A networked analysis model of the relationship between metacognitive ability and online learning engagement

The worldwide expansion of online learning has highlighted its persistently low completion rates; insufficient learning engagement is regarded as a central driver. Metacognitive ability-supporting self-regulation and strategic optimization in digital settings-has been theoretically linked to engagement, yet the fine-grained architecture of this association remains unclear. Employing network analysis, we mapped the joint structure of metacognitive and engagement items among 441 university students (emerging adults). Overall, 44.44% of the participants reported high online-learning engagement and 34.47% reported high metacognitive ability, both defined as an average item score of at least 3.5 on the 5-point response scale. Women displayed a higher prevalence of elevated metacognitive ability than men (40.08% vs. 26.26%, p < .01), whereas the two groups did not differ in the prevalence of high online-learning engagement (44.27% vs. 44.69%, p = .93); no significant differences emerged across grade levels (all p > .05). Within the estimated regularized partial-correlation network, OIN3 ("I can throw myself wholeheartedly into online learning") emerged as the super-hub (node strength = 1.82; betweenness = 0.23; standardized EI = 1.31). The strongest edge linked OIN14 ("I take focused notes") with MCOG3 ("I flexibly change learning methods") (regularized partial correlation r = 0.179, p < .001). Bridge-expected-influence analysis identified OIN10 ("I regain confidence when I meet setbacks") and OIN16 ("I try every possible way to overcome difficulties") as the principal bridges (bEI = 0.31 and 0.29, respectively) connecting the engagement and metacognitive clusters. The findings reveal an "affective engagement first, resilience bridges" mechanism and provide an evidence-based, targetable pathway for improving post-pandemic online instruction.