Learning Factors in AI-Augmented Education: A Comparative Study of Middle and High School Students

Abstract: The increasing integration of AI tools in education has led prior research to explore their impact on learning processes. Nevertheless, most existing studies focus on higher education and conventional instructional contexts, leaving open questions about how key learning factors are related in AI-mediated learning environments and how these relationships may vary across different age groups. Addressing these gaps, our work investigates whether four critical learning factors, experience, clarity, comfort, and motivation, maintain coherent interrelationships in AI-augmented educational settings, and how the structure of these relationships differs between middle and high school students. The study was conducted in authentic classroom contexts where students interacted with AI tools as part of programming learning activities to collect data on the four learning factors and students' perceptions. Using a multimethod quantitative analysis, which combined correlation analysis and text mining, we revealed markedly different dimensional structures between the two age groups. Middle school students exhibit strong positive correlations across all dimensions, indicating holistic evaluation patterns whereby positive perceptions in one dimension generalise to others. In contrast, high school students show weak or near-zero correlations between key dimensions, suggesting a more differentiated evaluation process in which dimensions are assessed independently. These findings reveal that perception dimensions actively mediate AI-augmented learning and that the developmental stage moderates their interdependencies. This work establishes a foundation for the development of AI integration strategies that respond to learners' developmental levels and account for age-specific dimensional structures in student-AI interactions.

• The paper demonstrates that learning factors (Experience, Clarity, Comfort, Motivation) interact holistically in middle school and modularly in high school.
• It employs quantitative correlations, non-parametric tests, and text-mining methods on data from 53 students to validate its findings.
• The research suggests tailored interventions with scaffolding for younger learners and autonomous tasks for older students in AI-augmented education.

Comparative Analysis of Learning Factor Interrelationships in AI-Augmented K-12 Computer Science Education

Introduction

The study "Learning Factors in AI-Augmented Education: A Comparative Study of Middle and High School Students" (2512.21246) conducts a rigorous examination of how key perceptual dimensions—Experience, Clarity, Comfort, and Motivation—interrelate in AI-mediated learning environments, and how these relationships are modulated by developmental stage. Employing a cross-sectional quantitative and text-mining methodology in authentic computer science classroom contexts, the research establishes evidence for distinct dimensional interaction patterns between middle and high school cohorts. The findings have salient implications for the design and deployment of AI educational interventions.

Methodological Overview

A total of 53 students (25 middle school, 28 high school) enrolled in extracurricular computer science programs participated. Both groups engaged with AI tools—Teachable Machine for middle schoolers and advanced LLM platforms (Gemini, Claude) for high schoolers—integrated into Python programming instruction. Data were gathered via a validated 17-item instrument (eight Likert-scale and nine open-ended items) administered through Wooclap, reliably capturing the targeted dimensions ($\alpha_{total}=0.85$).

Quantitative data were subjected to correlation matrix analysis and non-parametric group comparisons (Mann-Whitney U, Bonferroni-corrected), while qualitative responses underwent word frequency and semantic co-occurrence network analysis to triangulate attitudinal trends.

Dimensional Structures: Holistic vs Modular Perceptual Patterns

Statistical analyses revealed generally positive attitudes towards AI tools across both developmental stages. However, only perceived ease of use (Q3: $U=192$, $p=0.0013$) was significantly higher in high school students. Middle schoolers displayed more dispersed ratings, indicating less established evaluative criteria or digital literacy heterogeneity.

The correlation matrices (Figure 1) highlighted profound differences in the coherence of perception:

Figure 1: Correlation matrices between questionnaire items by school level, revealing markedly different patterns of association between middle and high school students.

For middle school students, strong positive correlations exist across all surveyed dimensions (e.g., Experience-Clarity $\rho=0.79$), confirming a holistic, mutually reinforcing evaluation structure. A positive perception in one factor typically elevated all others, suggesting that for younger students, positive experiences are global rather than dimension-specific.

Conversely, high school students exhibited attenuated or null correlations between dimensions not directly related (e.g., Experience-Motivation $\rho=0.02$), signifying a modular, discriminative appraisal mode characteristic of greater developmental maturity. Notably, the only strong correlation in this group was within the Clarity domain (Q5-Q9, $\rho=0.61$), indicating that clear AI explanations directly affected only subdimensions of understanding, not the broader experiential or motivational domains.

Dimension-level matrix aggregation visualized these age-dependent structural decouplings (Figure 2):

Figure 2: Correlations between aggregated dimensions (Experience, Clarity, Comfort, Motivation) for each school level, showing substantial differences in evaluation coherence between age groups.

These results underscore that developmental stage modulates not just mean perception scores but the fundamental structure of how perceptions are formed and interrelated during AI-mediated learning.

Semantic Analysis of Perceptual Focus

Text mining further substantiated the quantitative findings. WordCloud and co-occurrence network analysis of responses regarding the perceived impact of AI tools (Q7) distinguished clear semantic differences:

Figure 3: WordCloud and co-occurrence network for Q7 responses, showing student perceptions of the AI tool's impact on learning for middle and high school. The WordCloud highlights the most frequent terms, while the co-occurrence network reveals semantic relationships between key concepts.

Middle school students emphasized basic comprehension and generalized aid ("learn," "understand," "helped"), with co-occurrence networks centered around broad knowledge acquisition. High school students, however, foregrounded analytic and technical impacts ("code," "explanation," "think") and described the AI's instructional value in more granular terms. Semantic centrality for the term "understand" was preserved across both groups but linked to more diverse and higher-order concepts for older students.

Implications and Prospective Directions

This research provides robust empirical support for the assertion that Experience, Clarity, Comfort, and Motivation constitute intertwined yet dynamically structured learning factors in AI-augmented secondary education. The findings directly contradict the assumption of developmental invariance in perception dimensionality: middle school learners benefit from design strategies that amplify clarity and comfort to holistically uplift overall experience, whereas high school populations require differentiated interventions that target motivational salience and autonomous critical engagement.

Practically, the results suggest curriculum developers should implement structured, scaffolded AI approaches for younger students and increasingly autonomous, technically challenging tasks for older students. The data also highlight the necessity of explicit prompt engineering education, as observed difficulties in leveraging AI tool functionality were prevalent at all stages.

Theoretically, these findings prompt new research into how developmental cognitive structures interact with evolving student–AI interaction paradigms, supporting further investigation into personalized learning systems that adapt to age-linked perceptual architectures. The integration of longitudinal and behavioral log analysis is posed as essential for future inquiry, to map perception–interaction–outcome trajectories over time.

Conclusion

This comparative study delineates that the interrelationships among core perceptual learning factors in AI-mediated environments are fundamentally modulated by developmental stage. Middle school students exhibit a holistic, globally positive perception coupling, while high school students form nuanced, dimensionally differentiated assessments. These findings inform the design of age-adaptive AI educational systems and point towards the necessity of both prompt engineering literacy and developmentally responsive curricular scaffolding. Future research integrating longitudinal perceptual data and system usage traces is essential to optimize AI’s pedagogical efficacy across the K-12 continuum.