E-learning service quality and student satisfaction in public universities: a quantitative hierarchisation of determinants through systematic review

Introduction This systematic review provides comprehensive hierarchisation of e-learning service quality determinants predicting student satisfaction in public universities worldwide. Public universities face unprecedented challenges delivering quality e-learning while managing severe resource constraints and equity mandates. Unlike previous reviews aggregating findings across institutional types through qualitative synthesis, this study provides effect size stratification enabling evidence-based resource allocation for public sector contexts systematically underrepresented in literature. Methods A systematic bibliographic search of Scopus and Web of Science (June 2025) identified 15 empirical studies (2016–2025) examining e-learning quality-satisfaction relationships in public universities. Studies were selected through purposive sampling based on predefined eligibility criteria: public university contexts, quantitative designs reporting standardised effect sizes ( β coefficients), and peer-reviewed publications (2016–2025) in English/Spanish. Following PRISMA 2020 guidelines, studies underwent Mixed Methods Appraisal Tool (MMAT) 2018 quality assessment (mean score: 3.67/5). Standardised regression coefficients ( β ) were extracted, tabulated, and stratified by effect magnitude. Synthesis employed Synthesis Without Meta-analysis (SWiM) guidelines given methodological heterogeneity precluding meta-analysis. Results Fifteen studies encompassed 9,847 students across 12 countries. Hierarchisation revealed perceived value as dominant predictor ( β = 0.626, R 2 = 0.568), surpassing instructor competence by 5.35×. Four-tier classification emerged: Critical ( β 0.40), High ( β = 0.35–0.44), Moderate ( β = 0.20–0.34), Low ( β 0.20). Geographic heterogeneity was extreme (online interaction β = 0.197–0.768, coefficient of variation CV = 74.3%). Institutional quality dimensions remained systematically absent (80.0% omission). Methodological limitations included absolute cross-sectional dominance (100%) and socioeconomic status omission (0%). Discussion Findings reveal value-centric paradigm challenging technology-centric frameworks, wherein students prioritise outcome-oriented dimensions employability, return on investment (ROI) over process-oriented attributes. System/information quality function as hygiene factors—absence guarantees dissatisfaction, presence insufficient beyond functional thresholds. Strategic implications: prioritise value articulation over technological sophistication, achieve technological parity rather than pursuing “arms races”, invest in compensatory offline infrastructure, and contextualise frameworks through local validation. Future research should pursue longitudinal designs, geographic diversification, and equity-oriented conceptualisations incorporating socioeconomic moderators.