Student Engagement as a Bridge between Support and Learner Satisfaction in Blended Learning Environments at Higher Education

<p dir="ltr">INTRODUCTION Blended learning integrates online and face-to-face activities, replacing some in-person teaching with meaning online interactions (Dziuban et al., 2016). It enhances efficiency, teaching practices, and outcomes but demands student independence and active engagement (Broadbent & Poon, 2015). Effective design should support autonomy, competence, relatedness, and engagement (Chiu, 2021). While its potential to foster engagement is recognized (Halverson et al., 2012), research on digital support in higher education remains limited (Halverson & Graham, 2019). This study explores a model where teacher and digital support predict engagement (i.e., behavioural, cognitive, emotional, agentic), which subsequently predicts learner satisfaction across interaction dimensions (Gao, 2020). </p><p dir="ltr">METHODS The study involved 674 first- and second-year undergraduate students (mean age = 22.83, SD = 3.53) from Singapore-based university, sampled from nine modules across the Engineering and Health and Social Sciences clusters. The participants completed a cross-sectional survey at the end of the trimester in November 2023. The survey included demographic information and validated scales measuring task value, perceived difficulty, workload, teacher/digital support, student engagement, and learner satisfaction. Data analyses included descriptive statistics, multi-group invariance tests, confirmatory factor analysis (CFA), and structural equation modelling (SEM) using SPSS 27.0 (IBM Corporation, 2020) and Mplus 8.0 (Muthen & Muthen, 1998-2017). </p><p dir="ltr">RESULTS & DISCUSSION All scales showed acceptable to excellent internal reliability (a = .79-.90). The measurement model showed good fit indices: c2 = 2607.379, df = 1310, p < .001, CFI = .934, TLI = .925, RMSEA = .038, SRMR = .040, with gender invariance confirmed. SEM analyses accounted for shared variance, yielding a good fit after removing three latent variables: c2 = 1986.321, df = 995, p < .001, CFI = .942, TLI = .932, RMSEA = .038, SRMR = .041. Results revealed that teacher autonomy and digital relatedness support predicted agentic engagement, while digital competence and relatedness support predicted emotional engagement. Emotional and agentic engagement predicted all learner satisfaction facets except for learner-instructor and learner-technology interaction, respectively. Bootstrapping analysis showed that emotional and agentic engagement mediated the relationships between various dimensions of support and learner satisfaction. </p><p dir="ltr">IMPLICATIONS & CONCLUSION This study highlights that simply having appropriate teacher and digital support is insufficient in promoting learner satisfaction. It is how instructors create emotionally engaging digital environments and foster student agency that matters (Kok et al., 2025). Instructors can enhance emotional engagement by building a sense of community through regular interaction, collaboration, real-world connections, and timely feedback (e.g., Chiu, 2023; Kok, 2024). Supporting agency involves enabling students to set goals, plan, monitor, and assess their learning (e.g., Broadbent & Poon, 2015). Future research should examine the long-term effects of engagement on outcomes like academic performance and self-regulated learning. Additionally, developing and testing interventions to enhance emotional and agentic engagement could further improve blended learning experiences.</p>