AI-Powered Adaptive Tutoring System for Middle-School Mathematics

1. Overview

This study proposes the development and evaluation of an AI-powered adaptive tutoring system designed to improve learning outcomes in middle-school mathematics.

Scientific rationale & importance: Prior studies show that personalized learning interventions significantly increase student engagement and retention [1–3]. However, current systems are expensive and closed-source. An open, data-driven alternative could broaden access to adaptive education.
Funding/facilities/ethics: Funding of 1,000 USD is requested to cover participant recruitment, software development, and cloud computing costs. We will seek ethics approval from the [University/IRB placeholder].
Timeline: 4 months — data collection (Month 1), intervention deployment (Month 2), evaluation (Month 3), analysis and reporting (Month 4).
Data sharing: All anonymized datasets, code, and analysis scripts will be made available on GitHub and OSF.

2. Introduction

Background & literature. Adaptive tutoring systems leverage machine learning to provide individualized instruction, but most remain proprietary [4,5]. Open-source approaches are rare, limiting reproducibility and access.

Research questions.

RQ1: Can an open-source AI tutoring system significantly improve student performance in mathematics compared to a control group?
RQ2: How does adaptive personalization affect student engagement?

Hypotheses.

H1: Students using the AI tutor will show greater improvements on standardized math assessments than the control group.
H2: Students will report higher engagement and satisfaction with personalized interventions compared to static lessons.

Impact statement. Regardless of the outcomes, this study will provide crucial evidence on the viability of open-source AI tutors and create reusable infrastructure for educational research.

3. Methods

Participants/Sample

Target sample: 100 middle-school students (ages 11–13).
Inclusion criteria: Enrolled in mathematics courses, parental consent obtained.
Exclusion criteria: Diagnosed severe learning disabilities that require specialized interventions outside scope.
Recruitment: Partnering schools in [City/Region].
Sample size justification: A priori power analysis (α=0.05, power=0.80, d=0.5) indicates N=64 needed; we aim for 100 to account for attrition.

Materials and Procedures

Intervention: AI tutor providing adaptive problem sets tailored to student progress.
Control: Standardized worksheet-based practice.
Procedure: Pre-test → 4-week intervention → post-test + survey.
Setting: School computer labs and monitored online sessions.

Planned Analyses

Preprocessing: Exclude incomplete datasets; normalize test scores.
Statistical tests:
- Pre-post gains: Mixed ANOVA (Group × Time).
- Engagement: Independent t-tests on survey scores.
Contingent decisions:
- IF dropout >20%, THEN conduct intention-to-treat analysis with multiple imputation.
- IF data non-normal, THEN use non-parametric alternatives (Wilcoxon, Kruskal–Wallis).

Ethics and Data Management

Ethics approval will be obtained from [Institution IRB].
Student data anonymized, stored securely on encrypted servers.
All anonymized data, scripts, and materials archived on OSF.

4. Pilot Data (Optional)

A small internal test (N=12) indicated feasibility: students completed >85% of tasks and reported the system easy to use. Gains were modest but consistent, suggesting full-scale deployment is viable.

5. Budget

	ItemEstimated Cost (USD)
Participant recruitment incentives	300
Cloud hosting & compute resources	250
Software development support	300
Data management & storage	100
Miscellaneous (printing, admin)	50
Total	1,000

6. References

Bloom, B. S. The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educ. Res. 13, 4–16 (1984).
Pane, J. F. et al. Continued progress: Promising evidence on personalized learning. RAND Corporation (2015).
VanLehn, K. The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educ. Psychol. 46, 197–221 (2011).
Knewton. Adaptive learning platform overview. (2014).
Xu, Y. et al. A review of adaptive learning systems. Comput. Educ. 173, 104276 (2021).

AI-Powered Adaptive Tutoring System for Middle-School Mathematics

AI-Powered Adaptive Tutoring System for Middle-School Mathematics

Funding Progress

Funders

Related Funding

2-NEW RFP For Neuroimaging Studies

Nonprofit Recipient

Topics

DOI