AI-driven Virtual Field Trips for Agricultural Education Evaluation

Thesis Abstract

The study addresses the pressing need to assess agricultural education effectiveness in environments constrained by physical field trip opportunities, rising costs, and limited access to diverse agro-ecologies by leveraging AI-driven virtual field trips (VFTs) to evaluate learning outcomes, engagement, and practical competencies. The aim is to determine how AI-enabled VFTs influence student understanding of agronomic concepts, decision-making skills, and attitudes toward sustainable farming, compared with traditional field-based experiences. Specific objectives include (1) evaluating learning gains in domain knowledge (taxonomy, physiology, and crop management) using pre-post tests, (2) examining changes in practical competencies and procedural understanding through performance rubrics, (3) exploring student engagement, motivation, and perceived credibility via validated scales, (4) identifying the moderating effects of prior ICT literacy and farm exposure on learning gains, and (5) developing a parsimonious model linking AI-driven VFT exposure to educational outcomes grounded in constructivist and situated learning theories. A mixed-methods research design will be employed, combining quasi-experimental and interpretive approaches. The population comprises secondary and tertiary-level agricultural education learners enrolled in eight institutions across two regions with comparable curricula. A total sample of 320 students will be recruited, with 160 assigned to the AI-driven VFT intervention and 160 to conventional field trips or standard classroom simulations. Data collection instruments include (i) a validated knowledge assessment comprising 40 multiple-choice and short-answer items aligned to agronomy and agroecology outcomes, (ii) a practical skills rubric assessing crop scouting, soil assessment, and irrigation planning applied to simulated scenarios, (iii) a Likert-scale engagement and motivation questionnaire adapted from the User Engagement in Learning Environments scale, (iv) a demographic and ICT-literacy survey, and (v) semi-structured interview guides for a purposive subsample of 40 participants to capture in-depth experiences. AI techniques will underpin the VFT system, employing computer vision to track interaction patterns, natural language processing to analyze discourse during tasks, and adaptive analytics to tailor scenario difficulty in real time. Validity and reliability of instruments will be established through content validity indices (CVI > 0.78), pilot testing (n=40), and Cronbach’s alpha (>0.80 for scales). Data will be analyzed using a combination of inferential and qualitative methods (i) ANCOVA to compare post-test outcomes while controlling for baseline scores and ICT-literacy, (ii) multivariate regression to model learning gains as a function of exposure duration, engagement, and prior experience, (iii) thematic analysis of interview transcripts to identify perceived affordances, barriers, and contextual factors, and (iv) machine-learning derived feature importance to reveal which interaction metrics most strongly predict outcomes. A theoretical framework grounded in constructivist learning theory (Piaget) and situated learning (Lave and Wenger) will guide interpretation, with AI-driven personalization incorporated as a mediating factor. Expected findings include statistically significant improvements in topic knowledge and practical competencies for the VFT group, higher engagement and motivation scores, and a positive shift in attitudes toward ICT-enabled agricultural education. It is anticipated that greater exposure to AI-adaptive scenarios will correlate with enhanced diagnostic reasoning and situational problem-solving abilities, particularly among learners with moderate prior ICT literacy. The study’s contribution to knowledge lies in providing robust empirical evidence on the effectiveness of AI-driven VFTs as substitutes or complements to traditional field experiences, identifying key interaction metrics that predict learning gains, and detailing a scalable implementation framework for integrating AI-enhanced VFTs into agricultural curricula. The main conclusion is that AI-driven VFTs can substantially augment agricultural education by delivering diverse, authentic, and scalable field experiences that foster conceptual understanding and practical skills, even when physical field trips are restricted. Recommendations include investing in AI-driven VFT development with emphasis on ecological validity, expanding access to reliable hardware and high-speed connectivity, providing professional development for educators on interpreting AI-derived feedback, and conducting longitudinal studies to assess retention and transfer of learning to real-world farming contexts.