Digital Fatigue, Sustainability Behaviour, and Energy Awareness Among Generation Z: The Role of Cognitive Resources and Education

Abstract

This study investigates how digital lifestyles and cognitive fatigue influence sustainable behaviour and energy awareness among Generation Z. Drawing on environmental psychology and social science perspectives, it explores behavioural and cognitive mechanisms linking digital overexposure with pro-environmental engagement. A cross-sectional survey conducted among 683 Polish secondary-school students examined the relationships between digital activity, fatigue, self-regulation, and sustainability practices such as waste segregation, reuse, and consumption moderation. The results show that higher digital fatigue and problematic online use are negatively associated with sustainability engagement, supporting the view that cognitive overload reduces individuals’ capacity for mindful, sustainability-oriented action. Using k-means clustering and robust regression analyses based on ordinary least squares (OLS), this study identifies distinct sustainability behaviour profiles among Generation Z and examines how digital fatigue and problematic online use predict lower engagement in pro-environmental practices. Importantly, educational level moderated this effect, suggesting that energy and sustainability literacy can buffer the adverse consequences of digital exhaustion. The findings contribute to the growing field of digital sustainability and highlight the need to integrate digital well-being and environmental education into youth and social policy frameworks.

1. Introduction

In the face of global sustainability challenges and the increasing social impacts of climate change, research has progressively shifted from technological optimisation toward the study of human behaviour and everyday practices. While traditional approaches focus on physical systems and efficiency metrics, contemporary social science perspectives emphasise that the success of sustainability transitions depends equally on transforming daily habits, consumption patterns, and individual cognition. Within this context, Generation Z occupies a crucial position—both as digital natives and as future sustainability citizens.

At the same time, the deep digitalisation of everyday life raises new questions about how technology use influences sustainable behaviour and social well-being. Emerging evidence suggests that excessive digital engagement may contribute to cognitive overload and self-regulatory fatigue, thereby reducing individuals’ capacity for mindful and sustainability-oriented decision-making.

Sustainable development is widely conceptualised as a multidimensional construct encompassing environmental, social, and economic pillars. While the present study adopts this broader understanding, its empirical focus is deliberately placed on the behavioural–environmental dimension of sustainability, operationalised through everyday practices that are observable and relevant for adolescent populations.

The present study integrates insights from environmental psychology, behavioural social science, and digital media studies into a unified framework for analysing sustainable behaviour among young people. It examines how digital lifestyles, cognitive resources, and educational experiences interact to shape environmental sustainability awareness and behaviour among Generation Z, understood as one component of the broader sustainability framework. Furthermore, it explores the moderating role of demographic factors—gender, place of residence, and education—and introduces a composite indicator, the Sustainability Behaviour Index (SBI), to measure youth engagement in sustainable actions.

By combining behavioural and socio-cognitive perspectives, this research contributes to the growing body of literature that links human behaviour, digital well-being, and environmental responsibility. It highlights the importance of viewing sustainability not only as a technological challenge but also as a cognitive and social process, thereby offering a conceptual bridge between digital sufficiency, energy awareness, and sustainable citizenship.

The growing energy demand of digital infrastructures, including data centres and AI-driven services, underscores the relevance of linking digital sustainability with physical energy systems. This connection opens an important interface between sustainability research and physics- and STEM-oriented education.

2. Sustainable Behaviour in the Digital Era: A Literature Review

In sustainability research, behavioural indicators are often employed to capture specific dimensions of sustainable development rather than its full conceptual scope. In the context of youth studies, environmental behaviours are frequently used as empirically accessible proxies of sustainability engagement, while acknowledging that social and economic dimensions remain equally important but methodologically distinct.

Research linking pro-environmental behaviour with digital technology use remains an emerging, interdisciplinary field that integrates insights from environmental psychology, media and communication studies, life-cycle assessment (LCA), and studies on digital overload and media addiction. This approach reflects a growing recognition that sustainability in the digital age cannot be understood solely through technological efficiency but must also account for cognitive, motivational, and social processes underlying everyday digital practices (Ramírez-Correa et al. 2025; Sá et al. 2023).

2.1. The Energy Intensity of Digital Technologies and the Environmental Footprint of ICT

With regard to the environmental dimension of digitalisation, the review by Belkhir and Elmeligi (2018) remains one of the most frequently cited analyses of the emission footprint of information and communication technologies (ICT). The authors demonstrated that, even under optimistic scenarios, the ICT sector may account for over 14% of global emissions by 2040, surpassing those from air transport (Belkhir and Elmeligi 2018). This macro-level perspective justifies treating daily digital activity time (WZC) as an indirect indicator of environmental demand—even if users remain unaware of the ecological consequences of their screen-based routines (Freitag et al. 2021).

Recent LCA-based studies of network infrastructure, data centres, and end-user devices indicate a strong rebound effect: improvements in device efficiency are often offset by intensified usage (Freitag et al. 2021; Istrate et al. 2024; Malmodin et al. 2024). Lange et al. (2020) even question whether digitalisation results in a genuine reduction in total energy demand or merely relocates it into an immaterial sphere (Lange et al. 2020).

This ambivalence highlights the need to combine quantitative data on device-use time (WZC) with self-reported environmental attitudes to empirically assess the trade-offs between digital convenience and sustainability values (Hassoun et al. 2025).

2.2. Digital Consumption as Sustainable Behaviour: From Macro-LCA to Micro-Level Indicators

Life-cycle analyses and energy-use estimates for the ICT sector consistently demonstrate that digital services constitute a growing share of global emissions. This supports using WZC (digital activity time) as an indirect measure of environmental intensity.

Empirical and meta-analytical studies provide frameworks for assessing the ecological impact of digital consumption in terms of emissions and resource use—essential for linking individual choices to macro-level outcomes (Istrate et al. 2024; Malmodin et al. 2024; Nosratabadi et al. 2023).

In behavioural research, there has been a clear shift from classical approaches—such as feedback and goal-setting (Abrahamse et al. 2007; Composto and Weber 2022)—toward dynamic, “living review” models that continuously integrate the effects of informational, social, and financial stimuli (Khanna et al. 2024). However, few studies explicitly examine behavioural interventions within digital contexts; for example, few treat screen-time reduction itself as a pro-environmental or well-being-oriented action.

2.3. Behavioural Change and Environmental Interventions: From Households to Digital Environments

Theoretical models in environmental psychology—such as the Value-Belief-Norm (VBN) model and the Theory of Planned Behaviour (TPB)—remain fundamental to interpreting pro-environmental actions (Abrahamse et al. 2005; Steg 2023).

Complementary research on household energy behaviour (Iweka et al. 2019) identifies multiple intervention strategies, including feedback, financial incentives, and social comparison. These approaches show heterogeneous results, depending on the social and technological context.

In the present study, such findings justify the use of regression (Tobit/OLS) and cluster models to detect user profiles differing in their sensitivity to environmental and digital cues (Iweka et al. 2019).

2.4. Environmental Psychology and Barriers to Sustainable Action

Wang et al. (2021) provide a comprehensive classification of barriers to low-carbon behaviour, encompassing cognitive and motivational factors (e.g., low self-efficacy) as well as structural and social constraints. These frameworks align with the factorial variables used in this study (attitudes, knowledge, situational constraints) and reinforce the assumption that interactions between psychological traits and digital contexts are crucial for explaining sustainability behaviours among youth (Wang et al. 2021).

2.5. Digital Fatigue, Self-Regulation, and Addictive Use as Cognitive Resource Depletion

Psychological research has long examined cognitive fatigue, self-regulatory depletion, and media addictions—constructs closely related to those operationalised here as WNS (digital fatigue) and WPU (problematic use) (Freed et al. 2025; Qin et al. 2024; Sweigart et al. 2025).

“Cognitive-resource” models, such as ego-depletion and self-regulation theory, suggest how intensive digital engagement may drain the mental energy required for sustainable, goal-directed behaviour (Paterna et al. 2024). Validated instruments like the Bergen BSMAS and other scales of problematic use (Baumeister et al. 1998; Elhai et al. 2017) can be effectively adapted for quantitative analyses linking digital behaviour and sustainability outcomes.

An additional strand of research highlights online risk exposure and gambling-related behaviours as an emerging dimension of problematic digital engagement among adolescents. Risk-oriented online activities, including exposure to gambling content, betting platforms, and gamified reward systems, have been shown to intensify impulsivity, cognitive overload, and self-regulatory depletion, particularly among young users. Recent studies indicate that such patterns are closely linked to reduced self-control and impaired goal-oriented decision-making in digital environments (Paterna et al. 2024; Sweigart et al. 2025). In this study, this dimension is captured as WRH (Online Risk and Gambling Exposure) and treated as a complementary digital-behavioural construct relevant to sustainability-related self-regulation.

2.6. The Concepts of Digital Sustainability and Digital Sufficiency

In response to the growing environmental footprint of online activity, the paradigms of digital sustainability and digital sufficiency have emerged, advocating the inclusion of digital user behaviour in mainstream sustainability strategies (Farrelly 2025; Istrate et al. 2024).

Santarius et al. (2023) define digital sufficiency as a focus on usage intensity and behavioural patterns, not merely technological performance, as key determinants of the ecological impact of online activities (Gujral et al. 2025; Santarius et al. 2023). Similarly, Istrate et al. (2024), in the first full-scale LCA of digital content consumption, argue that micro-behaviours such as autoplay, multitasking, and scrolling should be integrated into environmental modelling.

This approach shifts attention from innovation alone toward behavioural moderation—encouraging “thinking before the next stream or scroll”—and the design of technologies that promote long-term sufficiency rather than short-term efficiency. Review studies further indicate that digitalisation’s effects are systemic and complex: expected savings are often offset by rebound effects, as the convenience of online services increases total resource use (Mouthaan et al. 2023).

2.7. Education, Energy Awareness, and the Role of Youth as Sustainability Citizens

Studies on youth energy literacy confirm that education moderates the link between environmental attitudes and behaviours, justifying its inclusion as a key moderator in this research (Elhai et al. 2017; Jaradat et al. 2024; Ji et al. 2023; Kellberg et al. 2025). Additionally, a large-scale European study found that digital media engagement among youth significantly predicted pro-environmental behaviours across countries (Hansen et al. 2025).

Research shows that learning ecologies foster energy awareness not only through formal instruction but also through participation in digital and community practices (Ji et al. 2023; Todorovic et al. 2025). Other studies highlight that young people become sustainability citizens when they possess both cognitive competence and agency regarding energy infrastructures (Jaradat et al. 2024). In the context of the present study, this concept is applied specifically to the environmental–behavioural dimension of sustainability. While the notion of “sustainability citizenship” encompasses social and economic agency in a broader theoretical sense, the empirical focus here is limited to environmentally oriented practices and energy awareness that are measurable and developmentally appropriate for adolescent respondents.

Education also helps bridge the intention–behaviour gap by enhancing self-efficacy and personal norms, with studies showing that interventions combining information and behavioural feedback yield lasting effects even under high cognitive load (Khanna et al. 2024), confirming earlier findings from student-focused meta-analyses (Abrahamse et al. 2007; Pabian and Pabian 2023).

Furthermore, recent reports (Shengjergji et al. 2024) warn that the environmental cost of digital education may produce an “educational paradox”—students expressing high climate concern while generating large digital footprints. This underscores the need to integrate digital sustainability awareness into curricula, reinforcing the assumption that education not only promotes pro-environmental action but also moderates the link between WNS and SBI.

2.8. Research Gaps and Integration of Behavioural and System-Level Assessments

The convergence of environmental psychology and digital sustainability research highlights the necessity of integrating behavioural indicators with systemic assessments. This supports a conceptual approach of linking the SBI with digital variables (WZC, WNS, WPU, WRH) and employing hybrid analytical models, including principal component analysis (PCA), clustering, and moderated regression.

An emerging focus concerns algorithmic influence—so-called algorithmic nudging—which challenges the assumption that behaviour results solely from conscious beliefs (Kovács-Szépvölgyi et al. 2025; Yu et al. 2024). Interdisciplinary oversight of algorithmic systems has been called for, as these systems shape users’ attention and decisions, thereby influencing digital energy consumption (Schmauder et al. 2023).

Empirical findings also indicate a cognitive gap: users rarely estimate the emissions associated with cloud use, streaming, or app engagement, which hinders self-regulation and sufficiency-oriented habits (Frick et al. 2021).

Policy and research agendas should therefore extend environmental psychology frameworks to include digital contexts—connecting LCA with behavioural analysis, accounting for algorithmic effects, and promoting interventions that encourage digital moderation and sufficiency.

Representative works include:

• Santarius et al. (2023)—Digital Sufficiency.
• Mouthaan et al. (2023)—Systemic Sustainability Effects of Digitalisation.
• Schmauder et al. (2023)—Algorithmic Nudging.

Parallel discussions in medical and educational literature point to the cognitive and motivational costs of excessive technology use, including fatigue and declines in learning performance (Rudroff 2025). Educational reports (Shengjergji et al. 2024) similarly caution that the digital transformation of schooling through EdTech may generate a new form of educational digital emissions.

These health-related and environmental perspectives have rarely been integrated, offering new research opportunities to explore whether self-regulation (WNS/WPU) mediates the relationship between digital overstimulation and ecological motivation.

Accordingly, the selection of indicators, constructs, and analytical strategies in this study—PCA for SBI construction, standardisation of digital variables, cluster analysis, and regression models with interaction terms and heteroskedasticity correction—aligns with current best practices and supports the testing of hypotheses H1–H5 within an interdisciplinary social-scientific framework (Huebner et al. 2021).

3. Materials and Methods

The present study adopts a behavioural and cognitive sustainability framework to examine how members of Generation Z manage resources and attention across two interrelated domains: (1) sustainable everyday practices and (2) digital engagement. While physical sustainability systems involve measurable resource flows, individual decisions and cognitive processes also represent micro-level systems of regulation, where attention, motivation, and self-control can be interpreted as cognitive resources. Excessive digital activity, in turn, implies indirect environmental impact through ICT infrastructures and mental depletion through sustained cognitive load. Within this framework, sustainable actions such as waste segregation, reuse, and moderate consumption are treated as indicators of behavioural sustainability, whereas digital fatigue and problematic online use are understood as forms of cognitive imbalance. The study aims to identify links between these domains, contributing to the understanding of sustainability behaviour and digital awareness among young people.

3.1. Study Design and Participants

A cross-sectional online survey (Computer-Assisted Web Interview, CAWI) was conducted among secondary-school students representing Generation Z in Poland. The questionnaire was distributed through direct invitations to schools and posts on educational forums and student social-media groups. Data were collected between 1 March 2025 and 30 June 2025. The study was conducted within an academic research project coordinated by the Institute of Economics and Finance, The John Paul II Catholic University of Lublin (Lublin, Poland).

All participants received an information sheet and provided electronic informed consent. Participation was voluntary and anonymous.

In total, N = 683 respondents completed the survey and met quality criteria (valid core demographics). The mean age was M = 17.38 years (SD = 1.41). The sample consisted of 55.5% males, 37.2% females, and 7.3% undisclosed1, attending various types of secondary schools in both rural and urban areas.

Missing data were handled as follows: for scale construction (e.g., the Sustainability Behaviour Index, SBI), respondents were included if at least 75% of items were complete. For multivariate analyses, listwise deletion was applied. Consequently, the final analytic subsamples were as follows:

▪

Behavioural profiles + demographics (N = 428);

▪

Extended models with digital variables (N = 475)2.

Sensitivity analyses confirmed robustness of effects across missing-data strategies. Additional analyses confirmed that respondents excluded due to listwise deletion did not differ significantly from the retained sample in terms of age or gender. This reduces the likelihood of systematic bias resulting from missing data handling.

3.2. Measures

3.2.1. Sustainable Behaviour Indicators

Four behavioural items captured everyday sustainable practices:

▪

Waste segregation;

▪

Reuse of materials or products;

▪

Purchasing second-hand items;

▪

Frequency of new clothes purchases (reverse-coded for moderation).

Responses were recorded on a five-point Likert scale (1 = never, 5 = always).

Principal Component Analysis indicated a one-factor solution explaining 61.8% of variance (Cronbach’s α = 0.83). The resulting Sustainability Behaviour Index (SBI) served as the main dependent variable, rescaled to 0–100 for interpretability. These indicators capture routine pro-environmental practices and should be interpreted as partial behavioural measures of sustainability. They do not encompass the social or economic dimensions of sustainable development, which were beyond the scope of the present empirical design.

3.2.2. Digital Behaviour and Cognitive Variables

Four constructs represented digital-behavioural and cognitive aspects:

▪

WZC (Digital Activity Time)—Self-reported daily hours online, proxy for digital intensity;

▪

WNS (Digital Fatigue)—Seven items assessing perceived overload and attention loss;

▪

WPU (Problematic Use)—Six items on difficulty limiting online activity and impulsivity;

▪

WRH (Online Risk and Gambling Exposure)—Four items on gambling or risky online behaviour.

All scales were z-standardised before analysis.

The scales measuring Digital Fatigue (WNS; α = 0.87) and Problematic Use (WPU; α = 0.81) showed high internal consistency, as assessed by Cronbach’s alpha.

3.2.3. Demographic and Moderating Variables

Controls included gender, age, residence (rural–urban), and education level. Education was also tested as a moderator of the relationship between digital fatigue and sustainable behaviour.

3.3. Analytical Strategy

The analytical design combined exploratory and predictive methods. All analyses were conducted in IBM SPSS Statistics 29.0 (IBM Corp., Armonk, NY, USA), with p < 0.05 as the significance threshold.

• PCA—Validation of SBI structure.
• k-means clustering (k = 2–6)—Identification of behavioural profiles.
• OLS regression—Demographic predictors of SBI.
• Multinomial logistic regression—Demographic differentiation of sustainability profiles.
• Gender-specific models—Differences in determinants of sustainable behaviour.
• Correlation and PCA of digital indicators (WZC, WNS, WPU, WRH).
• Extended regressions—Predicting SBI with digital variables.
• Moderation—Testing whether education mitigates the negative link between digital fatigue and sustainable behaviour.

Continuous variables were mean-centred; robust standard errors were corrected for heteroskedasticity.

3.4. Research Hypotheses

The hypotheses refer specifically to behavioural–environmental sustainability practices, understood as one dimension of sustainable development. This study explores how sustainability behaviour and digital engagement intersect among members of Generation Z. It assumes that both domains rely on common psychological mechanisms such as attention, motivation, and self-regulation, which shape behavioural efficiency in everyday life.

Five hypotheses were formulated:

H1. 

Everyday sustainable practices (e.g., waste segregation, reuse, moderate consumption) form a coherent construct of behavioural sustainability.

H2. 

Distinct behavioural profiles emerge, reflecting varying orientations toward sustainability and digital balance.

H3. 

Socio-demographic factors (gender, age, residence, education) predict sustainable behaviour, with females and urban youth expected to show higher engagement.

H4. 

Higher levels of digital activity, fatigue, and problematic use negatively predict sustainable behaviour, revealing a behavioural trade-off between digital intensity and sustainability.

H5. 

Education moderates the relationship between digital fatigue and sustainable behaviour—individuals with greater sustainability literacy and digital awareness are less affected by cognitive overload.

Together, these hypotheses position sustainability as both a behavioural and cognitive process rooted in the digital lifestyles of young people.

3.5. Expected Contribution

This study contributes to social and sustainability sciences by extending the notion of efficiency from physical systems to human cognition and behaviour. By linking sustainable lifestyles with digital wellbeing, it introduces a novel framework for understanding how young people balance environmental awareness with digital engagement.

At the theoretical level, it proposes an integrative behavioural-sustainability model, connecting sustainable practices with cognitive and digital dynamics.

At the empirical level, the study:

▪

Quantifies behavioural sustainability among Generation Z;

▪

Reveals how digital engagement and fatigue affect sustainable practices;

▪

Identifies demographic moderators shaping youth behaviour;

▪

Supports design of educational programmes promoting digital balance and sustainability literacy.

Ultimately, it frames energy transition as a social and cognitive process, emphasising digital wellbeing and sustainability education as key components of a socially just transition.

4. Results and Statistical Analyses

This section presents the empirical results structured according to the study’s main hypotheses. Analyses were conducted to identify behavioural patterns, socio-demographic predictors, and the relationships between digital engagement, cognitive variables, and sustainability practices. The findings provide an integrated picture of how Generation Z balances digital lifestyles and sustainable behaviour, contributing to a broader understanding of digital well-being and sustainability literacy among youth.

4.1. Descriptive Statistics and Measurement Validation

Table 1. PCA loadings for four sustainability behaviour items.

Variable (Behavioural Indicator)	PC1
Waste segregation	0.851
Reuse of multi-use items	0.879
Second-hand clothing	0.812
Frequency of new clothes (rev.)	0.743
Explained Variance (%)	61.8

Note: $SB{I}_{\mathrm{i}}={\lambda }_{1}WasteSegregatio{n}_{\mathrm{i}}+{\lambda }_{2}Reus{e}_{\mathrm{i}}+{\lambda }_{3}SecondHan{d}_{\mathrm{i}}+{\lambda }_{4}ConsumptionModeratio{n}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$; PCA revealed a one-dimensional solution. The loadings represent the correlation between each item and the principal component (SBI). Factor loadings (λ₁–λ₄) correspond to the correlations between each observed variable and the latent construct.

presents the results of the Principal Component Analysis (PCA) for the four sustainability-related items: waste segregation, reuse, purchasing second-hand products, and moderation of consumption frequency. The analysis revealed a unidimensional factor solution explaining 61.8% of total variance.

All four variables loaded strongly on this factor (λ = 0.74–0.88), confirming their shared conceptual core: habitual pro-environmental behaviour. The extracted factor was thus defined as the Sustainability Behaviour Index (SBI), reflecting the frequency and internalisation of sustainable practices among respondents.

The internal consistency of this dimension (Cronbach’s α = 0.83) confirmed a robust and coherent construct suitable for further analyses.

The next step involved classifying participants into homogeneous groups according to their sustainability engagement. As shown in

Table 2. Silhouette scores for K-Means cluster solutions (k = 2–6).

Number of Clusters (k)	Silhouette Score
2	0.41
3	0.47
4	0.44
5	0.42
6	0.39

Note: Optimal cluster solution: k = 3 (highest silhouette score = 0.47).

, silhouette coefficients for k = 2–6 indicated that the optimal clustering solution was achieved for k = 3 (silhouette = 0.47), supporting clear and interpretable segmentation.

The resulting profiles were:

▪

Cluster 1—Eco-Active: respondents with high engagement across all sustainable behaviours,

▪

Cluster 2—Moderate Participants: selectively engaged (e.g., recycling but limited reuse),

▪

Cluster 3—Passive Consumers: low sustainability and higher consumption frequency.

This pattern supports the conceptual validity of differentiated sustainability orientations within Generation Z and provides a foundation for subsequent comparisons.

Regression analysis further explored socio-demographic predictors of sustainable behaviour (

Table 3. Robust OLS regression predicting Sustainability Behaviour Index (SBI).

Variable	Coefficient (β)	Std. Error	t	p-Value
Constant	58.213	3.521	16.55	<0.001
Female	4.672	1.824	2.56	0.011
Urban residence	3.942	1.576	2.50	0.013
Age	0.184	0.089	2.07	0.039

Note: $SB{I}_i={\beta }_0+{\beta }_{1}Femal{e}_i+{\beta }_{2}Urba{n}_i+{\beta }_{3}Ag{e}_i+{\epsilon }_i$; R² = 0.164; Adjusted R² = 0.152 (Standard errors are HC3-robust).

). The robust OLS model was significant (F = 9.84, p < 0.001), explaining 16.4% of SBI variance. Gender emerged as the strongest predictor: female respondents reported significantly higher sustainability scores than males (β = 4.67, p = 0.011). Urban residence also had a positive effect (β = 3.94, p = 0.013), while age showed a modest upward trend (β = 0.18, p = 0.039). These findings are consistent with prior social research indicating that women and urban youth tend to show greater sustainability engagement due to normative influences, educational exposure, and access to pro-environmental infrastructure.

To examine more nuanced patterns, multinomial logistic regression was used to predict membership in the three clusters, using the Eco-Active group as a reference (

Table 4. Multinomial logistic regression predicting cluster membership.

Variable	Cluster 1 (Passive) OR 95% CI	p-Value	Cluster 2 (Moderate) OR 95% CI	p-Value
Constant	0.48 0.22–1.05	0.065	0.35 0.16–0.78	0.010
Gender	2.13 1.38–3.29	0.001	1.19 0.79–1.79	0.402
Place of Residence
Suburban	1.15 0.68–1.94	0.601	1.30 0.80–2.11	0.290
Urban (Ref: Rural)	0.61 0.37–1.01	0.055	0.62 0.39–0.98	0.041
Age	1.04 0.87–1.25	0.623	1.08 0.91–1.28	0.385

Note: $lnP(Y_{\mathrm{i}}=j)/P(Y_{\mathrm{i}}=EcoActive)={\alpha }_{\mathrm{j}}+{\beta }_{1\mathrm{j}}Mal{e}_{\mathrm{i}}+{\beta }_{2\mathrm{j}}Urba{n}_{\mathrm{i}}+{\beta }_{3\mathrm{j}}Ag{e}_{\mathrm{i}}$; Pseudo R² = 0.217 (Note: Coefficients represent log-odds/OR compared to the Eco-Active reference cluster).

).

Male gender significantly increased the odds of being classified as a Passive Consumer (OR = 2.13, p = 0.001). Urban residence, compared to rural, reduced the odds of belonging to the Moderate Participants group (OR = 0.62, p = 0.041). Age showed no significant effect, suggesting that within this relatively narrow Generation Z cohort, maturity plays a limited role in differentiating sustainability engagement. Overall, these results reinforce gender and residential context as key differentiators of sustainability profiles among young people.

Further analysis by gender (

Table 5. Gender-specific OLS regression results (male vs. female subsamples).

Variable	Male β	p-Value	Female β	p-Value
Constant	55.602	<0.001	60.117	<0.001
Urban	4.118	0.032	3.528	0.048
Age	0.162	0.081	0.195	0.046
R2	0.138		0.151
N	183		245

Note: For males: $SB{I}_{\mathrm{i}}={\beta }_{0\mathrm{m}}+{\beta }_{1\mathrm{m}}Urba{n}_{\mathrm{i}}+{\beta }_{2\mathrm{m}}Ag{e}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$; for females: $SB{I}_{\mathrm{i}}={\beta }_0\mathcal{f}+{\beta }_1\mathcal{f}Urba{n}_{\mathrm{i}}+{\beta }_2\mathcal{f}Ag{e}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$.

) revealed that urban environment and age significantly predicted higher sustainability engagement among females (β = 3.53, p = 0.048; β = 0.20, p = 0.046). Among males, urban residence also showed a significant positive effect (β = 4.12, p = 0.032), while age effects were weaker (p = 0.081). These differences indicate that environmental motivation and behavioural consistency may be gender-specific, reinforcing the need for tailored communication and educational strategies that address distinct motivational patterns.

4.2. Inferential Analyses: Behavioural Profiles, Predictors, and Cognitive–Digital Interactions

To test the hypothesised relationship between digital engagement and sustainability, correlation analyses were performed (

Table 6. Correlations among digital-energy indicators and SBI.

Variable	WZC_Hours	WNS_Mean	WPU_Mean	WRH_Score
WZC_hours	0.412 ***	-	-	-
WPU_mean	0.298 ***	0.526 ***	-	-
WRH_score	0.154 *	0.128	0.085	-
SBI_pct	−0.213 **	−0.327 ***	−0.241 **	−0.092

Note: Significance levels: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Negative correlations indicate that greater digital exposure and fatigue correspond with lower sustainability behaviour.

). The results revealed a coherent pattern: greater digital intensity (WZC) and fatigue (WNS) correlated with higher problematic use (WPU; r = 0.30–0.41) and lower sustainability behaviour (r = −0.09 to −0.33). These associations confirm that heavy digital exposure is systematically accompanied by fatigue, diminished self-control, and reduced engagement in sustainable practices. The findings empirically support the concept of cognitive-resource dispersion—that is, digital overstimulation limiting capacity for mindful, sustainability-oriented action.

Table 7. PCA on combined variables (four digital-energy indicators + SBI)—loadings.

Variable	PC1	PC2	PC3	PC4
WZC_hours	0.558	0.217	−0.789	0.146
WNS_mean	0.667	−0.391	0.120	0.614
WPU_mean	0.601	−0.312	0.450	−0.585
WRH_score	0.234	0.873	0.091	−0.412
SBI_pct	−0.434	0.012	0.396	0.792

Note: Explained variance ratio: PC1 = 0.356, PC2 = 0.296, PC3 = 0.182, PC4 = 0.165 (Cumulative ≈ 1.000).

presents the results of the Principal Component Analysis combining the digital-behavioural indicators (WZC, WNS, WPU, WRH) with the Sustainability Behaviour Index (SBI). Two dominant components explained approximately 65% of total variance. The first component reflected a digital fatigue–sustainability trade-off, characterised by strong positive loadings on digital intensity and negative loadings on SBI. The second component, dominated by WRH, represented a risk-oriented digital behaviour dimension.

This latent structure supports the study’s conceptual model, illustrating that digital and sustainable behaviours are interrelated yet distinct domains. The inverse relationship between digital engagement and SBI provides clear empirical evidence for the cognitive and behavioural trade-off proposed in the theoretical framework.

To further examine these effects, a robust OLS model was estimated (

Table 8. OLS regression predicting SBI (dependent variable: SBI_pct)—robust HC3 SEs.

Variable	Coef. (β)	Std. Err.	t	p-Value	95% CI (Lower–Upper)
Constant	72.421	2.984	24.28	<0.001	66.556–78.286
WZC_hours	−0.912	0.211	−4.32	<0.001	−1.326–−0.498
WNS_mean	−9.843	1.124	−8.76	<0.001	−12.048–−7.638
WPU_mean	−6.351	1.432	−4.44	<0.001	−9.160–−3.542
WRH_score	−1.842	0.863	−2.14	0.033	−3.535–−0.149
Male	−3.217	1.012	−3.18	0.002	−5.212–−1.222
Age	−0.124	0.206	−0.60	0.548	−0.528–0.280
Urban	2.118	0.987	2.15	0.032	0.178–4.058
Class_year	0.321	0.411	0.78	0.434	−0.487–1.129

Note: $SB{I}_{\mathrm{i}}={\beta }_0+{\beta }_{1}WZ{C}_{\mathrm{i}}+{\beta }_{2}WN{S}_{\mathrm{i}}+{\beta }_{3}WP{U}_{\mathrm{i}}+{\beta }_{4}WR{H}_{\mathrm{i}}+{\beta }_{5}Mal{e}_{\mathrm{i}}+{\beta }_{6}Ag{e}_{\mathrm{i}}+{\beta }_{7}Urba{n}_{\mathrm{i}}+{\beta }_{8}ClassYea{r}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$; Model fit: R² (unadjusted) ≈ 0.289 (Note: robust SEs used).

).

The inclusion of digital-behavioural predictors significantly increased explanatory power (R² = 0.289). All four indicators had negative coefficients:

• WZC (β = −0.91, p < 0.001);
• WNS (β = −9.84, p < 0.001);
• WPU (β = −6.35, p < 0.001);
• WRH (β = −1.84, p = 0.033).

These findings demonstrate that digital overload, psychological fatigue, and impulsive behaviours are strong negative predictors of sustainability engagement, even when controlling for gender, age, and residence.

Demographic results confirmed that male participants scored significantly lower (β = −3.21, p = 0.002), while urban students exhibited slightly higher sustainability (β = 2.12, p = 0.032). In sum, the model validates that digital self-regulation capacity is a key determinant of sustainable behaviour among young people.

Finally, a moderation analysis tested whether education level (approximated by class year) buffers the negative impact of digital fatigue (WNS) on sustainable behaviour (

Table 9. Moderation analysis—does education (Class_year) moderate WNS → SBI?

Variable	Coef. (β)	Std. Err.	t	p-Value	95% CI
Constant	71.102	3.112	22.86	<0.001	64.961–77.244
WNS_mean	−10.230	1.215	−8.42	<0.001	−12.612–−7.848
Class_year	0.410	0.512	0.80	0.425	−0.593–1.413
WNS_mean × Class_year	0.428	0.137	3.12	0.002	0.159–0.697
Male	−3.105	1.005	−3.09	0.002	−5.082–−1.128
Age	−0.098	0.201	−0.49	0.621	−0.494–0.298
Urban	2.037	0.985	2.07	0.038	0.104–3.970

Note: $SB{I}_{\mathrm{i}}={\beta }_0+{\beta }_{1}WN{S}_{\mathrm{i}}+{\beta }_{2}ClassYea{r}_{\mathrm{i}}+{\beta }_3(WN{S}_{\mathrm{i}}\times ClassYea{r}_{\mathrm{i}})+{\beta }_{4}Mal{e}_{\mathrm{i}}+{\beta }_{5}Age\mathrm{i}+{\beta }_{6}Urba{n}_{\mathrm{i}}+{\epsilon }_{\mathrm{i}}$; OLS with interaction term WNS_mean × class_year.

).

The interaction term WNS × Class_year was positive and significant (β = 0.43, p = 0.002), indicating that educational advancement reduces the adverse influence of digital exhaustion on sustainability engagement.

This pattern suggests that educational maturity and sustainability literacy enhance cognitive resilience, enabling students to maintain sustainable practices despite high digital intensity.

The result aligns with the hypothesis that energy and attention management skills improve through education, mitigating the cognitive overload associated with constant digital use.

4.3. Summary of Findings and Interpretation

Across all analyses, the results consistently support the main hypotheses. Sustainable behaviour among Generation Z forms a coherent construct shaped by both social context (gender, residence, education) and cognitive-digital dynamics (fatigue, problematic use). The data reveal a clear behavioural trade-off between digital intensity and sustainability orientation—a finding that underscores the psychological dimension of sustainability in the digital era.

From a social-scientific perspective, these results highlight the importance of integrating digital well-being education with sustainability curricula. Strengthening students’ capacity for self-regulation, critical media use, and cognitive balance appears essential for promoting long-term sustainable citizenship among youth.

5. Discussion

Importantly, the discussion focuses on environmental sustainability behaviours as analytically distinct from, yet conceptually embedded within, the broader social and economic sustainability framework. This study set out to examine the behavioural sustainability of Generation Z, exploring the interrelations between sustainable practices, digital lifestyles, and cognitive resources. The findings provide strong empirical support for the proposed behavioural-energy framework, confirming that energy management extends beyond physical systems into the domains of everyday behaviour, cognition, and digital consumption.

5.1. Interpretation of Key Findings

The results confirmed the first hypothesis (H1). Principal Component Analysis demonstrated that the four key sustainable actions—waste segregation, reuse, second-hand purchasing, and consumption moderation—formed a coherent, unidimensional construct, the Sustainability Behaviour Index (SBI). This validates the conceptualisation of behavioural sustainability as a latent trait reflected across multiple pro-environmental practices.

This conclusion aligns with prior work showing that diverse sustainable actions often stem from shared psychological foundations such as biospheric values, self-identity, and personal norms (Balundė et al. 2020; Bhattarai et al. 2024; Denault et al. 2024; Muyulema-Masaquiza and Ayala-Chauvin 2025). Treating these four behaviours as manifestations of a single latent construct is therefore both empirically justified and theoretically grounded.

Cluster analysis (H2) revealed three distinct behavioural profiles—Eco-Active, Moderate Participants, and Passive Consumers—confirming that sustainability engagement varies along a continuum rather than representing a binary state.

This typology parallels previous segmentation research on sustainable consumption. Similar threefold divisions of consumers—such as Doers, Conscious, and Reluctant—have been reported in studies on food consumption (Gazdecki et al. 2021), while other analyses distinguish convenience-driven and value-driven energy users (Słupik et al. 2021). Comparable attitudinal clustering has also been observed in environmental behaviour research, identifying groups ranging from Environmentalists to Inactivators (Yoon and Ahn 2020), which underscores the need for differentiated rather than uniform intervention strategies. The existence of the “Moderate” cluster in this study further suggests that partial or selective engagement is typical among youth who support sustainability ideologically but not behaviourally.

The analysis of socio-demographic predictors (H3) confirmed that gender and place of residence are significant determinants of sustainability engagement, whereas age played a minor role within this narrow age group. Female respondents and urban residents reported higher SBI scores—consistent with previous evidence that women exhibit stronger environmental concern and more frequent pro-environmental actions (Abrahamse et al. 2005; Li et al. 2022; Steg 2023; Tien and Huang 2023; Vicente-Molina et al. 2018).

This gender difference is consistent with socialisation theory, which posits that women are more often oriented toward care and community responsibility. Similarly, studies indicate that urban residents display higher sustainability engagement, supported by environmental infrastructure and social norms (Alcock et al. 2020; Cheng and Mao 2024; Dąbrowski et al. 2022). The multinomial regression analysis further confirmed that men were significantly more likely to belong to the “Passive Consumers” cluster, underscoring persistent gendered differences in environmental responsibility.

The strongest empirical support concerns H4, linking digital activity and sustainability. Negative correlations and regression coefficients for all digital-energy variables (WZC, WNS, WPU, WRH) reveal a clear behavioural trade-off: higher digital intensity and cognitive fatigue consistently predict lower sustainability engagement. This finding substantiates the concept of cognitive energy dispersion, suggesting that the mental and attentional demands of the digital environment deplete the very self-regulatory resources required for mindful, pro-environmental decisions (Baumeister et al. 1998; Freed et al. 2025; Sweigart et al. 2025). Empirical parallels are found in research on social media fatigue and cognitive depletion, where digital overload reduces attention and self-control capacities (Wang et al. 2025). Similarly, studies on digital distraction and knowledge work show that multitasking consumes cognitive resources and limits sustained behavioural engagement (Jarrahi et al. 2023).

The PCA results (Table 7) visualised this antagonism—placing high digital load and low sustainability scores on opposite poles of the same latent dimension—confirming that digital overload can hinder sustainable behaviour among youth.

The moderation analysis confirmed H5, indicating that education level mitigates the negative relationship between digital fatigue and sustainability. This finding resonates with research showing that educational attainment enhances energy literacy and resilience in decision-making under cognitive load (Santillán and Cedano 2023). Similarly, recent work demonstrates that cognitive maturity and sustainability education foster resistance to digital overstimulation and strengthen environmental self-efficacy (Al Mulhim and Zaky 2023; Awdziej et al. 2023; Ji et al. 2023; Khanna et al. 2024).

Thus, education appears to serve as a protective factor, supporting self-regulation and preserving sustainable engagement despite digital fatigue.

5.2. Theoretical and Practical Implications

Theoretically, this study advances an integrative model combining environmental psychology, behavioural energy economics, and digital media research. It shows that the concept of “energy” can be extended beyond the physical to the behavioural and cognitive domains, capturing efficiency, conservation, and overuse within human decision-making.

This framework offers a conceptual bridge between sustainability and digital well-being, positioning cognitive regulation as a central mechanism of sustainable citizenship in the digital age (Hassoun et al. 2025).

Practically, the findings highlight the need to incorporate behavioural segmentation into sustainability policy and education. Previous energy research emphasises that segmentation enhances the effectiveness of conservation programmes by aligning interventions with behavioural profiles (Muyulema-Masaquiza and Ayala-Chauvin 2025).

The observed negative link between digital overuse and sustainability further suggests that promoting digital sufficiency and digital well-being (Santarius et al. 2023) can yield dual benefits: reducing cognitive strain and strengthening environmental responsibility. Interventions should therefore extend beyond informational campaigns and focus on developing self-regulation skills that enable young people to manage digital energy use consciously.

Moreover, the moderating effect of education underlines the importance of integrating digital energy awareness and critical reflection on ICT’s environmental cost directly into school curricula (Santillán and Cedano 2023; Shengjergji et al. 2024). Such integration would align with the broader UN agenda on Education for Sustainable Development.

Addressing the “educational paradox” requires not only reducing digital activity but also improving its efficiency. Examples include promoting data-light practices such as sharing cloud links instead of duplicating files, optimising storage use, and raising awareness of the hidden energy costs of digital convenience.

5.3. Limitations and Future Research

This study has several methodological and contextual limitations that should be considered when interpreting the findings: First, its cross-sectional design does not allow the establishment of causal relationships between digital behaviours (such as fatigue or problematic use) and sustainable practices. Longitudinal or experimental studies would be necessary to clarify the temporal dynamics and identify potential bidirectional effects. Longitudinal and experimental designs could further test whether cognitive resource competition between digital engagement and sustainability behaviour persists over time and across cultural or generational contexts. Second, the sample was limited to Polish secondary-school students, which constrains the generalizability of results to broader or cross-cultural populations. Future research should include comparative studies involving diverse educational contexts or different segments of Generation Z across multiple countries, which is increasingly important in globalised digital environments (Istrate et al. 2024).

Third, the reliance on self-reported measures introduces the possibility of social desirability bias or underreporting, particularly in sustainability-related behaviours. Combining self-reports with objective measures, such as screen-time logs or learning analytics datasets, would improve the validity of the assessment (Manner 2023). Additionally, while the Sustainability Behaviour Index (SBI) demonstrated strong psychometric qualities, it could be expanded in future studies to cover other critical sustainability domains, including mobility, household energy use, and civic participation in climate actions.

Finally, future research would benefit from integrating behavioural data with system-level modelling approaches, enabling the exploration of how micro-behaviours of young citizens scale up to impacts on energy systems or environmental footprints, thus linking individual digital practices with global sustainability transitions (Istrate et al. 2024; Manner 2023). Such mixed-method approaches would build upon emerging frameworks in energy and sustainability sciences that merge behavioural insights with resource accounting and life-cycle analysis, creating a more comprehensive understanding of digital sustainability across scales.

6. Conclusions

This study contributes to the growing body of research on the behavioural dimensions of sustainability by highlighting how digital fatigue and problematic online use may limit pro-environmental engagement among adolescents. By combining k-means clustering with robust OLS regression, it offers a methodological framework for identifying multidimensional behavioural profiles and their predictors within youth populations. This extends prior work on eco-behavioural segmentation and bridges a gap in sustainability research that has often overlooked the role of digital strain in everyday ecological practices.

The findings also advance the literature by showing that digital well-being is not merely a psychological or cognitive concern, but one that has tangible consequences for environmentally responsible behaviour. Unlike previous studies that treat sustainability and digital life as parallel domains, this article empirically anchors their intersection, revealing that excessive or draining digital engagement can detract from pro-ecological actions, especially among younger generations.

From a policy perspective, the results underscore the need to integrate digital well-being into sustainability education programmes. As schools and policymakers in Europe increasingly support initiatives such as the EU’s Digital Education Action Plan and the GreenComp competence framework, balancing digital innovation with mindful consumption becomes an imperative. Targeted interventions—such as digital wellness curricula or eco-digital literacy campaigns—can equip young people with the capacity to navigate digital environments without compromising sustainable lifestyles.

In light of ongoing digital transformations, these findings call for collaborative action among educational institutions, technology developers, and policymakers to foster digital cultures that are both mentally healthy and ecologically responsible. Research in physics education and STEM pedagogy can play a crucial role in translating abstract energy concepts into concrete digital sustainability practices for students. Future research should continue to explore the intersection of youth digital behaviour, policy developments, and sustainability transitions on a global scale. Future research should extend this framework by incorporating social and economic sustainability dimensions, enabling a more comprehensive assessment of sustainable citizenship in the digital age.