Exploring the influence of teachers' motivating styles on college students' agentic engagement in online learning: The mediating and suppressing effects of self-regulated learning ability

The study addresses how college students can sustain high agentic engagement in online learning, a key determinant of learning effectiveness that emphasizes students’ proactive, constructive contributions to their learning. In online contexts, temporal/spatial separation, limited interaction, and technical issues can dampen engagement. Drawing on self-determination theory (SDT), the authors focus on teachers’ motivating styles—autonomy-supportive versus controlling—as salient external factors that may shape engagement. They also consider students’ self-regulated learning (SRL) ability as a crucial internal factor for navigating the freedoms and challenges of online learning. The study aims to clarify relationships among teachers’ motivating styles, students’ SRL ability, and agentic engagement in online settings, and to determine whether SRL mediates these relationships. Seven research questions (Q1–Q7) probe associations and mechanisms linking autonomy support and control to agentic engagement, the links between motivating styles and SRL, the link between SRL and agentic engagement, and whether SRL mediates the effects of both motivating styles on agentic engagement.

Theoretical background centers on SDT and its distinction between autonomy-supportive and controlling motivating styles. Autonomy-supportive teaching acknowledges students’ perspectives, provides rationale and choice, and fosters internal motivation; controlling style uses pressure and coercion, directing students’ thoughts/behaviors. Prior work generally links autonomy support to better engagement, motivation, and well-being, while control is often detrimental, though cultural and contextual nuances exist. Agentic engagement is framed as a fourth dimension of engagement that captures students’ proactive contributions to instruction and environment shaping, with demonstrated links to achievement. SRL is defined as learners’ proactive regulation of goals, strategies, monitoring, and evaluation; it is vital in both online and offline contexts and can be influenced by teacher practices and the learning environment. Prior evidence suggests autonomy support relates positively to SRL and engagement, and that SRL often predicts engagement and achievement and may mediate links from teaching practices to engagement.

Design and participants: A cross-sectional survey targeted college students in China with online learning experience. Using random sampling via the WJX online platform, 702 students responded; 681 valid responses were retained. The sample comprised 611 females (89.7%) and 70 males (10.3%); 206 first-year (30.3%), 225 second-year (33.0%), and 250 third-year (36.7%) students. Ethical approval was obtained from South China Normal University; participation was anonymous and voluntary. Instruments: A questionnaire included demographics and four 5-point Likert scales (1=strongly disagree to 5=strongly agree): (a) Teachers’ motivating styles adapted from Lauermann & Berger (2021), with autonomy-supportive style (ASS; 8 items) and controlling style (CS; 5 items). Example items: “My teachers encourage me to find solutions to problems on my own in online learning”; “During the online learning process, my teachers almost never consider my opinions.” (b) Agentic engagement (AE) adapted from Reeve (2013) to the online context (9 items; e.g., “I try to make what I learn interesting.”). (c) Online SRL adapted from Barnard et al. (2009) covering six subdimensions: goal setting, environment structuring, strategy use, time management, help-seeking, and self-evaluation (e.g., “I’m not going to lower the quality of my learning just because it’s online.”). Translation and pilot: Scales were translated into Chinese and adapted by experts. A pilot with five teachers and ten students verified clarity and contextual appropriateness. After refinement, data collection proceeded. Reliability and validity: Data were split randomly for EFA and CFA. Internal consistency and composite reliability were high: Cronbach’s alpha and CR were all >0.90; AVE and factor loadings exceeded 0.50, indicating good convergent validity. Model fit indices for CFA met thresholds (e.g., χ2/df=5.668, GFI=0.914, AGFI=0.883, NFI=0.959, CFI=0.966, RMSEA=0.076). Harman’s single-factor test indicated common method bias was minimal (largest single factor=26.471% <40%). Analysis: SPSS 21.0 was used for EFA, descriptive statistics, and Pearson correlations. AMOS 24.0 conducted CFA. Mediation was tested using PROCESS (Model 4) with bootstrap confidence intervals; mediation deemed significant when 95% CI excluded zero.

Descriptive statistics: Means (M) and standard deviations (SD): ASS M=3.906, SD=0.743; CS M=2.328, SD=1.104; AE M=3.739, SD=0.785; SRL M=3.888, SD=0.679. Correlations: ASS and AE r=0.777 (p<0.01); SRL and AE r=0.832 (p<0.01); ASS and SRL r=0.785 (p<0.01); CS and SRL r=-0.109 (p<0.01); CS and AE r=-0.003 (ns). Mediation with ASS (predictor) → SRL (mediator) → AE (outcome): - Total effect of ASS on AE: positive (report indicates substantial effect). - Direct effect (without mediator): β=0.342, t=10.101, p<0.001. - ASS → SRL: β=0.705, t=32.100, p<0.001. - SRL → AE: β=0.660, t=17.692, p<0.001. - Indirect effect via SRL: significant (β≈0.465; 95% CI [0.360, 0.577]). - Conclusion: SRL partially mediates the positive effect of ASS on AE. Mediation with CS (predictor) → SRL (mediator) → AE (outcome): - Direct effect of CS on AE (without mediator): β=0.063, t=4.199, p<0.001 (positive). - CS → SRL: β=-0.067, t=-2.857, p<0.001 (negative). - SRL → AE: β=0.973, t=39.755, p<0.001 (positive). - Indirect effect via SRL: β=-0.065, 95% CI [-0.121, -0.014] (significant, negative). - Total effect: β=-0.002, 95% CI [-0.056, 0.052] (ns), indicating a suppression effect whereby the negative indirect path offsets the positive direct path. Overall: Both autonomy-supportive and controlling styles show positive direct relations with agentic engagement in online contexts, but controlling style undermines SRL, which in turn suppresses its overall effect on agentic engagement. ASS enhances SRL, which boosts AE.

Findings answer the seven research questions: (Q1) Autonomy-supportive style positively relates to and enhances agentic engagement; (Q2) Controlling style also shows a positive direct association with agentic engagement in this online, Chinese higher-education context; (Q3) Autonomy support positively predicts SRL; (Q4) Controlling style negatively predicts SRL; (Q5) SRL positively predicts agentic engagement; (Q6) SRL partially mediates the ASS→AE link; (Q7) SRL exerts a suppressing (negative indirect) effect on the CS→AE link, offsetting its positive direct effect. The unexpected positive direct effect of controlling style on agentic engagement may reflect cultural/contextual dynamics (e.g., classroom collectivism, normative use of control for managing large classes, and the logistical challenges of online class management). Nevertheless, because control weakens SRL—an essential driver of agentic engagement—the net influence is suppressed. Practically, teachers should flexibly calibrate motivating styles: employ autonomy support to foster SRL and deeper agentic engagement, while applying control judiciously to manage online classes without undermining students’ self-regulation.

This study clarifies how teachers’ motivating styles and students’ SRL ability jointly shape agentic engagement in online learning. Autonomy-supportive teaching enhances both SRL and agentic engagement and exerts a positive indirect effect on engagement via SRL (partial mediation). Controlling style shows a positive direct relation with engagement but concurrently lowers SRL, producing a suppressing mediation that neutralizes its total effect. The contribution lies in demonstrating SRL’s dual role as mediator and suppressor in full-time online learning contexts and in highlighting cultural/contextual nuances in motivating styles’ effects. Implications include encouraging educators to strengthen autonomy-supportive practices and explicitly develop students’ online SRL skills to bolster agentic engagement. Future research should broaden sampling across institutions, consider additional engagement dimensions (behavioral, emotional, cognitive), and use mixed methods (e.g., interviews, qualitative analyses) to triangulate and deepen insights.

The sample was drawn from a single university, limiting generalizability to the broader population of Chinese college students. Only four variables were examined, and data relied solely on self-report questionnaires, which may introduce bias. The cross-sectional design precludes causal inference.