AI evidence extraction

The impact of teachers’ innovative teaching behaviors on vocational college students’ learning engagement in the Chinese context: the mediating roles of AI trust and learning satisfaction

Tang Zhiwen, Ying Lin, Che Jing Shang · Frontiers in Psychology · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

6/10

Relevance

3/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.3389/fpsyg.2026.1646199

Methodology & findings

Study design

Cross-sectional survey study with structural equation modeling (SEM).

Sample

N = 513, 7 groups

Primary method

Confirmatory Factor Analysis (CFA) in AMOS to examine measurement properties. Ordinary least squares (OLS) regression-based mediation analysis using PROCESS (Model 4, Hayes) with bias-corrected bootstrap confidence intervals (5,000 resamples). Composite reliability (CR) and average variance extracted (AVE) indices computed for convergent and discriminant validity. Heterotrait-monotrait (HTMT) ratio assessed for discriminant validity. Cronbach's alpha and McDonald's omega (ω) reported for internal consistency.

Main result

The study found that "ITB significantly and positively predicted LE" and that "Both AIT and LS acted as partial mediators in this relationship, suggesting that enhancing students' satisfaction yields slightly stronger gains in engagement than focusing on AI trust alone." The research revealed that "teachers' innovative practices enhance engagement both directly and indirectly by fostering learning satisfaction and cultivating calibrated, critical trust in AI—that is, confidence in AI's usefulness combined with awareness of its limitations."

Reports effect sizes.

Research paradigm

Positivist/empiricist quantitative approach with structural equation modeling

Author conclusions

The authors conclude: "This study is among the first to model AI trust and learning satisfaction as parallel mediators between innovative teaching behaviors and learning engagement, thereby enriching our understanding of how pedagogical innovation and students' psychological perceptions jointly shape engagement in the AI era." They further state: "These results imply that higher education should not only encourage innovative, AI-supported teaching designs but also deliberately nurture students' emotional experiences (satisfaction) and their thoughtful, critical trust in AI tools to maximize learning engagement." The authors emphasize that "teachers' innovative practices enhance engagement both directly and indirectly by fostering learning satisfaction and cultivating calibrated, critical trust in AI—that is, confidence in AI's usefulness combined with awareness of its limitations."

Risk of bias

Selection bias: convenience sampling from vocational institutions in Guangdong Province, may not be representative of all vocational students in China; Self-report bias: all measures based on student self-reports of perceptions and behaviors; Reverse causality: cross-sectional design precludes causal inference; students already engaged may rate teachers' behaviors differently; Common method variance: all data collected via single online questionnaire using Likert scales; Social desirability bias: students may over-report engagement and satisfaction; Selection bias: Convenience sampling rather than random assignment; Geographic/contextual bias: Single province (Guangdong) with advanced digital infrastructure may not represent broader Chinese vocational education; Cross-sectional design precludes causal inference; Self-report measurement: All data collected via questionnaire, vulnerable to social desirability bias; Non-independent submissions identified and removed, but data quality screening may have introduced selection bias; Selection bias: convenience sampling from vocational institutions in Guangdong Province limits geographic generalizability; Common method variance: all measures are self-reported by students; Cross-sectional design prevents causal inference; Regional specificity: Guangdong has advanced digital infrastructure and early AI adoption, which may not be representative of other regions; Potential social desirability bias in student ratings of teacher behaviors

Limitations

The authors acknowledge that "this study examined a cross-sectional design, which precludes causal inference
While we modeled ITB as predictors of AIT, LS, and LE based on theoretical reasoning, the direction of causality cannot be definitively established from cross-sectional data alone
It is theoretically plausible that students who are already highly engaged might perceive their teachers' methods as more innovative, or that satisfied students might rate their trust in AI and their teachers' behaviors more favorably (reverse causality)." Additionally, the study notes limitations regarding self-reported measures and the specific context of Guangdong Province, suggesting findings may not generalize to other regions or educational systems with different levels of AI adoption.

Open questions raised

The authors identify the need for: (1) experimental or quasi-experimental designs to more rigorously identify causal processes; (2) longitudinal studies to examine temporal dynamics; (3) broader geographic sampling beyond Guangdong Province; (4) investigation of how institutional governance, enforcement credibility, and peer climates condition AI trust formation; (5) systematic evaluation of how AI-enhanced teaching affects students' trust, satisfaction, and engagement over time.
The authors identify the need for longitudinal and experimental designs to establish causal pathways; broader geographic sampling beyond Guangdong; examination of how institutional governance, enforcement credibility, and peer climates condition AI trust formation; and systematic evaluation of AI-enhanced teaching effects over time.
The authors identify the following gaps and future directions: (1) the need for longitudinal or experimental designs to establish causality rather than associations; (2) expansion of research beyond Guangdong Province to other geographic regions and institutional contexts; (3) investigation of how peer climates, governance clarity, and enforcement credibility moderate the relationship between innovative teaching and AI trust; (4) examination of how AI trust and learning satisfaction may influence each other over time in longitudinal frameworks; (5) systematic evaluation of how AI-enhanced teaching affects students' trust, satisfaction, and engagement over time.

Data: The authors state: 'The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.' No public dataset repository or URL provided.; The authors state: "The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author." However, no publicly accessible dataset URL provided.; Data availability statement indicates: 'The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.' No public dataset repository is explicitly named.Code: None mentioned.; None mentioned in the paper.; Not mentionedExtracted from: pdfAgreement 43%

Explore related topics

AI Literacy Training AI Governance Survey