Self-regulated learning and student engagement as predictors of environmental awareness in higher education

1 Introduction

Contemporary university education faces the challenge of integrating the cognitive, emotional and ethical development of students in a context characterized by the intensive use of digital technologies and by an environmental crisis of global scope (Leal Filho et al., 2018). In this scenario, higher education must not only promote conceptual learning, but also favor psychological dispositions that guide students toward responsible practices with the environment, particularly in the face of the environmental impact associated with the operation of technological infrastructures such as data centers, energy consumption and intensive use of water (Masanet et al., 2020). From this perspective, environmental awareness is configured as a complex educational outcome, influenced by deep psychological processes and not only by the acquisition of declarative knowledge, which requires rethinking the traditional explanatory models used in environmental education and moving toward more integrative approaches (Otto and Pensini, 2017).

Recent literature has shown that emotions play an important role in the transformation of environmental knowledge into attitudinal and behavioral dispositions (Brosch, 2021). Studies in environmental psychology and education have shown that positive emotions such as interest or empathy can strengthen involvement with environmental problems, while negative emotions such as worry or guilt can favor change-oriented reflective processes when activated in structured educational contexts (Ojala, 2012; Verplanken and Roy, 2013). In this sense, emotions are not conceived as isolated learning outcomes, but as psychological mechanisms that mediate the way in which students process, interpret and resignify environmental information, indirectly influencing the development of environmental awareness (Clayton et al., 2015).

This emotional dimension is closely linked to the processes of self-regulation of learning (Ben-Eliyahu, 2021). From the theoretical framework of the Motivated Strategies for Learning Questionnaire (MSLQ), self-regulation is understood as the student’s ability to plan, monitor and evaluate their goals, effort and cognitive-emotional states in the face of complex tasks (Pintrich, 1988, 2004). Previous research has shown that students with higher levels of self-regulation tend to exhibit greater persistence, critical reflection, and coherence between attitudes and behaviors (Dent and Koenka, 2016), including those related to sustainability. From this perspective, self-regulation not only optimizes academic performance, but also facilitates the internalization of environmental values and the translation of knowledge into consistent behavioral dispositions, which allows us to understand environmental awareness as an extension of well-established self-regulatory processes.

In a complementary way, the emotional and behavioral commitment of the student, conceptualized from the Student Engagement (SED) framework proposed by Skinner, provides a key experiential dimension to explain active involvement with educational content (Skinner et al., 2008). Academic engagement has been associated with higher levels of participation, interest, and sense of belonging, as well as a greater willingness to act in a manner consistent with the values promoted in the educational environment (Fredricks et al., 2016; Kahu and Nelson, 2018). However, evidence suggests that its effect is more consistent when it is articulated with deep cognitive processes, such as self-regulation of learning, which allow reflection to be sustained and emotional involvement to be given continuity in the face of environmentally complex content.

In this context, educational experiences mediated by collective technological visualization, including extended reality (XR) resources and shared visual representations, emerge as pedagogical devices with a high potential to activate relevant emotional and motivational processes (Radianti et al., 2020). Although not all students interact individually with immersive environments, evidence indicates that the collective visualization of environmental scenarios such as the operation of data centers, water consumption, or the ecological footprint of digital technology is sufficient to generate significant emotional responses that influence the perception of the environmental impact of technology use and pro-environmental attitudes (Nelson et al., 2020). These experiences make it possible to make visible environmental externalities that are usually invisible, connecting the daily use of technology-supported environmental learning with its ecological consequences and favoring student involvement (Breves and Schramm, 2021).

Based on these elements, the present study proposes an explanatory model of a predictive nature, based on multiple linear regression, in which environmental awareness is conceptualized as a dependent variable, while self-regulation of learning (MSLQ) and emotional and behavioral engagement (SED) are established as predictor variables (Hair et al., 2019). Academic emotions are integrated into the model as theoretical mechanisms that contextualize the educational experience mediated by technological visualization, without being modeled as direct independent variables. This approach allows us to move beyond descriptive analyses, offering an empirically grounded explanation of how cognitive and motivational processes interact to shape environmental awareness in university students. In the following section, the theoretical framework that formally supports this analytical model is developed.

2 Theoretical framework

2.1 Self-regulation of learning as a cognitive-emotional basis of environmental awareness

Self-regulation of learning constitutes a theoretical framework to explain how students plan, monitor and evaluate their own cognitive and emotional processes in the face of complex tasks, including those related to socio-environmental problems (Panadero, 2017). From Pintrich (2004) proposal, self-regulation integrates motivational components, cognitive strategies and emotional control, which allows us to understand why some students manage to translate environmental knowledge into consistent behaviors, while others remain at a merely declarative level. In this sense, self-regulation not only improves academic performance, but also acts as a mechanism that facilitates the internalization of values and responsible decision-making, which is essential for the formation of a solid environmental awareness (Ben-Eliyahu, 2021). This self-regulatory ability prepares the student to process complex environmental information without avoiding it emotionally, thus connecting the cognitive dimension with the behavioral dimension.

Various studies have shown that students with high levels of self-regulation have greater persistence in the face of content that generates emotional dissonance, such as those associated with the environmental impact of human and technological activity (Virtanen et al., 2021). The regulation of goals, the control of effort and the management of negative emotions allow the student not to block out environmental information, but to process it in a reflective and action-oriented way (Lin et al., 2019). From this perspective, self-regulation becomes a key antecedent of environmental awareness, by enabling the student to maintain the cognitive and emotional commitment necessary to understand the environmental implications of the intensive use of digital technologies (Christopoulos et al., 2020; Li et al., 2019). This role explains why self-regulation is proposed in this study as a central predictor variable.

2.2 Emotional and behavioural engagement as a driver of pro-environmental involvement

Student engagement, conceptualized from Skinner’s Student Engagement model, provides a relational and motivational dimension that complements the self-regulation of learning (Sinatra et al., 2015; Skinner et al., 2008). Emotional and behavioral engagement is manifested in active participation, sustained interest, and student willingness to engage with educational content and experiences, elements that have been consistently associated with lasting attitudinal and behavioral changes (Wang and Eccles, 2013). In the environmental field, this commitment translates into greater sensitivity to ecological impacts and a greater likelihood of adopting responsible behaviour (Bouman et al., 2018). Thus, engagement does not act solely as an outcome of learning, but as a catalyst for learning (Kahu et al., 2020).

Empirical evidence suggests that emotional engagement intensifies the connection between the educational experience and the student’s personal values, facilitating the appropriation of environmental narratives and the perception of individual and collective responsibility (Bissing-Olson et al., 2016). When students feel emotionally involved, environmental information is no longer perceived as external or abstract, becoming a personally relevant issue (Schneider et al., 2021). This implication is especially relevant in technological contexts, where the environmental impact often remains invisible. In this way, student engagement emerges as a fundamental predictor of environmental awareness, complementing self-regulatory processes from a more experiential dimension (Reeve, 2012).

2.3 Academic emotions as mediators between technology-supported environmental learning and environmental awareness

Academic emotions play a critical mediating role in the relationship between learning, technology, and environmental behavior (Pekrun and Linnenbrink-Garcia, 2014). The literature on environmental psychology has shown that both positive and negative emotions have a differentiated influence on the willingness to act pro-environmentally (Taufik et al., 2015). Positive emotions such as interest or hope strengthen intrinsic motivation (Ojala, 2012), while negative emotions such as guilt or worry can activate reflective processes aimed at change when they are properly regulated. In educational contexts, these emotions are activated especially when students are confronted with the real consequences of their actions or decisions (Markowitz et al., 2018).

In this framework, educational experiences mediated by extended reality and collective technological visualization take on a strategic role, as they allow for a tangible representation of the environmental impacts associated with the use of digital technologies, such as water and energy consumption in data centers. Even when the experience is not immersive individually, shared on-screen viewing has proven to be sufficient to activate meaningful and socially shared emotional responses, amplifying their formative effect (Brosch, 2025). These emotions act as bridges between cognitive understanding and environmental awareness, reinforcing both self-regulation and commitment (Wamsler et al., 2021).

2.4 Environmental awareness as an integrative result of cognitive, emotional and technological processes

Environmental awareness, understood as the individual’s ability to recognize, evaluate and act in the face of the environmental impacts of their decisions, is configured as a multidimensional construct that integrates knowledge, emotion and behavior (Ardoin et al., 2020; Whitmarsh et al., 2021). Previous studies have shown that environmental awareness is strengthened when high levels of engagement, self-regulation, and emotional activation converge, particularly in educational contexts that use experiential and technological strategies (Luo et al., 2019; Trott, 2020). From this perspective, environmental awareness cannot be explained by a single isolated factor, but as the result of a system of interdependent variables (Gifford and Nilsson, 2014).

Beyond confirming previously documented associations, the incremental contribution of this study lies in jointly modelling self-regulated learning and student engagement as simultaneous predictors of environmental awareness within a single technology-mediated learning context. Whereas earlier research has typically examined self-regulation and pro-environmental or educational outcomes in isolation, the present specification estimates the relative weight of each construct when both are considered together, showing that self-regulatory processes rather than engagement carry the dominant explanatory role. This clarifies which psychological mechanism most warrants pedagogical attention in sustainability education and advances the literature from isolated bivariate relationships toward a parsimonious, empirically tested multivariable account.

Under this integrative approach, the present study proposes to analyze environmental awareness as a dependent variable, explained by the interaction between self-regulation of learning, emotional and behavioral engagement and emotional activation generated by educational technological experiences (Wamsler et al., 2021). This approach allows us to move toward an empirical model that recognizes the complexity of environmental learning in higher education, laying the foundations for the formulation of the regression model (Hayes, 2022) and the approach of the analytical model that is developed in the following section.

3 Approach to the analytical model

The central objective of this study is to analyze student environmental awareness as an explainable result from the interaction between self-regulatory and motivational cognitive processes, activated within the framework of technology-mediated environmental learning experiences (Panadero, 2017; Radianti et al., 2020). Unlike traditional descriptive approaches that have treated environmental awareness as an isolated or merely attitudinal construct, this paper is based on the assumption that this awareness is built from internal psychological mechanisms that guide the way in which students process, interpret and resignify environmental information in real educational contexts (Brosch, 2021). From this perspective, environmental awareness is conceived as a complex educational outcome, influenced by the self-regulation of learning and by the degree of emotional and behavioral involvement of the student (Sinatra et al., 2015).

In coherence with the contemporary literature in educational psychology and environmental education, the study adopts a quantitative approach of a correlational-predictive nature, using a multivariate regression model (Hair et al., 2019; Hayes, 2022). This approach allows estimating the relative contribution of different psychological predictors and evaluating their joint explanatory capacity, without assuming direct causal relationships (Fiedler et al., 2018; Tabachnick et al., 2019). In this way, the analysis is aimed at understanding how certain cognitive and motivational processes are associated with environmental awareness in university students, rather than establishing strict causal relationships.

3.1 Definition of the proposed regression model

In the proposed analytical model, environmental awareness is defined as the dependent variable (Y) and is measured by a study-specific scale, summarized in a global composite score, which assesses perceptions, attitudes, and dispositions related to the environmental impact of the use of digital technologies. The predictor variables (X) correspond to two psychological dimensions widely consolidated in the educational literature: (a) self-regulation of learning, operationalized through the MSLQ (Pintrich, 2004); and (b) the student’s emotional and behavioral engagement, as measured through the SED instrument proposed by Skinner et al. (2008).

These variables are integrated into an explanatory model that recognizes the differential influence of each predictor on environmental awareness, allowing their relative contribution to be analyzed within the same analytical framework (Tabachnick et al., 2019). Academic emotions are not incorporated as explicit independent variables, but are considered theoretical mechanisms that contextualize the educational experience mediated by technology, indirectly influencing the processes of self-regulation and student engagement (Pekrun and Linnenbrink-Garcia, 2014).

Formally, the multiple linear regression model is expressed as follows:

Environmental Awarenessi = β 0 + β 1 ( Self − Regulation MSLQi ) + β 2 ( SED Commitmenti ) + ε i

where represents the model constant, and the regression coefficients associated with each predictor variable, and the random error term β 0 β 1 β 2 ε i(Hayes, 2022). This model allows evaluating both the individual weight of each predictor and the global explanatory capacity of the set of independent variables, aligning with recent empirical approaches in environmental education and educational psychology that address pro-environmental behavior from integrative perspectives (Whitmarsh et al., 2021).

3.2 Role of academic emotions in the explanatory model

Although the statistical model is formulated as a direct regression between observable psychological variables, the present study explicitly recognizes the role of academic emotions as contextual mechanisms that influence the learning experience (Linnenbrink-Garcia and Pekrun, 2011). The literature has pointed out that positive emotions can strengthen intrinsic motivation and cognitive involvement (Reeve, 2012), while negative emotions, when they do not generate avoidance, can favor reflective processes aimed at environmental responsibility (Verplanken and Roy, 2013). In this sense, emotions are conceived as catalysts that amplify the effects of self-regulation and commitment on environmental awareness, without being treated as direct independent variables in the statistical model.

The educational experience analyzed, based on an immersive extended reality (XR) environment in which students interacted directly with three-dimensional scenarios depicting the environmental consequences of human technological activity, constitutes a context conducive to emotional activation (Pellas et al., 2021). This embodied and immersive interaction helps students to connect technological activity with its specific environmental consequences (Freitag et al., 2021). From this perspective, emotions provide an interpretative framework for reading the coefficients of the regression model and enrich the psychoeducational understanding of the results obtained (Schneider et al., 2021).

3.3 Analytical assumptions of the study

Based on the theoretical framework developed, this study is guided by three analytical assumptions that guide the construction of the model and the interpretation of the results (Gifford and Nilsson, 2014). First, it is assumed that self-regulation of learning is a relevant predictor of student environmental awareness, given its role in cognitive and emotional regulation in the face of environmentally complex content (De La Fuente et al., 2021). Secondly, it is considered that the emotional and behavioral commitment of the student contributes positively to environmental awareness, particularly when it is articulated with self-regulatory processes that allow reflection and involvement to be sustained over time (Kahu et al., 2020). Finally, it is assumed that the joint consideration of self-regulation and commitment will explain a significant proportion of the variance in environmental awareness, reflecting the multidimensional nature of the construct.

These assumptions are not presented as causal statements, but as analytical orientations that structure the predictive model adopted in the study. The following section presents the research methodology, detailing the design, the participants, the measurement instruments and the data analysis strategy used to empirically contrast the proposed model.

4 Methodology

4.1 Research design

The study adopts a quantitative, non-experimental design, with a correlational-predictive scope, with the aim of analyzing the explanatory capacity of psychological variables associated with learning about environmental awareness in university students (Creswell and Creswell, 2018). This type of design is appropriate when seeking to estimate associations and predictive relationships between latent constructs without direct manipulation of independent variables, particularly in real educational contexts where controlled experimental intervention is not feasible (Shadish et al., 2001). The predictive approach allows the relative weight of each variable to be identified in the explanation of the variance observed in environmental awareness, without assuming direct causal relationships (Hair et al., 2019).

The research was developed in an educational context mediated by technology, in which students participated in an immersive extended reality (XR) learning experience addressing the environmental consequences of human technological activity (Pellas et al., 2021). These experiences functioned as a common pedagogical context that favored emotional activation and environmental reflection, without constituting a formal experimental intervention, which reinforces the relevance of the non-experimental design and contributes to the ecological validity of the findings (Breves and Schramm, 2021).

4.2 Participants and study context

The sample was composed of 242 university students who answered three standardized psychometric instruments within the framework of an educational experience supported by immersive extended reality (XR) technology, a sample size considered adequate for multiple regression models with a reduced number of predictors (Hair et al., 2019). Participants were selected using non-probabilistic convenience sampling, a common strategy in correlational educational studies when working with intact groups of students (Etikan, 2016). All participants engaged with the same immersive XR materials under an identical protocol, which allowed the variability of the learning context to be partially controlled (Fiedler et al., 2018).

In demographic terms, participants had a mean age of 22.06 years (SD = 1.64; range 19–30). Regarding self-reported gender, 189 participants (78.1%) identified as male, 49 (20.2%) as female, 3 (1.2%) preferred not to disclose their gender, and 1 (0.4%) identified as a transgender woman. All participants were enrolled university students recruited from intact class groups. The predominance of male students and the convenience-based recruitment constrain the representativeness of the sample; these limitations are discussed explicitly in Section 6.1. The demographic characteristics of the sample are summarized in Table 1.

Table 1. Demographic characteristics of the sample (N = 242).

The educational environment was characterized by the use of immersive extended reality (XR) resources designed to make the environmental consequences of human technological activity tangible and experiential (Freitag et al., 2021). This context is particularly relevant for the analysis of environmental awareness, since it connects technological activity with its environmental externalities, favoring the emotional activation and critical reflection of the student (Van Der Linden et al., 2015).

To improve transparency and reproducibility, the educational experience is described in detail as follows. The intervention was conducted during regular class periods of the course taught by the lead instructor, who developed the educational materials and generously provided access to her student groups for the purposes of this experiment. Because multiple class groups participated, several sessions were conducted to ensure that all enrolled students could take part in the activity. Each session lasted approximately one hour and followed the same instructional protocol.

Unlike conventional audiovisual presentations, the intervention employed an extended reality (XR) environment in which students interacted directly with three-dimensional educational animations using head-mounted displays and motion-tracking gloves. The immersive learning materials were specifically designed by the course instructor to support environmental education through experiential interaction. The animations focused on sustainability-related topics and allowed students to manipulate virtual objects and explore environmental phenomena from a first-person perspective. For example, one of the scenarios simulated the accumulation of space debris orbiting the Earth, enabling students to observe floating debris around the planet, physically reach toward these objects, grasp them using the motion-tracking gloves, and reposition them within the virtual environment.

The XR experience was complemented by instructor-guided explanations and collective reflection on the environmental consequences of human technological activities. Students were encouraged to explore the virtual scenarios actively, discuss their observations, and connect the simulated experiences with real-world environmental challenges. The combination of immersive visualization, embodied interaction, and guided pedagogical discussion was intended to strengthen cognitive understanding while simultaneously fostering emotional engagement and environmental awareness. All sessions were delivered using the same hardware configuration, educational animations, and instructional sequence to ensure consistency across participant groups.

4.3 Measuring instruments

Data collection was carried out through three validated questionnaires widely used in the international literature, which guarantees the reliability and comparability of the results obtained. All items were rated on a five-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). The complete list of subscales and items for each instrument is provided in Appendix A. Composite scores were computed as the mean of the relevant items, and the internal consistency of every scale was estimated in the present sample using both Cronbach’s alpha and McDonald’s omega (reported below and summarized in Table 2).

Table 2. Descriptive statistics and internal consistency of the study measures (N = 242).

For statistical analyses, global composite scores were constructed by means of the average number of items per scale; In the case of student engagement, items previously recoded in the database were used, so that higher scores represented higher levels of emotional and behavioral engagement (DeVellis, 2016). Environmental awareness was operationalized using a specific scale of the study, calculated as a global composite score (average of items), where higher values indicate greater environmental awareness (Hayes, 2022).

4.3.1 Self-regulation of learning (MSLQ)

Self-regulation was measured using the MSLQ, developed by Pintrich and De Groot (2012), which evaluates motivational components, cognitive strategies and regulation of effort. The MSLQ has demonstrated adequate levels of internal consistency and validity in various educational contexts and has been used to analyze the relationship between self-regulation of learning, motivation and behaviors toward sustainability (Credé and Phillips, 2011). In this study, aggregate MSLQ scores were used as a global indicator of self-regulation of learning (Lavuri et al., 2023). In the present sample, the MSLQ composite showed excellent internal consistency (Cronbach’s alpha = 0.95; McDonald’s omega = 0.96).

4.3.2 Emotional and behavioral engagement (SED)

Student engagement was assessed using the SED instrument proposed by Skinner, which measures emotional and behavioral dimensions of academic engagement (Skinner et al., 2008). This instrument has been widely used to analyze student involvement and its relationship with sustainable attitudes and behaviors (Kahu et al., 2020). In the present study, the SED allowed to capture the level of emotional and behavioral involvement of students during the technology-mediated environmental learningexperience. Consistent with the scoring protocol of the instrument, disaffection items were reverse-coded so that higher values reflect greater engagement, and a composite engagement score was computed as the mean of the behavioural-engagement, emotional-engagement, and reverse-coded behavioural- and emotional-disaffection items. This composite showed good internal consistency (Cronbach’s alpha = 0.90; McDonald’s omega = 0.91).

4.3.3 Environmental Awareness Scale

Environmental awareness was assessed using an adapted version of the instrument reported by Yusliza et al. (2020), originally developed and validated to examine pro-environmental behaviour and sustainable development perceptions among adult populations. The instrument was selected because its theoretical foundation is consistent with contemporary conceptualizations of environmental awareness as a multidimensional construct encompassing cognitive understanding, affective concern, and behavioural disposition toward environmental sustainability (DeVellis, 2016; Yusliza et al., 2020).

The original scale developed by Yusliza et al. (2020) was grounded in the literature on environmental responsibility, sustainability, and pro-environmental behaviour. During its development, items were generated from established theoretical constructs and subjected to empirical validation procedures to assess their psychometric adequacy. For the present study, the items were reviewed by the research team to ensure linguistic clarity and contextual suitability for undergraduate students participating in the intervention. The adaptation process did not alter the underlying constructs measured by the original instrument.

The final version used in this study consisted of eight items assessing awareness of environmental issues, recognition of the environmental consequences of human activities, and perceptions of individual responsibility toward sustainable practices. Responses were recorded using a five-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). The resulting score was used as a dependent variable in the analytical model.

Additional evidence supporting the appropriateness of the scale was obtained from the current sample. The Environmental Awareness Scale demonstrated good internal consistency, with Cronbach’s alpha of 0.89 and McDonald’s omega of 0.91. Furthermore, item-total correlations were consistently positive and of moderate-to-high magnitude, indicating satisfactory homogeneity among items and supporting the internal structure of the instrument within this population. Because the scale was adapted from a previously validated instrument, the complete wording of the items used in the present study is reported in Appendix A to facilitate transparency and reproducibility.

4.4 Data collection procedure

The questionnaires were applied at the same time point, after the immersive extended reality (XR) educational experience (Creswell and Creswell, 2018; Setia, 2016). The students answered the instruments individually, guaranteeing anonymity and confidentiality of the information (Podsakoff et al., 2003; Tabachnick et al., 2019). Prior to the application, the participants were informed about the academic objectives of the study and the exclusively scientific use of the data, following basic ethical principles of educational research (Israel and Hay, 2006).

The simultaneous application of the instruments made it possible to capture in an integrated way the levels of self-regulation of learning, student engagement and environmental awareness, minimizing biases derived from contextual or temporal changes (Setia, 2016). This procedure is consistent with previous research that has analyzed relationships between psychological and environmental constructs in real educational contexts.

4.5 Data analysis strategy

The data analysis strategy was developed in several stages. First, descriptive analyses were performed to examine the distribution of the variables and verify basic assumptions of normality and absence of outliers (Hair et al., 2019). Secondly, the internal consistency of the instruments was evaluated by means of reliability coefficients, in order to ensure the psychometric quality of the scales used (Dunn et al., 2014).

Subsequently, Pearson correlation analyses were performed to explore the bivariate associations between self-regulation of learning, emotional and behavioral engagement, and environmental awareness (Cohen, 2013; Schober et al., 2018). Finally, a multiple linear regression model was estimated, in which environmental awareness acted as a dependent variable and self-regulation of learning (MSLQ) and student engagement (SED) as predictor variables. The model allowed estimating standardized beta coefficients and the coefficient of determination (R²), providing empirical evidence on the explanatory capacity of the proposed analytical model (Hair et al., 2019; Hayes, 2022). The results of this analysis are presented in the following section. Prior to interpreting the regression model, its statistical assumptions were examined: multicollinearity (variance inflation factors, VIF, and tolerance), independence of residuals (Durbin-Watson statistic), homoscedasticity (Breusch-Pagan test), normality of residuals (Shapiro–Wilk test and inspection of skewness and kurtosis), and linearity (inspection of partial-regression and residual plots). For all analyses, unstandardized (B) and standardized (beta) coefficients are reported together with their standard errors and 95% confidence intervals; confidence intervals for the correlation coefficients were obtained through Fisher’s z transformation.

5 Results

5.1 Descriptive statistics and reliability of the measures

Before testing the hypothesized relationships, descriptive statistics and internal-consistency coefficients were computed for the three composite measures. As shown in Table 2, all scales reached acceptable to excellent reliability (Cronbach’s alpha and McDonald’s omega > = 0.89), supporting the psychometric adequacy of the instruments in this sample.

5.2 Correlational analysis between the study variables

In order to explore the bivariate relationships between the central variables of the study, Pearson correlation analyses were performed between self-regulation of learning, emotional and behavioral engagement, and environmental awareness (Cohen, 2013; Schober et al., 2018). The results are presented in Table 3, which shows the pattern of associations between the constructs analyzed in a sample of 242 university students (Funder and Ozer, 2019).

Table 3. Pearson correlations between self-regulation of learning, student engagement, and environmental awareness (N = 242).

N = 242. *p < 0.05; ***p < 0.001. 95% confidence intervals (Fisher’s z): self-regulation-environmental awareness [0.26, 0.47]; engagement-environmental awareness [0.03, 0.27]; self-regulation-engagement [0.21, 0.43].

As can be seen in Table 3, self-regulation of learning has a positive correlation of moderate magnitude and statistically significant with environmental awareness (r = 0.37, p < 0.001), indicating that higher levels of planning, effort control, and cognitive-emotional regulation are associated with greater awareness of the environmental implications of technology use (Panadero, 2017). This result suggests that self-regulation is a relevant factor for environmental understanding and reflection in technology-mediated educational contexts (Wamsler et al., 2021). The 95% confidence interval for this correlation (r = 0.37) was [0.26, 0.47], confirming a robust positive association of moderate magnitude (p < 0.001).

In contrast, emotional and behavioural engagement showed a positive but weak correlation with environmental awareness (r = 0.15, 95% CI [0.03, 0.27], p = 0.019). Although statistically significant at the bivariate level, the magnitude of this association is small, and - as shown below - it becomes non-significant once self-regulated learning is entered simultaneously into the regression model. Self-regulated learning and student engagement were moderately intercorrelated (r = 0.32, 95% CI [0.21, 0.43]).

5.3 Regression model for predicting environmental awareness

Based on the results of the correlational analysis, a multiple linear regression model was estimated with the aim of evaluating the joint predictive capacity of learning self-regulation and emotional and behavioral engagement on environmental awareness. In this model, environmental awareness was established as a dependent variable, while self-regulation (MSLQ) and student engagement (SED) were included as predictor variables, allowing the estimation of standardized beta coefficients and global adjustment indicators such as the coefficient of determination (R²) (Tabachnick et al., 2019). The results of the model are presented in Table 4.

Table 4. Multiple linear regression model for prediction of environmental awareness (N = 242).

R² = 0.14; adjusted R² = 0.14; F(2, 239) = 20.09; p < 0.001. N = 242. B = unstandardized coefficient (standard error in parentheses); β = standardized coefficient; VIF = variance inflation factor.

As shown in Table 4, the regression model was statistically significant and accounted for approximately 14% of the variance in environmental awareness [R² = 0.14; adjusted R² = 0.14; F(2, 239) = 20.09, p < 0.001]. Examination of the standardized coefficients showed that self-regulated learning was the only significant predictor and carried the greatest explanatory weight [β = 0.37, 95% CI (0.24, 0.49), p < 0.001], whereas emotional and behavioural engagement made a negligible and non-significant unique contribution once self-regulation was held constant [β = 0.03, 95% CI (−0.09, 0.16), p = 0.620]. These results indicate that environmental awareness is associated mainly with self-regulatory processes, with student engagement adding little incremental explanatory value within the model (Sinatra et al., 2015; Whitmarsh et al., 2021).

The comparatively low explained variance merits explicit interpretation. Several factors account for it. First, environmental awareness is a multiply determined construct: it is shaped by environmental values and worldviews, prior environmental knowledge, connectedness to nature, affect and personality, and socio-demographic and cultural background, so a model built on only two psychological predictors is expected to capture a limited share of its variance. Second, the model was intentionally theory-driven and parsimonious - it tested the specific contribution of self-regulated learning and student engagement rather than maximizing prediction by adding many covariates - so a modest coefficient of determination is a design consequence, not a flaw. Third, effect sizes of this order (R² of roughly 0.10 to 0.20) are common and regarded as meaningful in educational and psychological research on complex attitudinal outcomes, where single predictors rarely explain large proportions of variance (Cohen, 2013). Fourth, part of the unexplained variance is likely methodological: the constructs were measured by self-report at a single time point, the environmental-awareness scores were high and somewhat restricted in range (which attenuates observable associations), and measurement error further lowers the observed R². Finally, the cross-sectional design captures concurrent associations rather than the cumulative developmental influences that build environmental awareness over time. Taken together, a modest R² does not weaken the central finding that self-regulated learning is a significant predictor; rather, it situates self-regulation as one meaningful contributor among the many determinants of environmental awareness, and indicates that future models should incorporate the omitted variables noted above.

Regression diagnostics supported the adequacy of the model. Multicollinearity was negligible (VIF = 1.12 for both predictors; tolerance = 0.89), and the residuals were independent (Durbin-Watson = 1.92). The Breusch-Pagan test indicated a slight departure from homoscedasticity (LM = 6.11, p = 0.047); accordingly, the analysis was cross-checked using heteroscedasticity-consistent (HC3) standard errors, which did not change the pattern of significance. The residuals showed some deviation from normality (Shapiro–Wilk W = 0.92, p < 0.001; skewness = −1.48; kurtosis = 9.26), driven by a small number of students reporting very low environmental-awareness scores; given the sample size (N = 242) and the robustness of ordinary least squares to non-normality under the central limit theorem, the estimates are considered stable. Finally, the comparatively modest explained variance (R² = 0.14) indicates that approximately 86% of the variance in environmental awareness is attributable to factors not captured by the model. This is consistent with the multifactorial nature of the construct and points to relevant omitted variables - such as environmental values, prior environmental knowledge, connectedness to nature, and socio-demographic background - that should be incorporated in future models.

5.4 Relationships between variables

Figure 1 shows the relationship between self-regulation of learning and environmental awareness, represented by a scatter plot with a linear trend line, a visual resource recommended to illustrate associations between continuous variables (Cohen, 2013).

The distribution of the data shows a clear positive trend, indicating that, as the levels of self-regulation increase, the levels of environmental awareness also increase, reinforcing the statistical results obtained in the correlational and regression analysis. This pattern is consistent with evidence linking self-regulatory processes with greater cognitive and behavioral integration in the face of complex environmental problems (Wamsler et al., 2021).

Figure 2 presents the relationship between emotional and behavioral engagement and environmental awareness, illustrated by a scatter plot that allows us to visualize the intensity and direction of the association between the two variables (Schober et al., 2018).

Although a general positive trend is observed, the dispersion of the data is greater, reflecting a weaker and more heterogeneous relationship between both variables. This visual pattern is consistent with the lower explanatory weight of engagement in the regression model and is consistent with research indicating that engagement, on its own, does not always predict complex attitudinal outcomes without the support of deep cognitive processes (Kahu et al., 2020; Sinatra et al., 2015).

Consequently, the results suggest that the contribution of commitment depends to a greater extent on its articulation with self-regulatory processes, reinforcing the integrative perspective that raises the interaction between motivational and cognitive dimensions in the explanation of environmental awareness (Panadero, 2017).

Overall, the empirical results support the proposed analytical model and confirm that student environmental awareness is mainly associated with self-regulation of learning, complemented by emotional and behavioral engagement. This finding is consistent with research highlighting the central role of self-regulation in internalizing values and translating knowledge into consistent behavioral dispositions (DeVellis, 2016), as well as studies that underline the complementary role of engagement in consolidating meaningful educational outcomes (Fredricks et al., 2016).

Likewise, these results provide an empirical basis for interpreting the role of educational experiences mediated by technological visualization as contexts that favor the activation of key psychological processes for environmental training in higher education, in line with the literature that indicates that technological environments can enhance emotional involvement and critical reflection when articulated with self-regulatory cognitive mechanisms (Brosch, 2021; Radianti et al., 2020).

6 Discussion

The results obtained show that environmental awareness in university students can be explained in a significant way from psychological variables linked to learning, particularly self-regulation of learning and, to a lesser extent, emotional and behavioral engagement, when these are activated in educational contexts mediated by technology (Lavuri et al., 2023). The estimated regression model explains 14.7% of the variance of environmental awareness (R² = 0.15), a value consistent with previous studies in environmental education and educational psychology, where behavior and pro-environmental awareness are recognized as complex and multifactorial phenomena (Gifford and Nilsson, 2014). This result supports the relevance of the proposed integrative approach and reinforces the idea that environmental awareness does not emerge from a single isolated factor, but from the interaction between cognitive and motivational processes (Ardoin et al., 2020).

In particular, self-regulation of learning showed the highest explanatory weight within the model (β = 0.37), as well as a positive and statistically significant correlation with environmental awareness (r = 0.37, p < 0.001). This finding confirms that students with greater capacities to plan, monitor and regulate their goals, effort and emotional states are more willing to recognize and reflect on the environmental implications of the use of technology (Verplanken and Roy, 2013). The moderate magnitude of this relationship is consistent with the literature that indicates that self-regulation facilitates cognitive persistence in the face of content that generates emotional dissonance, such as those related to the environmental crisis, allowing information to be transformed into critical reflection and potential pro-environmental action (Bissing-Olson et al., 2016; Li et al., 2019).

In contrast, students’ emotional and behavioural engagement did not make a significant unique contribution to the regression model (β = 0.03, p = 0.620). Although engagement correlated weakly but significantly with environmental awareness at the bivariate level (r = 0.15), its effect was fully absorbed by self-regulated learning when both predictors were modelled simultaneously. This pattern suggests that the influence of engagement on environmental awareness operates largely through, or jointly with, self-regulatory processes rather than independently of them (Sinatra et al., 2015).

The visualization of the bivariate relationships, represented in Figures 1, 2, provides graphical evidence that supports these statistical results. Figure 1 shows a positive trend between self-regulation of learning and environmental awareness, with a consistent distribution of data reinforcing the stability of this relationship (Nelson et al., 2020; Trott, 2020). On the contrary, Figure 2 shows a greater dispersion in the relationship between student engagement and environmental awareness, which reflects the heterogeneity of the effect of commitment and explains its lower explanatory weight in the regression model (Virtanen et al., 2021; Wang and Eccles, 2013). These visual representations allow contextualizing the coefficients obtained and strengthen the interpretation of the model from a psychoeducational perspective.

A central aspect of the discussion is the indirect role of academic emotions activated during the educational experience mediated by technological visualization. Although emotions were not modeled as explicit independent variables, the results should be interpreted considering that the experience of collective visualization of the environmental impacts associated with the use of digital technologies functions as an emotional trigger that favors student involvement and the activation of self-regulatory processes (Christopoulos et al., 2020; Pellas et al., 2021). Previous studies have shown that emotions such as worry, empathy, or even guilt, when presented in a structured educational setting, can strengthen moral reflection and environmental awareness rather than generate avoidance (Gil et al., 2021). In this sense, technology-supported environmental learning operates as an emotional mediator that makes environmental externalities that are usually invisible visible.

From a theoretical perspective, the findings support integrative models that articulate the theory of self-regulation of learning (Pintrich, 2004), academic engagement (Skinner et al., 2008) and environmental psychology (Yusliza et al., 2020), providing empirical evidence on their interaction within the same explanatory framework. Unlike studies that have addressed these constructs in isolation, the present work demonstrates that environmental awareness can be understood as an emerging result of complex educational systems, where cognitive regulation plays a central role and emotional engagement acts as a contextual amplifier (De La Fuente et al., 2021; Pekrun and Linnenbrink-Garcia, 2014). This approach contributes to bridging the gap between educational psychology and sustainability education at the higher level.

These findings also resonate with recent interdisciplinary work showing how learning and knowledge construction can integrate cognition and emotion through modelling-based pedagogical tools. Perissi (2025), for instance, illustrates how visual and dynamic modelling can act as a path to knowledge that couples affective engagement with conceptual understanding - a mechanism consistent with the role attributed here to technology-mediated visualization in activating environmentally relevant academic emotions and supporting self-regulated processing of sustainability content.

A terminological clarification is also warranted. Although the terms “prediction” and “predictors” are used throughout this article, the cross-sectional and correlational nature of the design means that they denote statistical (associative) prediction rather than causal or temporal prediction. The reported coefficients should therefore be read as the relative strength of association among constructs measured at a single time point, not as evidence of causal or directional influence.

6.1 Limitations

Several limitations should be considered when interpreting these findings. First, the cross-sectional design and the simultaneous self-report assessment of all constructs at a single time point make the results susceptible to common method bias and social desirability effects, which can inflate the shared variance among measures (Podsakoff et al., 2003). Future research should separate the measurement of predictors and outcome in time, incorporate marker variables or validity scales, and combine self-reports with behavioural or multi-source indicators. Second, because all constructs were measured concurrently, the associations should be interpreted as correlational rather than causal. Third, the study relied on a convenience sample of 242 students that was predominantly male and drawn from a limited institutional context, which restricts the representativeness and generalizability of the findings; replication with larger, more balanced, and multi-institutional samples is needed. Fourth, the model accounted for a modest proportion of variance (R² = 0.14), indicating that relevant explanatory variables - such as environmental values, prior environmental knowledge, connectedness to nature, and socio-demographic characteristics - were not included and should be examined in future models. Finally, although the Environmental Awareness Scale was adapted from a previously validated instrument (Yusliza et al., 2020) and showed good internal consistency in this sample, its adaptation to the present educational context and student population warrants further validation across independent samples.

7 Conclusion

The present study confirms that environmental awareness in university students can be explained in a significant way from psychological processes associated with learning, particularly self-regulation of learning and, to a lesser extent, emotional and behavioral engagement, when these are activated in educational contexts mediated by technology. From a multiple linear regression model applied to a sample of 242 students, it was evidenced that the joint model explains a relevant proportion of the variance in environmental awareness, with self-regulation as the main predictor and commitment as a complementary positive contribution.

Specifically, self-regulation of learning emerged as the predictor with the greatest explanatory weight, showing a positive and statistically significant relationship at both the correlational and multivariate levels. This finding reinforces the idea that the student’s ability to regulate goals, effort, and cognitive-emotional processes constitutes a central component in transforming environmental knowledge into critical reflection and behavioral disposition. In this sense, environmental awareness cannot be understood only as an informative result, but as an extension of self-regulatory processes that allow cognitive commitment to be sustained in the face of environmentally complex and emotionally demanding problems.

Emotional and behavioral engagement showed a more moderate positive effect, suggesting that student engagement, while relevant, does not act as a strong independent predictor of environmental awareness when analyzed in isolation. However, their contribution within the multivariate model indicates that commitment functions as a complementary factor that enhances the effect of self-regulation, particularly in contexts where educational experiences generate emotional activation and meaningful participation. This result provides empirical evidence to nuance approaches that assume engagement as a direct predictor of pro-environmental behavior, underlining the need to integrate it with deeper cognitive processes.

From an educational perspective, the findings of the study suggest that higher education institutions can promote the development of critical environmental awareness by designing pedagogical experiences that explicitly strengthen the self-regulation of learning and foster the emotional involvement of the student. In particular, collective technological visualization experiences are presented as a viable pedagogical strategy to make visible the environmental externalities associated with the intensive use of digital technologies, activating emotional mechanisms that favor reflection without necessarily requiring high-cost individualized immersive interventions.

In theoretical terms, the study contributes to the literature by proposing and empirically validating an integrative model that articulates the theory of self-regulation of learning, student engagement and environmental psychology in the same explanatory framework. This contribution allows us to move toward a more complex understanding of environmental awareness as an emerging result of educational systems that combine cognition, emotion and technological mediation, overcoming one-dimensional approaches focused exclusively on declarative knowledge.

Overall, the need to rethink education for sustainability at the higher level is reinforced, not only in terms of curricular content, but also in relation to the psychological processes that are activated during learning. Integrating technology-supported environmental learning, self-regulation and student engagement is thus emerging as a strategic way to train professionals capable of critically understanding and responsibly assuming the environmental impacts of technology in contemporary society. Finally, given the cross-sectional and correlational design, these conclusions are framed in associative rather than causal terms; experimental and longitudinal studies are needed to establish the directionality of the relationships observed.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frsus.2026.1877871/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Erkut Altındağ, Doğuş University, Türkiye

Reviewed by: Ilaria Perissi, Istituto di Chimica dei Composti Organo Metallici Consiglio Nazionale Delle Ricerche Sezione di Firenze, Italy

Zeyad Aly Khalil, Obour Higher Institute for Management and Information Technology, Egypt