Evaluating the Influence of Cognitive Engagement, Academic Performance, and Self-Regulated Learning on the Adoption of Artificial Intelligence–Driven Adaptive Learning Systems among University Students in Sri Lanka
Jul 2026· Critical Journal of Social Sciences· Vol 2· 0 citations· 22 references
TL;DR
The results indicated that AI adoption had a significant and positive effect on self-regulated learning and academic performance as opposed to cognitive engagement which was found to have a less significant effect on AI adoption in the regression model.
Abstract
Adaptive learning systems (with the role of artificial intelligence, AI) are increasingly becoming a part of higher education to provide personalization of the content, dynamic learning pathways, and timely feedback on student progress. This paper investigates the effect of cognitive engagement, academic performance and self-regulated learning in adopting AI-based adaptive learning systems among students in a Sri Lankan university. The analysis relies on the theory of educational technology adoption and learner-centered theory, the study examines the possibility that the learning behaviors and learning attributes of students influence their adoption and use of AI-enhanced learning conditions. The survey data involving 100 students at a university was used in a quantitative cross-sectional study that was conducted to collect information on the topics of AI-based learning tools exposure and its effect on the students. The analysis of the data was carried out through descriptive statistics, reliability analysis, Pearson correlation, and multiple regression of SPSS. The results indicated that AI adoption had a significant and positive effect on self-regulated learning and academic performance as opposed to cognitive engagement which was found to have a less significant effect on AI adoption in the regression model. The model explains 48.1% of AI adoption, showing that self-directed learning ability and learning orientation influence AI use more than engagement in higher education.
This study aims to analyse the effects of learner autonomy, emotional intelligence, and student engagement on academic self-concept through the mediation of e-learning use amongst university students in Malang City, Indonesia. A quantitative approach with a cross-sectional survey design was employed on 944 students from 16 universities in Malang City, which are selected using purposive sampling. Data were collected using a closed questionnaire with a five-point Likert scale and analysed using ordinary least squares-based path analysis with four classical assumption tests. Results indicate that learner autonomy significantly affects e-learning use (β = .213, p < .001) and academic self-concept (β = .299, p < .001); emotional intelligence significantly affects e-learning use (β = .185, p < .001); student engagement is the strongest predictor of e-learning use (β = .348, p < .001) and directly affects academic self-concept (β = .311, p < .001); and e-learning use significantly affects academic self-concept (β = .282, p < .001). The main finding and novelty of this study is that emotional intelligence does not have a direct effect on academic self-concept (β = .011, p = .680), but exerts a significant and exclusive indirect effect through e-learning use (β = .052). The refined model yields R² = .559. This finding confirms e-learning as a psychological mediator that converts students’ emotional capacity into academic self-concept, which is a contribution that extends emotional intelligence theory in the context of digital higher education.
Jozua Ferjanus Palandi, Mardji, Eddy Sutadji et al.· Letters in Information Techn...· 0 citations
This study aimed to explore the correlation between the construct of digital self-efficacy and engagement of students in the context of AI-driven learning environments, while also investigating the potential mediation of academic motivation. The design applied was a descriptive correlational cross-sectional design. The study included a sample of 130 undergraduate students at Al Esraa University who had experience in using AI tools to help in the academic process. A 24-item questionnaire, assessed through three constructs - digital self-efficacy, academic motivation, and student engagement, was used for data collection. Means, standard deviations, Pearson's correlation coefficients, and mediation analysis using Hayes' PROCESS Macro, Model 4, were employed to analyze the data. The results indicated that students possessed high digital self-efficacy and were highly motivated and engaged in learning environments supported by AI. Academic motivation was the highest and positively significant with student engagement, and digital self-efficacy was positively and significantly associated with academic motivation and student engagement. The mediation analysis revealed that digital self-efficacy was partially mediated by academic motivation. The findings pointed out that students' confidence with the use of digital and AI-supported tools had a direct influence on the engagement of students, but also had an indirect effect through strengthening the academic motivation of the students. Overall, the study suggested steps to be implemented to enhance students' confidence, critical evaluation, and motivation while fostering meaningful and responsible interaction with the use of AI in their learning process.
Linda Ahmad Khateeb, R. Freihat· Journal of Ecohumanism· 0 citations
AI-mediated digital learning environments have transformed how students communicate with knowledge, access learning resources, and receive algorithmically generated feedback, thereby reshaping how they perceive the pace of academic progress.This study examines perceived academic acceleration as a learning outcome distinct from general achievement, focusing on students’ self-reported sense that they move through academic tasks and course materials more quickly when supported by digital tools. A cross-sectional survey model was used to test whether technology-enhanced learning frequency was associated with perceived academic acceleration directly and indirectly through self-efficacy, and whether learning motivation conditioned the technology-acceleration association. Item-level data from 420 undergraduate students were analyzed using reliability analysis, correlation analysis, measurement-model diagnostics, ordinary least squares regression, bootstrap indirect-effect estimation, moderation analysis, and collinearity diagnostics. Results indicated that technology-enhanced learning frequency, self-efficacy, and learning motivation were positively associated with perceived academic acceleration. The indirect association through self-efficacy was supported by bootstrap confidence intervals, suggesting that students who reported more frequent technology-enhanced learning within AI-mediated learning environments also reported stronger efficacy beliefs, which in turn were associated with stronger perceived academic acceleration. The interaction between technology-enhanced learning frequency and learning motivation was positive but did not reach conventional significance. The findings should therefore be interpreted as evidence of associations among students’ technology use, efficacy beliefs, motivation, and perceived acceleration, not as causal evidence that technology use produces faster academic progress.These findings contribute to the emerging literature on AI-mediated learning and digital communication by demonstrating how technology-supported learning experiences are associated with students’ perceptions of academic progress through psychological mechanisms.The study clarifies the conceptual boundary of perceived academic acceleration and highlights the need for longitudinal and behavioral validation in future research.
This study examines the influence of learning experiences in the Intelligent Systems concentration on Computer Engineering students’ work readiness in Artificial Intelligence (AI), with self-efficacy as a mediating variable. A quantitative approach with an explanatory research design was used. The participants were 40 Computer Engineering students from the 2022 cohort in the Intelligent Systems concentration, selected through purposive sampling. Data were collected using a four-point Likert-scale questionnaire measuring learning experience, self-efficacy, and work readiness. The instruments were tested for validity and reliability, and the data were analyzed using descriptive statistics and bootstrap-based mediation analysis with PROCESS Macro Model 4 in IBM SPSS Statistics 26. The results showed that learning experience had a positive and significant effect on self-efficacy, self-efficacy had a positive and significant effect on work readiness, and learning experience had a significant direct effect on work readiness. The indirect effect was also significant, with a coefficient of 0.8254 and a 95% bootstrap confidence interval of 0.4518 to 1.2224. These findings indicate that self-efficacy partially mediates the relationship between learning experience and work readiness. Meaningful AI-related learning experiences therefore strengthen both students’ competencies and confidence in preparing for AI-related careers.
Nur Annafiah, Mustari S. Lamada, Dwi Rezky Anandari Sulaiman· Jurnal MediaTIK· 0 citations
Learning Analytics is a key strategy to enhance the quality of education by converting the data generated by students into valuable information for educational decision making. Schools and colleges are increasingly using digital learning places that present huge amounts of data on attendance, assessment performance, interactions with the learning management system, participation and engagement. These data give the opportunity to identify learning patterns and predict academic outcomes prior to significant learning problems. The purpose of the present study is to investigate the relationship between learning behaviors and success of learners and to clarify the role of predictive models in improving learners' success. The quantitative research design was adopted and primary data were obtained from the respondents (i.e. undergraduate and postgraduate students) using a structured questionnaire and with the aid of institutional academic records. Some statistical methods such as descriptive analysis, correlation, regression, and predictive modelling were used to see the effect of learning engagement on academic achievement, digital participation on academic achievement, learning habits on academic achievement, and assessment performance on academic achievement. The results suggest that frequent engagement in online learning activities, frequent completion of assessment, and timely feedback significantly enhance the prediction of student performance. Predictive models were identified as effective predictors of student risk and allowed for personalized learning interventions, approaches and academic support services to be provided to students at risk. The study also identifies issues related to data quality, data privacy, ethical issues, and readiness of institutions for the implementation of learning analytics. The study highlights the potential of learning analytics to inform evidence-based learning and the need to make learning analytics a part of institutional planning and student support. The results hold much promise for education stakeholders who wish to increase student retention, student success, and learning results based on data-informed decisions regarding educational technology.
Dr. M. Vijayakumar, Dr. V. Nalini· Scriptora International Jour...· 0 citations