38. Technical issues, unsuitable home learning environments, reduced practical engagement, and declining academic performance	10	13.2	102	67.1	112	49.1
	39. Health problems, physical stress, weight gain, and long screen sitting	3	3.9	51	33.6	54	23.7
	40. Economic impact due to unemployment and losses in resources	26	34.2	14	9.2	40	17.5
Total NSIs:		83	23.64	268	76.35	351	22.61
Total NPIs, NBIs, NSIs		303	41.96%	1249	74.70%	1552	64.82

Source: Author.

revealed the positive and negative impacts of e-learning, focusing on three dimensions, (a) psychological, (b) behavioral, and (c) social, according to the following:

(a)

Psychological Impacts: The findings in Table 2 highlighted Positive Psychological Impacts (PPIs) among the two groups, showing notable differences. In G1, 52.6% reported high health awareness and self-care, compared to 4.6% in G2. Self-reliance and responsibility were significant in both (22.4% in G1, 19.7% in G2). Family and life appreciation were higher in G1 (27.6%) than in G2 (1.3%). Overall, PPIs affected (63.87% of G1), dropping to (36.12% in G2), reflecting reduced health awareness and self-care in the second year of e-learning. The findings revealed eleven Negative Psychological Impacts (NPIs). Isolation and social distancing were reported by 73.7% of G2 and 46.1% of G1, showing an increase in NPIs in the second year. Family pressures and domestic violence rose sharply, with 65.1% in G2 compared to 11.8% in G1. Mental health issues, such as disruption and lack of focus, were higher in G2 (57.2%) than in G1 (2.6%). Worry levels were (49.3% in G2 versus 22.4% in G1), while depression and negative emotions were affected (46.1% of G2 compared to 10.5% of G1). Fear of illness, failure, future uncertainties, and loss of self-confidence impacted both (41.4% in G2 and 44.7% in G1). The overall NPI score was significantly higher in G2 (79.59%), nearly triple that of G1 (20.40%), indicating deteriorating psychological well-being and adaptation difficulties in the second year of e-learning.

(b)

Behavioral Impacts: Five Positive Behavioral Impacts (PBIs) were recorded with notable differences. In G1, family communication increased significantly to 71.1%, but dropped to 25% in G2, indicating potential family tensions. Hobbies and skill development were higher in G1 (44.7%) than in G2 (9.2%), indicating greater personal growth in the first year. Changes in consumer behavior were more pronounced in G1 (26.3% fell to 1.3% in G2). Overall, G1 showed higher PBIs (66.20%) than G2 (33.79%), indicating better adaptation in G1. However, five Negative Behavioral Impacts (NBIs) were identified: laziness, boredom, reliance on others, reduced cooperation, creativity, and motivation. NBIs were significantly higher in G2 (63.2%) than in G1 (9.2%). Time management challenges affected (49.3% of G2, 30.3% of G1), with changes in the educational environment and reduced interactions (38.8% of G2 and 19.7% of G1). Disorganization was more common in G2 (29.6%) than in G1 (5.3%). Overall, NBIs were more prevalent in G2 (86.5%) compared to G1 (13.49%), reflecting behavioral changes in the second year of the e-learning experience.

(c)

Social Impacts: Ten Positive Social Impacts (PSIs) were recognized. Family cohesion was higher in G1 (64.5%) than in G2 (23.7%), indicating family conflicts in G2. Flexibility and ease of access (G2: 45.4% vs. 25% in G1) indicate adaptation to the opportuneness of e-learning. Technical skill improvement was greater in G2 (36.2%) than in G1 (26.3%), while economic benefits, like cost savings, were noted by (33.6% of G2 versus 19.7% of G1). Social benefits were prioritized, such as family cohesion for G1, while G2 focused on practical advantages, such as financial savings and skill development. PSIs accounted for 55.93% of total impacts (62.42% in G2, 37.57% in G1), followed by PBIs (25.71%) and PPIs (18.4%). G2 showed stronger integration through social networks, whereas G1 emphasized family interaction. Negative Social Impacts (NSIs), including technical issues, unsuitable home environments, loss of practical learning, and declining academic performance, were higher in G2 (76.35%) than in G1 (23.64%). Poor communication skills, social shyness, and difficulty forming relationships affected 66.4% of G2 students, emphasizing the importance of social interaction for well-being and academic success. Health issues, such as physical stress and weight gain, were more common in G2 (33.6%) than in G1 (3.9%), highlighting the adverse effects of prolonged e-learning.

Overall, the findings in Table 2 indicate a clear shift from relatively positive psychological, behavioral, and social outcomes in the first year of e-learning (G1) to predominantly negative impacts in the second year (G2). While G1 was characterized by higher levels of health awareness, family cohesion, and personal development, G2 showed a marked increase in psychological distress, behavioral disengagement, and social challenges. Despite some gains in technical skills and flexibility in G2, these benefits were outweighed by rising negative impacts, particularly isolation, reduced motivation, and weakened social interaction. Collectively, the results suggest that prolonged exposure to e-learning environments may undermine students’ well-being and adaptation over time, highlighting the need for more balanced and supportive educational models.

However, Figure 1 and Figure 2 showed NPIs as the most prevalent impacts (n = 171 for G1; n = 667 for G2; n = 838 combined).

The radar arrangement in Figure 1 shows a significant imbalance between positive and negative components of the e-learning experience. Negative psychological effects are the most prevalent axis across all groups, with the highest aggregate intensity observed in G2 and the combined sample. This suggests a significant increase in psychological strain over time. Negative behavioral and sociological consequences consistently raised values, implying that negative effects go beyond individual cognition to influence behavioral patterns and social relationships. Positive impacts, on the other hand, remain quite limited.

G1 contains considerably higher levels of positive psychological, behavioral, and social effects, indicating early adaptive responses such as enhanced self-awareness and involvement. However, these advantages drop significantly in G2, revealing decreased resilience and positive adaptation with extended e-learning. The overall polygon shape shows a clear transition from a more balanced effect profile in G1 to a negatively skewed distribution in G2, highlighting the cumulative and unsustainable nature of extended completely online learning circumstances. Figure 2 shows a comprehensive polar mapping of individual impact indicators, supporting the asymmetry observed in Figure 1.

The right-hand side of Figure 2 (positive effects) has modest and moderately variable values, which are principally connected with flexibility, autonomy, time efficiency, and digital skill acquisition. These elements illustrate the functional and instrumental benefits of e-learning, which contribute to the short-term flexibility and continuity of education. Conversely, the left-hand side (negative effects) has a larger frequency and intensity, with multiple peaks representing key difficulties such as social isolation, diminished interaction, psychological distress (e.g., anxiety, fatigue), decreased motivation, and cognitive overload. The grouping and magnitude of these negative signs point to a compounding effect, in which different stressors interact and reinforce each other. Importantly, the visual imbalance between the figure’s two sides demonstrates that negative effects are not only more numerous but also more intense than positive ones. This disparity highlights the limited viability of extended e-learning when implemented in the absence of sufficient psychosocial and pedagogical support systems during the pandemic situation or similar circumstances. Collectively, the statistics indicate that, while e-learning has operational benefits, its long-term reliance could negatively impact students’ well-being and learning quality, emphasizing the need for hybrid and human-centered educational models that are compatible with sustainable principles.

3.3. The Comprehensive Analysis of the Independent Sample t-Test

The independent sample t-test results, shown in

Table 3. Comprehensive analysis of the independent sample t-test.

		Levene’s Test for Equality of Variances		t-Test for Equality of Means
		F	Sig.	t	df	Significance		Mean Difference	Std. Error Difference	95% Confidence Interval of the Difference
						One-Sided p	Two-Sided p			Lower	Upper
Degrees	Equal variances assumed	4.155	0.042	3.124	268	<0.001	0.002	7.211	2.308	2.666	11.755
	Equal variances not assumed			3.675	250.871	<0.001	<0.001	7.211	1.962	3.347	11.075

Source: Author.

, indicate a statistically significant difference between the two groups for the measured variable. Levene’s test for equality of variances was significant (F = 4.155, p = 0.042), suggesting that the assumption of homogeneity of variances was not met; therefore, the unequal variances results were considered.

Table 3 reveals a significant difference between the two groups, with t (250.871) = 3.675 and p < 0.001. The average difference was 7.211 units, with a standard deviation of 1.962. The 95% confidence interval spanned from 3.347 to 11.075, indicating that the difference in group averages was statistically significant. Overall, the results show a distinct gap between the two groups, with one having much higher mean scores than the other.

Table 4. Results of independent sample t-test and differences between G1 and G2.

Psychological Impacts
Variable	Significance	Mean Difference	Group with Higher Score	Interpretation
1.  Depression and negative feelings	<0.001	0.355	Group 2	Experienced significantly higher depression and negative feelings.
Stress	<0.001	−0.237	Group 2	Reported significantly higher stress levels.
Worry	0.007	−0.270	Group 2	Recorded significant pessimism.
Psychological pressure (e.g., exams harder)	<0.001	−0.533	Group 2	Reported much higher psychological pressure.
Mental health, distraction, and lack of focus	<0.001	−0.543	Group 2	Experienced significantly more distractions and mental health issues.
2.  Behavioral & Social Adaptation
Health awareness and self-care	<0.001	0.480	Group 1	Recorded better health awareness and self-care.
Increase communication between family members	<0.001	0.461	Group 1	Showed more family communication.
Practicing hobbies and skills	<0.001	0.355	Group 1	Noted more engagement in hobbies and skills.
Practicing sport activities	<0.001	0.187	Group 1	Recorded more sports participation.
Enhance family interaction and cohesion	<0.001	0.414	Group 1	Presented higher family cohesion.
Adaptation, consumption behavior changes	<0.001	0.250	Group 1	Showed more adaptation.
3.  Social Isolation & Unhealthy Behavior Trends
Isolation and social distancing	<0.001	0.263	Group 1	Reported more isolation and social distancing.
Cheating and violating exam conduct	<0.001	−0.250	Group 2	Engaged more in cheating behaviors.
Laziness, boredom, indifference	<0.001	−0.548	Group 2	Showed much higher laziness and boredom.
4.  Educational Challenges and Technological Barriers
Technical problems and impractical tools	<0.001	−0.539	Group 2	Faced significantly more technical and learning difficulties.
Difficulty managing time and daily routine	0.006	0.191	Group 2	Struggled more with daily time management.
Poor ability to organize between study, food, and sleep	<0.001	−0.243	Group 2	Reported worse organizational skills.
Change in the perception of the educational environment	0.002	−0.191	Group 2	Perceived more negative educational changes.
5.  Economic & Environmental Awareness
Unemployment and economic impact perception	<0.001	0.249	Group 1	Perceived the economic impact as more severe.
Deprivation and unequal opportunities	<0.001	0.197	Group 1	Perceived more deprivation.
Reducing pollution, improving the environment	<0.001	0.178	Group 1	Showed more environmental awareness.
6.  Variables with No Significant Difference
Self-reliance and responsibility		0.025	No difference	No significant difference.
Self-confidence, ability to compete		−0.02	No difference	No significant difference.
Fear of illness, failure, and loss of self-confidence		0.07	No difference
Remote electronic communication skills		0.072	No difference
Ease of communication with professors		0.066	No difference
Getting to know colleagues via work groups		0.086	No difference
Sadness due to loss of relatives and friends	0.012	0.066	No difference

Key: Group 1 scored higher| Group 2 scored higher| No significant difference. Source: Author.

shows significant differences between the two groups across several dimensions (p < 0.001 in most situations). In terms of psychological and emotional effects, G2 reported greater levels of despair and negative emotions (mean difference = −0.355), stress (−0.237), and worry (−0.270, p = 0.007). Psychological stress was more significant in G2, with a significant negative mean difference of −0.533, and more issues in mental health, distraction, and lack of attention (−0.543).

Conversely, the findings in Table 4 showed that G1 has stronger behavioral and social adaptation. They reported better health awareness and self-care (0.480), greater communication among family members (0.461), and higher engagement in hobbies and skills (0.355). Participation in sports activities was also higher (0.187), as was family interaction and cohesion (0.414) and adaptation in consumption behaviors (0.250). Regarding social isolation and unhealthy behavior, G1 adhered more strictly to isolation and distancing guidelines (0.263), while G2 showed more troubling trends. They had a higher incidence of cheating and exam violations (−0.250) and reported significantly more laziness, boredom, and indifference, with a large mean difference of (−0.548). In the domain of educational and technological challenges, G2 again confronted more difficulties. They reported more technical issues and impractical learning tools (−0.539), poor time management (0.191, p = 0.006), and greater difficulty balancing study, sleep, and meals (−0.243). Their overall perception of the educational environment was also more negative (−0.191, p = 0.002). Fromeconomic and environmental issues, G1 showed more awareness. They perceived the economic impact more acutely (0.249), felt more deprived or affected by inequality (0.197, p = 0.002), and demonstrated a stronger orientation toward reducing pollution and protecting the environment. Finally, some variables showed no statistically significant differences between groups. These include self-reliance, self-confidence, fear of illness or failure, remote communication skills, ease of interaction with professors, ability to get to know peers through workgroups, and sadness over the loss of a family member.

Overall, the findings in Table 4 indicate a clear divergence between the two groups. G2 experienced significantly greater psychological distress, behavioral challenges, and academic difficulties, reflecting a more negative overall adaptation to the e-learning context. In contrast, G1 demonstrated stronger social and behavioral resilience, with higher engagement in health awareness, family interaction, and adaptive lifestyles. While some dimensions remained consistent across both groups, the findings show a shift over time toward increased vulnerability in G2, as well as a relative decline in positive coping mechanisms, supporting the assumption that e-learning has non-sustainable outcomes under certain circumstances.

3.4. Students’ Academic Performance Trends

The comparison of grade distributions between G1 (students who took the exam online from home) and G2 (students who took the exam on campus) reveals several remarkable differences in performance outcomes (see Figure 3). The findings for the final grades of students in the two groups showed the following:

(a)

Higher Grades (A and B+): G1 recorded a higher percentage of top-performing students, with 18.6% achieving an A and 20.93% a B+, totaling approximately 39.5%. In contrast, G2 had 12.13% earning an A and 13.87% a B+, totaling around 26%. This suggests that students in the online exam setting performed better at the upper end of the grade scale, possibly due to a more flexible or less stressful environment.

(b)

Mid-Range Grades (B to C): The proportions of students receiving B and C+ grades were relatively similar across both groups. G1 reported 16% in B and 18.6% in C+, while G2 had 15.02% in B and 13.29% in C+. However, G2 showed a higher percentage in the C category (17.91%) compared to G1 (13.95%), indicating a slight shift toward average performance in the on-campus group.

(c)

Lower Grades (D+, D, and F): G2 exhibited a notably higher proportion of lower grades, with 8.09% in D+, 13.29% in D, and 6.35% in F, totaling approximately 27.7%. Meanwhile, G1 showed 4.65% in D+, 8.13% in D, and 5.81% in F, totaling around 18.6%. This suggests that campus-based assessments may be associated with a higher incidence of low performance or failure.

Students who completed exams online from home, G1, demonstrated higher academic performance overall, with a greater share of top grades and fewer instances of poor performance. In contrast, students assessed on campus, G2, showed a broader spread of results with a higher concentration of lower grades. These differences suggest that exam conditions and delivery modes can significantly influence student outcomes, potentially due to factors such as exam stress, time constraints, and exam environment comfort. Figure 3 compares grade distributions between one online (Spring 2020–2021) and two on-campus exams (Fall 2021–2022). The online group shows a higher concentration of upper grades (A and B+), whereas the on-campus groups display greater proportions of lower grades (D and F). Overall, performance in the online format appears more positively skewed toward higher achievement, while campus-based exams show more dispersion and a stronger presence of lower-grade outcomes. The grade distribution analysis for the online exam was experimented with using both exponential and linear regression models. The linear model demonstrated a stronger fit (R² = 0.7862) compared to the exponential model (R² = 0.7086), indicating that approximately 79% of the variance in grade percentages across categories is explained by a linear decline from higher to lower grades. The corresponding correlation coefficient (r ≈ −0.89) indicates a strong negative relationship between grade category progression and student percentage distribution. In contrast, campus-based exams displayed more irregular and dispersed grade patterns, with relatively higher proportions in lower-grade categories (D and F). The stronger linear structure observed in the online exams suggests a more systematic performance gradient, whereas campus exams reflect greater differentiation and variability.

These findings in Figure 3 indicate statistically meaningful differences in grade distribution patterns between online and on-campus assessment formats. However, higher degrees in online exams do not automatically lead to higher educational performance.

3.5. Students’ Perceptions and Preferences for the Post-Pandemic Era

In response to the study’s third question, the research indicated students’ impressions of the post-pandemic era for G1 and G2. However, data on the fourth question, students’ post-pandemic preferences, was collected only from G2, which included 152 students in the second year of e-learning from various years, during more stable pandemic conditions, and reflected shifting attitudes in the second year of e-learning and lockdown conditions. G1 was excluded because this question was inappropriate in the first year of e-learning; this experience was fresh, and it was unpredictable whether the pandemic would finish.

Table 5. Students’ G2 preferences for the post-pandemic era.

Variables		On-Campus		E-Learning		Hybrid		No Opinion		Total
Preference		n	%	n	%	n	%	n	%	n
		64	42.1	23	15.13	6	3.94	59	38.81	152
Gender	Female	42	27.6	14	9.2	4	2.6	32	21.1	92
	Male	22	14.5	9	15.0	2	1.3	27	17.8	60
Academic year	First year	29	19.1	9	5.9	1	0.7	21	13.8	60
	Second year	22	14.5	11	7.2	2	1.3	13	8.6	48
	Third year	8	5.3	2	1.3	1	0.7	16	10.5	27
	Fourth year	5	3.3	1	0.7	2	1.3	9	5.9	17
Branch	Sharjah	35	23.0	12	7.9	2	1.3	27	17.8	76
	Al Dhaid	16	10.5	4	2.6	1	0.7	14	9.2	35
	Khorfakkan	11	7.2	7	4.6	2	1.3	16	10.5	36
	Kalba	2	1.3	0	0	2	1.3	1	0.7	5
Specialization	College of Communication	24	15.8	7	4.6	2	1.3	30	19.7	63
	Department of Sociology	40	26.3	16	10.5	4	2.6	29	19.1	89

Source: Author.

outlines three key post-pandemic preferences: on-campus, e-learning, and hybrid models. Many students experienced uncertainties shaped by their e-learning experiences and personal circumstances during the pandemic, according to the following:

3.5.1. Preference for On-Campus Education

Table 5 shows that 42.1% of students preferred on-campus learning, with a higher preference among females (27.6%) than males (14.5%). Sociology students (26.5%) were more inclined than communication students (15.6%), primarily because of the value they place on face-to-face interaction. Students emphasized that in-person learning fosters communication and cognitive skills. As RE 93 stated, “Education is more than just lectures… it fosters communication skills.” Direct interaction was seen as vital for critical thinking and engagement (RE 121, RE 84). Many reported better comprehension and time management on campus (REs 78, 178), noting that online learning disrupted routines and reduced academic focus. Technical issues were a major barrier to e-learning, including unstable connections and a lack of equipment (REs 119, 129). Others highlighted privacy concerns and family pressures during remote exams (REs 89, 82, 131), which led to stress and isolation. Some experienced academic decline linked to these difficulties (RE 124), and others mentioned mental health challenges such as depression and anxiety (RE 95, RE 170). For many, the university environment was seen as essential for motivation, social connection, and well-being. As RE 11 summarized, “On-campus education is now an urgent need… behaviors and habits have changed”.

3.5.2. E-Learning Preference

The findings in Table 5 show that 15.13% of students preferred e-learning, mainly for its flexibility and practical benefits. For many, it saved time, reduced commuting, and allowed a better balance between work, study, and family. As RE 81 noted, “E-learning has reduced my commute time… aiding in maintaining work commitments and saving time and money.” Others appreciated the university’s support in managing multiple roles. RE 136 explained, “Balancing work, family, and studies was manageable, thanks to the university’s support.” RE 120, working 172 km from campus, shared that “remote learning allowed me to manage both my job and education effectively.” Students also valued the comfort of learning from home (RE 114) and the inclusive environment it created. RE 105 stated, “Distance education allowed students to interact more freely… anytime from anywhere.” For many, especially during crises, e-learning proved successful despite its drawbacks, as RE 117 affirmed: “Despite its drawbacks, e-learning has been successful in the UAE”.

3.5.3. Hybrid: On-Campus and Online Learning

Hybrid education was the least preferred option, chosen by only 3.94% of students. Some valued it for combining the strengths of both modes. RE 125 said, “I prefer hybrid education to interact with professors in person and flexible communication with peers,” while RE 87 noted that it offers “benefits without complete reliance, ensuring societal stability.” Others, such as RE 104, appreciated its balanced nature, noting that it “reduces psychological and behavioral effects.” Despite these advantages, hybrid learning was less clearly understood or embraced. A large portion, 38.81%, expressed no clear preference, especially among first-year students, 13.8%, likely due to limited experience. No major differences were found by gender, field, or location.

Overall, the quantitative and qualitative data show a strong preference for on-campus learning, driven by the perceived benefits of increased contact, engagement, and academic success. While e-learning was commended for its adaptability and practical benefits, it was constrained by technological, social, and psychological barriers. Hybrid learning, while providing a balanced approach, remained less favored. These findings imply that, despite the rising importance of digital education, students continue to value face-to-face learning as critical to academic performance and well-being. Overall, the results indicate that learning preferences are rather constant across gender, specialty, and branch, with no statistically significant variances detected. However, the significant linear trend for the academic year suggests that students’ preferences may progressively develop as they advance through their studies. This suggests that academic advancement (experience) may be more significant in influencing choices than demographic or institutional variables.

4. Discussion

This study investigates the multiple impacts of the pandemic’s shift to e-learning, with an emphasis on the psychological, behavioral, social, and academic results of two independent groups after two years of lockdown-driven online education. Using data from 228 students, it identifies variations in views and choices for post-pandemic education, indicating important trends and patterns. These findings contribute to existing discussions about the sustainability of e-learning by providing a longitudinal, lived perception-based investigation of its psychological, behavioral, and social implications in critical crisis circumstances. The results show a significant change toward less favorable student experiences over time, notably during the prolonged exposure period (G2). However, given the complicated and overlapping settings under which the data were collected, these findings necessitate careful contextualized interpretation. A fundamental question is the extent to which the reported unfavorable results may be attributed solely to e-learning. This study does not demonstrate a clear causal relationship; rather, it examines students’ lived experiences within a larger context of concurrent stresses. The findings provide light on the complicated relationship between educational practices and their compatibility with post-pandemic sustainability goals, as well as future crises in a similar context.

4.1. The Multidimensional Impacts of E-Learning During the Pandemic

The e-learning experience during the pandemic reshaped students’ social, behavioral, and psychological well-being, with negative impacts outweighing positive ones. The descriptive findings in Figure 1 showed that Negative Psychological Impacts (NPIs) increased from 667 in the first year to 838 in the second, supporting stress theory and the link between prolonged stress and chronic harm. These findings are consistent with the existing literature [49,50,51,52,53]. Research conducted during the pandemic period has consistently documented increased levels of stress, anxiety, and disengagement among students, often linked to the combined effects of academic disruption and broader societal pressures [56,57,58,59,60]. Isolation, depression, and inequality were widespread, while the lack of psychological support constituted a long-term risk, described as a “psychological pandemic” [17,44,82]. Stress was especially elevated among women, aligning with previous studies that reported 43% increased work-related stress, 55.7% at home, and 27.3% psychological strain [25].

Despite the Positive Social Impacts (PSIs) being recorded in the G1, the family dynamics and cohesion reported by 64.5% in G1 dropped to 23.7% in G2, like the findings of previous studies, which noted 45.1% increased family support [25]. However, prolonged remote learning led to rising family conflict and even domestic violence, which supports stress theory [50,83]. Technological skills improved slightly (26.3% G1; 36.2% G2), and 45.4% of G2 appreciated e-learning’s flexibility. Environmental benefits were noted—reduced pollution, energy use, and traffic were compatible with other studies [84,85], though interest declined from 22.4% (G1) to 4.6% (G2), suggesting shifting to academic priorities.

E-learning’s cost-effectiveness was recognized by 19.7% in G1 and 33.6% in G2, aligning with other findings [8,35]. Negative Behavioral Impacts (NBIs) were prominent (n = 363), exceeding PBIs (n = 216), with time management issues rising from 23% to 49.3%. Reports of cheating, absenteeism, boredom, and laziness jumped from 9% (G1) to 63.2% (G2), echoing the findings [61,86,87]. Reduced interaction with professors and peers—from 19.7% (G1) to 38.8% (G2)—highlighted worsening organizational habits [88]. While e-learning’s accessibility and economic value are well-documented, its psychological and behavioral downsides are often minimized [33,36]. In contrast, pre-crisis studies present a more differentiated perspective, highlighting the potential benefits of e-learning in non-crisis settings—such as flexibility, autonomy, and enhanced self-regulation—when implemented under stable conditions and supported by effective pedagogical strategy [38]. This divergence reinforces the argument that contextual factors significantly mediate the outcomes of e-learning according to different contexts.

The independent sample t-test in Table 4 revealed significant group differences and critical changes across psychological, behavioral, and educational domains (p < 0.001 in most cases). G2 experienced more psychological distress, reporting higher levels of depression, stress, and distraction (e.g., mean diff. in psychological pressure = −0.533). In contrast, G1 demonstrated better behavioral and social adaptation, including stronger health awareness (0.480), family interaction (0.461), and engagement in hobbies and sports. G1 also adhered more to isolation guidelines and showed fewer signs of boredom or academic misconduct. G2 faced greater educational and technical challenges, including time management issues and negative perceptions of the e-learning circumstances during the pandemic. Economically and environmentally, G1 showed more awareness and concern. However, some variables—such as self-confidence, fear of illness, and communication with professors—did not differ significantly, indicating shared experiences despite differing overall impacts. Although some studies observed post-pandemic resilience, these findings reveal that prolonged stress has significantly altered student attitudes over time [89]. However, the transition from relatively balanced—or moderately positive—perceptions in G1 to a predominantly negative profile in G2 suggests that duration and cumulative exposure to fully online learning environments play a critical role. Importantly, this pattern should be understood as the result of an interaction between sustained digital learning conditions and external stressors, rather than as an inherent limitation of e-learning as a modality [90,91,92].

4.2. Academic Performance Between Resilience and Sustainability

The comparison between G1 (online exams) and G2 (on-campus exams) revealed notable performance differences among students. G1 achieved more top grades (A and B+), totaling 39.5%, versus 26% in G2, while G2 recorded more low grades (27.7% vs. 18.6%), suggesting that the online setting may have provided a less stressful or more flexible environment. Mid-range grades were similar, though G2 showed a slight increase in C-level scores. These outcomes support the idea that exam conditions significantly influence academic performance. The findings align with previous studies, which observed higher pass rates in online exams, and that students in scientific fields—especially male medical students—faced greater challenges during the pandemic, affecting study time and outcomes [15,18,27,93]. While other studies highlighted the disadvantages of e-learning, unequal access to technology contributed to performance declines [37,58,94]. These findings emphasized the need for balanced strategies addressing digital access, learning environments, and educational resilience for long-term sustainability by ensuring good education, well-being, and equality for all students under similar settings.

4.3. E-Learning for the Post-Pandemic Era and in Times of Crisis

The findings revealed that 42.1% of students in G2 preferred on-campus education, particularly those who began university online during the pandemic. The findings align with other studies, which confirmed that 69.65% of students preferred face-to-face, 23.08% hybrid, and 7.27% remote education [10]. This preference echoes findings on the challenges faced by new students adapting to online formats [90]. On-campus settings were favored for better communication, skill development, and social interaction, aligning with previous studies [91]. Only 15.13% preferred e-learning for its flexibility and cost-effectiveness, reflecting broader trends [92], while 3.94% chose hybrid models, citing unclear benefits and challenges in line with previous studies [48]. Notably, 38.81% showed no clear preference, highlighting the need for adaptable strategies that address diverse student needs and integrate digital technologies to support academic performance and well-being.

4.4. The Sustainability of E-Learning Redefining Strategies Toward a New Sustainable Educational Paradigm

This study applies a systems-oriented approach to sustainability, utilizing known proxy indicators and outcomes—student well-being (SDG 3), learning conditions (SDG 4), and perceived inequality (SDG 10)—rather than direct SDG measures. As a result, the findings do not measure SDG accomplishment per se, but rather examine how e-learning environments change important underlying mechanisms related to these objectives. The findings therefore indicate short-term variations in health, learning, and equitable circumstances, which may have long-term consequences for sustainable paths, particularly in terms of strengthening or reducing existing inequities.

Resilience, sustainability, and change management policies are essential for adapting education systems in times of crisis, such as pandemics, natural disasters, and wars. While an e-learning environment is significant, it alone cannot ensure sustainable education, well-being, and academic success. Achieving this requires robust support for students and faculty that aligns with the SDGs, promoting global health (SDG 3) and education quality (SDG 4) and (SDG 10) reducing inequality [23]. This approach enriches learning experiences and aligns education with societal goals. Key competencies for sustainable development include cognitive (knowledge and thinking skills for SDGs), behavioral (action and teaching methods), and socio-emotional (collaboration, communication, self-reflection, and values) aspects, which are essential for a successful life and a well-functioning society. Key competencies such as systems thinking, strategic and critical thinking, self-awareness, and problem-solving are strongly recommended [23,95]. It facilitates profound analysis and sound decision-making. However, soft skills such as “integrity, communication, civility, accountability, social skills, a positive attitude, professionalism, adaptability, collaboration, and work ethics” are widely acknowledged as vital in higher education, augmenting disciplinary knowledge and improving both academic performance and employability [96]. However, in today’s digital and mixed learning contexts, adjustment capacity and emotional intelligence are crucial for dealing with uncertainty and sustaining good work [97]. Also, leadership and digital literacy, including understanding how to use AI technologies, are important for getting ahead in today’s academic and work environments [98,99]. These skills work together to provide more well-rounded learning outcomes, better prepare students for changing social and economic needs and comprehensive e-learning patterns, and ensure the provision of soft skills. However, emerging studies have revealed that while general-purpose GenAI technologies can improve task performance in e-learning contexts, they do not always yield substantial learning benefits. Offloading cognitive work to chatbots increases the risk of metacognitive disengagement and superficial learning, which may impede long-term skill development in digital and hybrid educational models [100]. UNESCO (2017) announced a new strategy, “Education for Sustainable Development,” for ensuring learners acquire skills to promote sustainable practices, including human rights, gender equality, peace, and global citizenship, based on: (a) a learner-centered approach, (b) action-oriented learning, and (c) transformative learning [23]. Based on this vision, this study highlights the need for resilience, digital transformation, and sustainability in post-pandemic education, especially with new artificial intelligence systems. UNESCO (2020) emphasized skill development, social–emotional learning, and mental health support to enhance academic outcomes, align with SDGs, and prepare education systems for future challenges, promoting a sustainable and equitable world for all [63].

5. Conclusions

The pandemic disrupted in-person education, posing challenges to achieving the UN’s 17 SDGs. Universities transitioned to e-learning, often unprepared, creating both challenges and opportunities. Firstly, this study examined the psychological, behavioral, and social impacts of e-learning on students during the pandemic, offering insights to improve future educational practices. E-learning provided flexibility and education continuity during the COVID-19 pandemic in 2020–2023, but affected mental health, increasing stress and depression due to reduced social connections, challenging the achievement of SDGs by 2030. Negative impacts, particularly in the second year, outweighed positives, supporting stress theory and studies linking prolonged stress to chronic issues [49]. Behaviorally, students faced difficulties with motivation and focus, which affected performance despite improvements in digital and learning skills [46,48]. Findings showed changes in students’ behavior, reflecting challenges in maintaining study habits in an online environment. Socially, family cohesion initially improved but later declined, leading to conflicts and long-term psychological impacts [83]. Lockdown during e-learning and isolation policies reduced engagement with peers and teachers, impacting “quality of life”, “social well-being”, “collaborative learning”, and “soft skills learning”.

The transition to e-learning yielded mixed results, with some students benefiting from its flexibility, while others struggled with increased workloads and a lack of psychological support due to different forms of inequality. Continuous stress negatively shifted attitudes toward e-learning, with a clear preference for on-campus education, particularly among those who started their studies online. Secondly, the study revealed that e-learning impacted academic performance, raising concerns about “inequality between students”. Students from various faculties, not just scientific ones, faced significant challenges that negatively impacted their “well-being” and “academic performance” [56]. These findings reveal the “unsustainable nature of e-learning” as the only model relying on quantitative approaches, emphasizing the need for “qualitative strategies” like “resilience”, “equality”, “digital transformation”, “curriculum revision”, and “sustainability approaches”. Thirdly, the study recommends that future education strategies should integrate SDGs into curricula, fostering cognitive, behavioral, and socio-emotional skills [23]. Highlighting the need for an adaptable and sustainable educational paradigm, the UOS adopted in the academic year 2024–2025 a “hybrid teaching model” in autumn, and in spring a “face-to-face teaching model,” paving the way for further studies on its impacts on students and academic staff to ensure well-being and good public health for all.

6. Implications of the Study and Suggestions

This study demonstrates that education is a public health and well-being issue and reveals, at the same time, a clear sustainability paradox in e-learning environments. While initial exposure to digital learning supported behavioral adaptation, family cohesion, and personal development, prolonged reliance on e-learning significantly increased psychological stress, academic challenges, and social inequalities. These results suggest that the sustainability of digital education depends not only on technological accessibility but also on its long-term impacts on student well-being, mental health, and learning effectiveness. Overall, the results indicate that e-learning should not be conceptualized as inherently detrimental. Rather, it is a context-dependent educational modality whose outcomes are shaped by implementation quality, duration, and surrounding socio-environmental conditions. The observed decline in positive indicators and the intensification of negative experiences over time underscore the risks associated with prolonged reliance on fully online learning environments with the absence of adequate pedagogical, psychological, and social support structures.

From a sustainability perspective, these findings highlight the necessity of resilient and adaptive e-learning systems. E-learning sustainable models should move beyond emergency remote teaching toward structurally integrated approaches that combine digital flexibility with meaningful social interaction. Hybrid learning models appear essential for mitigating the negative effects associated with prolonged isolation and disengagement. In addition, institutional strategies should incorporate continuous monitoring of student well-being, alongside pedagogical innovations that enhance engagement and reduce cognitive overload. Without such adjustments, extended reliance on fully online learning risks generating diminishing returns in both educational quality and student well-being, thereby undermining the long-term sustainability of digital education systems. In conclusion, while e-learning maintained educational continuity during the crisis, it revealed significant areas needing improvement in mental health support, engagement strategies, and teaching methods. This study demonstrates a clear temporal shift toward more negative student perceptions with prolonged exposure to e-learning. However, these findings must be interpreted within a broader contextual framework that accounts for concurrent stressors and methodological constraints. By adopting a balanced and critically reflective perspective, the study contributes to a more nuanced understanding of e-learning sustainability and emphasizes the importance of context-sensitive, adaptive educational strategies. This study builds on global research while offering a new multidisciplinary perspective. It highlights the urgent need for a resilient, sustainable education model that supports students academically, psychologically, and socially, promoting well-being and understanding. These findings are crucial for policymakers, planners, and researchers, especially considering ongoing war-related crises in the Gulf (March 2026) and the shift toward e-learning. E-learning, as implemented during the pandemic, remains highly relevant today, informed by accumulated experiential evidence. Integrating these lessons can drive innovative strategies that address immediate challenges and strengthen both short- and long-term outcomes. Aligned with sustainable education, such efforts support SDGs 3, 4, and 10, and accelerate progress toward all 17 SDGs by 2030, and well-being for all.

7. Limitations and Future Studies

The reliance on self-reported data constitutes both a limitation and a methodological strength. While subjective perceptions may not fully align with objective performance indicators, they provide essential insight into students’ lived experiences, as recommended by similar previous studies [56], which are central to evaluating the sustainability and acceptability of educational systems in crisis, such as pandemics and wars, when there is an obligation to move to online learning. Prior research demonstrates that perceived stress, motivation, and engagement are themselves critical determinants of academic success and persistence [43]. Nonetheless, the potential divergence between perceived and lived outcomes underscores the need for future research to adopt mixed-method designs that integrate direct physical observation -based data with behavioral and performance indicators. Such an approach was not feasible in the present study due to pandemic-related lockdown restrictions, which precluded direct physical observation. Despite this limitation, the study was able to capture meaningful shifts in students’ attitudes, behaviors, academic performance, and overall well-being based on examining two different groups of students (G1 and G2) representing the first and second years of e-learning, and yielded meaningful insights. The absence of financial support constrained further investigation and timely publication. A key limitation is the absence of direct physical observation-based data, as pandemic-related lockdown restrictions precluded in-person direct observation. Nonetheless, proxy indicators—such as proctored online examinations and variations in students’ academic performance across online and on-campus assessments—partially capture behavioral and performance dimensions. Furthermore, future research should explore the impacts of GenAI on teaching and learning outcomes to ensure the sustainability of education systems in different environments via on-campus and e-learning strategies.