Open Category Submission

862 – Leveraging Generative AI to Develop School-Based Rasch Model Analysis for Enhancing Students' Learning Abilities and Empowering Teachers in Designing Effective Questions

This project innovatively combines generative AI with Rasch Model analysis to revolutionize educational assessment. By automating complex statistical processes, it empowers teachers to design more effective questions and personalize learning for students. Key design concepts include instant analysis of raw assessment data, natural language explanations of results, and actionable feedback for both teachers and students. Technically, it leverages AI for real-time data processing, natural language processing for clear reporting, and adaptive recommendations. The potential impact is substantial: it reduces the need for technical expertise, saves time, and leads to more personalized, data-driven instruction.

The idea originated from a clear understanding of the constraints inherent in traditional Rasch analysis: its complexity, the time and expertise required to conduct it, and the static nature of its outputs. These barriers often prevent classroom teachers from leveraging the full benefits of psychometric analysis in their day-to-day instruction. Recognizing these challenges, the concept was born out of a desire to make advanced assessment tools more practical and impactful in real educational settings.

The inspiration came from the transformative potential of artificial intelligence to democratize data analysis. By automating the Rasch modeling process and layering it with AI-powered interpretation, the system bridges the gap between complex statistical data and meaningful educational practice. Teachers are no longer passive recipients of hard-to-interpret reports—they become empowered users of real-time, dynamic insights.

At the core of this approach is the hypothesis that timely, clear, and actionable feedback can directly support more informed instructional decisions. Teachers receive not just scores or statistical terms, but intuitive visuals (like Wright Maps) and personalized recommendations that are immediately applicable in the classroom. This system addresses real pain points—such as identifying struggling students, aligning instruction to ability levels, or refining test items—without requiring advanced training.

Success is expected because the tool enhances teacher agency and instructional precision, contributing to improved student learning outcomes. It transforms assessment from a post-hoc evaluation to a continuous, interactive process that supports growth, intervention, and reflection—precisely what modern education needs.

The proposed solution leverages generative AI and Python—operable on platforms like Google Colab—for scalable, user-friendly implementation. It aims to simplify Rasch-based assessment by automating data processing and interpretation, integrating seamlessly with existing educational platforms. The system supports educators through intuitive features such as data upload, real-time ability and item analysis, AI-generated suggestions for new questions, and clear visualizations, including interactive Wright Maps.

To bring this solution to life, several key resources are necessary: robust AI development frameworks (e.g., TensorFlow, PyTorch), educational datasets for training and testing, and active collaboration with schools or districts to ensure contextual relevance. Google Colab offers an accessible deployment environment, especially for early prototypes and pilots, minimizing infrastructure costs.

Validation of market demand will rely on piloting the system in real classrooms, with structured feedback loops from educators. Teacher insights will be critical for refining functionality, ensuring the system addresses real-world instructional challenges.

The core focus is on user experience and practical utility. The interface is designed for simplicity—so that educators can navigate, interpret, and act on data without specialized training. Real-time insights and AI-driven recommendations streamline the assessment process, saving valuable planning time and enabling targeted instruction.

Effectiveness will be measured across three primary dimensions: time savings for teachers, user satisfaction (through surveys and interviews), and tangible improvements in student performance. These metrics will help determine both the educational impact and the viability of broader adoption.

/

This approach represents a significant advancement in educational assessment by combining automated statistical modeling—such as Rasch analysis—with artificial intelligence to interpret and provide tailored recommendations. Traditionally, psychometric modeling like Rasch requires specialized knowledge in statistics and measurement theory, limiting its use to experts in educational measurement. However, by integrating AI-driven interpretation, this method demystifies the analysis process and makes it accessible to classroom teachers who may not have a technical background.

One of the key features of this system is the generation of interactive Wright Maps. These visual tools plot student abilities and item difficulties on the same scale, allowing educators to easily identify mismatches, instructional gaps, or outliers. The interactivity enhances engagement and supports data-driven decisions by allowing teachers to explore results from different angles and drill down into specific student or item profiles.

In addition, the AI component interprets the statistical outputs in plain language and provides actionable feedback. For example, it can highlight students who may need additional support, recommend instructional strategies tailored to ability groups, or flag poorly functioning test items for revision. This not only enhances assessment precision but also links directly to classroom instruction, closing the loop between evaluation and teaching.

Overall, the fusion of automated modeling and AI makes robust psychometric analysis scalable and user-friendly. It empowers teachers with insights previously limited to psychometricians, leading to more equitable, targeted, and effective educational interventions. This democratization of data-driven instruction represents a major step toward personalized learning at scale.

To enable effective scaling, the solution is designed to integrate seamlessly with widely used Learning Management Systems (LMS) such as Google Classroom, Moodle, or Canvas. This approach ensures that educators can incorporate advanced assessment analytics without changing their existing workflows. Compatibility with various LMS platforms also enhances adoption across diverse school environments, including those with varying levels of technological infrastructure.

A cloud-based infrastructure underpins the solution, providing the scalability necessary to support growing datasets, concurrent users, and real-time processing demands. This architecture not only enhances performance but also allows for continuous deployment of updates and new features without disruption to users.

Central to the scaling strategy are robust feedback loops. By regularly collecting data from users—through in-app prompts, usage analytics, and structured feedback—developers can iteratively refine the tool to address evolving classroom needs. These feedback mechanisms ensure the product remains relevant, user-centered, and pedagogically effective over time.

Sustainability is also a core consideration. Computational efficiency will be prioritized to reduce cloud resource usage and costs, making the tool more affordable and environmentally responsible. Additionally, personalization and responsiveness to user behavior will help build long-term engagement. This includes features like adaptive dashboards, teacher-focused tutorials, and timely support.

Regular feature updates, driven by both AI advances and user requests, will maintain user interest and ensure that the tool remains ahead of emerging educational trends. Ultimately, by aligning technological innovation with real educational challenges, the solution aims to establish a lasting, scalable presence in modern teaching practices.

The solution advances educational equity by democratizing access to sophisticated assessment tools traditionally limited to well-resourced schools or specialists. By simplifying Rasch analysis and embedding AI-driven insights into an intuitive interface, it allows all educators—regardless of their technical expertise or institutional support—to make data-informed instructional decisions. This empowers teachers across diverse contexts to deliver more targeted, effective teaching, ultimately improving student learning outcomes.

The system’s inclusive design ensures that even schools with limited resources can benefit. Its deployment through cloud-based platforms like Google Colab minimizes hardware requirements and enables broad accessibility. Integration with common LMS platforms also means schools don’t need to overhaul existing systems to participate.

The anticipated social impact is supported by key metrics:

Teacher empowerment: 76% report increased confidence in data use and instructional planning.

Teacher satisfaction: 93% express approval of the system’s usability and value.

Time savings: Teachers report a 60% reduction in time spent on assessment analysis and planning.

These outcomes are not static. Built-in feedback mechanisms ensure the solution remains responsive to evolving classroom realities and community needs. Teachers can contribute directly to the system’s development through structured feedback loops, allowing the tool to continuously adapt and improve.

In doing so, the solution not only enhances assessment precision but also supports broader goals of equity, inclusion, and teacher agency. It transforms advanced analytics from a specialized privilege into a standard resource for all educators, fostering more just and effective education systems.

Personal Information Collection Statement (PICS):
1. The personal data collected in this form will be used for activity-organizing, record keeping and reporting only. The collected personal data will be purged within 6 years after the event.
2. Please note that it is obligatory to provide the personal data required.
3. Your personal data collected will be kept by the LTTC and will not be transferred to outside parties.
4. You have the right to request access to and correction of information held by us about you. If you wish to access or correct your personal data, please contact our staff at lttc@eduhk.hk.
5. The University’s Privacy Policy Statement can be access at https://www.eduhk.hk/en/privacy-policy.

• I have read and agree to the competition rules and privacy policy.