📄 social_science

An Integrative Model of Artificial Intelligence Competencies and a Multi-theoretical Approach to the Implementation of Inclusive STEM Learning

This quantitative study of 1,000 Indonesian inclusive school teachers demonstrates that both AI competencies and a multi-theoretical pedagogical approach significantly and synergistically enhance the implementation of inclusive STEM learning, validating an integrative model for teacher readiness and policy development.

Original authors: I Wayan Sumandya, I Wayan Widana, A.A. Istri Mirah Dharmadewi, Gde Iwan Setiawan

Published 2026-06-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: I Wayan Sumandya, I Wayan Widana, A.A. Istri Mirah Dharmadewi, Gde Iwan Setiawan

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to build a house that is welcoming to everyone, regardless of their size, shape, or abilities. This is what Inclusive STEM Learning is like: it's about teaching Science, Technology, Engineering, and Math in a way that every single student can participate and succeed.

This paper is essentially a blueprint for how teachers can build that house successfully. The authors, a team from Universitas PGRI Mahadewa Indonesia, argue that you can't just hand a teacher a pile of bricks (technology) and expect a great house. They need two specific things to work together: AI Skills and Educational Theory.

Here is the breakdown of their findings using simple analogies:

1. The Two Essential Ingredients

The researchers studied 1,000 teachers in Indonesia to see what makes them successful at this task. They found two main "ingredients" are needed:

AI Competency (The Toolbox): This isn't just about knowing how to turn a computer on. It's like a teacher having a high-tech, magical toolbox. They need to know how to use AI to create lessons that adapt to different students, understand how the AI "thinks," and know the ethical rules (like not letting the AI be unfair).
Multi-Theoretical Approach (The Architect's Plan): This is the teacher's knowledge of how people learn. Imagine a master architect who knows that some people learn by doing, some by watching, and some by talking. A "multi-theoretical" teacher doesn't rely on just one rulebook; they mix different learning theories (like how kids build knowledge themselves, how they learn from each other, and how to make everyone feel safe) to design the perfect lesson.

2. The Magic of Mixing Them (The Synergy)

The most exciting part of the study is what happens when you combine these two.

Think of AI as a powerful engine in a car, and Learning Theory as the steering wheel and map.

If you have a powerful engine (AI) but no steering wheel (Theory), the car goes very fast but crashes or goes in circles. The teacher might use cool tech but miss the students' needs.
If you have a great map (Theory) but a weak engine (No AI), you know exactly where to go, but you move too slowly to get there efficiently.
The Study's Finding: When you have both a powerful engine and a skilled driver with a map, the car flies. The researchers found that when teachers had both strong AI skills and a deep understanding of learning theories, their ability to teach inclusively was significantly stronger than if they had just one or the other. They called this a "synergistic effect."

3. What the Numbers Say

The researchers used a sophisticated statistical method (like a very precise calculator) to measure this.

AI Skills alone helped teachers be about 40% more effective.
Learning Theories alone helped teachers be about 32% more effective.
But when combined, they boosted each other, making the overall success even higher.

The study explains about 43% of why some teachers are great at inclusive STEM and others aren't. The rest depends on other factors not measured in this specific study (like school funding or student home life).

4. The Bottom Line

The paper concludes that to make STEM learning fair and accessible for everyone (including students with special needs), we cannot just train teachers to be "techies." We also cannot just train them to be "theorists."

We need hybrid teachers: those who are comfortable with the latest AI tools and who deeply understand the human side of learning. When these two worlds collide, they create a learning environment that is not only smart but also kind, fair, and adaptable to every single student.

In short: You need the right tools (AI) and the right plan (Theory) to build a classroom where everyone belongs.

Technical Summary: An Integrative Model of Artificial Intelligence Competencies and a Multi-theoretical Approach to the Implementation of Inclusive STEM Learning

Problem Statement
The integration of Artificial Intelligence (AI) into Science, Technology, Engineering, and Mathematics (STEM) education within inclusive schools faces a critical gap: teacher readiness is often limited to technological proficiency, lacking a comprehensive pedagogical theoretical framework. While AI offers potential for personalization and accessibility, its implementation in inclusive contexts risks creating disparities if educators view it merely as a technical tool rather than a conceptual instrument for data-driven, ethical, and inclusive pedagogy. Existing literature lacks integrative models that explicitly combine teachers' AI competencies with multi-theoretical approaches (e.g., constructivism, social-cognitive theory, inclusive education theory) as predictors of successful inclusive STEM implementation. Furthermore, methodological limitations in prior studies, such as a reliance on qualitative or simple survey designs without simultaneous causality testing between latent variables, hinder a robust understanding of these structural relationships.

Methodology
This study employed a quantitative explanatory survey design to test an integrative structural model.

Participants: The sample consisted of 1,000 active teachers from inclusive schools in Indonesia (Bali, West Nusa Tenggara, and East Nusa Tenggara), selected via purposive sampling. Criteria included at least two years of teaching experience and prior involvement in AI or STEM programs.
Instrument: Data were collected using a five-point Likert scale questionnaire measuring three primary constructs:
1. AI Competence: Encompassing conceptual understanding, operational capabilities, innovation, collaboration, and confidence.
2. Multi-theoretical Approach: Measuring the conformity of theory and practice, theoretical synthesis, contribution to inclusion, feasibility, and professional development.
3. STEM Implementation Capabilities: Assessing planning, adaptive curriculum, active strategies, AI integration, inclusive evaluation, and resource support.
Analysis: The data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) via SmartPLS software. The analysis followed a two-stage approach:
1. Measurement Model: Evaluated for reliability (Cronbach's Alpha, Composite Reliability) and validity (Convergent Validity via AVE and Factor Loadings; Discriminant Validity via Fornell-Larcker criteria and HTMT).
2. Structural Model: Tested for multicollinearity (Inner VIF), path coefficients (hypothesis testing), effect sizes ( $f^2$ ), and predictive relevance ( $R^2$ , $Q^2$ ). Bootstrapping was used to determine statistical significance.

Key Results
The structural model demonstrated excellent construct validity and reliability, with all factor loadings exceeding 0.70, Composite Reliability values above 0.90, and Average Variance Extracted (AVE) values above 0.65. Discriminant validity was confirmed as the square root of AVE for each construct exceeded the inter-construct correlations.

The hypothesis testing yielded the following significant findings:

AI Competence: Teachers' AI competencies have a significant positive direct effect on the implementation of inclusive STEM learning ( $\beta = 0.408$ , $p < 0.001$ ).
Multi-theoretical Approach: The multi-theoretical approach has a significant positive direct effect on inclusive STEM implementation ( $\beta = 0.320$ , $p < 0.001$ ).
Interaction Effect: There is a significant positive interaction effect between AI competence and the multi-theoretical approach ( $\beta = 0.213$ , $p < 0.001$ ), indicating a synergistic relationship where the combination of both factors enhances implementation effectiveness.
Model Fit: The structural model explains 42.7% of the variance in inclusive STEM implementation ( $R^2 = 0.427$ ), categorized as moderate predictive power. The model showed no multicollinearity issues (Inner VIF < 5) and a Standardized Root Mean Square Residual (SRMR) indicating good model fit.

Key Contributions

Theoretical: The study proposes and validates an "Integrative Model" (referred to in the discussion as the AI-Integrated Multi-theoretical Model or AIMTM) that explicitly links AI competencies and multi-theoretical pedagogical approaches as dual predictors of inclusive STEM success. It extends constructivist and social-cognitive theories into a techno-pedagogical context, emphasizing that AI must be grounded in ethical and humanistic educational theories.
Methodological: By utilizing PLS-SEM, the research provides a robust causal and predictive analysis of latent variables, addressing the methodological gaps of previous exploratory or qualitative studies in this domain.
Practical: The findings offer strategic implications for teacher professional development, suggesting that training must move beyond technical AI skills to include the integration of learning theories, ethical reflection, and collaborative competencies.

Significance and Claims
The authors claim that this research provides strong empirical evidence that the successful implementation of inclusive STEM learning relies on the synergy between technological mastery (AI competence) and epistemological depth (multi-theoretical approach). The study asserts that AI alone is insufficient; it requires a "pedagogical alignment" where teachers act as mediators between algorithms and human values.

The paper concludes that the integration of these dimensions creates an "intelligent learning ecosystem" capable of fostering equitable, adaptive, and sustainable STEM education. The authors position their work as a foundational step for policymakers and curriculum developers to design interventions that prevent the "directionless technologist" phenomenon, ensuring that AI serves as a tool for equity rather than a source of new educational gaps. The study modestly notes that while the model is robust, future research should explore multi-level contextual influences and cross-cultural validations across Southeast Asia.

1. The Two Essential Ingredients

2. The Magic of Mixing Them (The Synergy)

3. What the Numbers Say

4. The Bottom Line

Technical Summary: An Integrative Model of Artificial Intelligence Competencies and a Multi-theoretical Approach to the Implementation of Inclusive STEM Learning

More like this