Artificial Intelligence for Math Learning in Schools
Narrative Description of Project:
Convergent was commissioned to execute a comprehensive research study for MIT and J-PAL that investigated the cognitive processes used by Grade 9 students while solving written mathematics problems, with a specific focus on linear equations and algebraic reasoning. The study engaged 10,000 students across multiple states, who independently attempted a structured set of problems in a controlled environment without external assistance. Convergent designed the implementation methodology to enable systematic analysis of students' problem-solving approaches, detailed categorization of error types, and identification of the underlying thought patterns that influenced their mathematical reasoning. The analysis conducted by Convergent revealed critical insights into students' conceptual understanding and strategic approaches, identifying specific cognitive gaps and persistent misconceptions. These findings provided valuable evidence that highlighted key areas where targeted pedagogical interventions and curriculum adjustments could significantly enhance learning outcomes in secondary mathematics education.
Description of Actual Services Provided:
Convergent was responsible for the end-to-end field execution and data management of this large-scale project. Our services included:
- Full-scale Study Execution: Managed the entire on-ground research process for 10,000 students.
- Data Collection: Administered standardized assessments and collected all answer sheets.
- Data Digitization: Scanned and digitized all student responses for analysis.
- Data Delivery: Securely transferred the complete dataset to MIT and J-PAL investigators.
Research Methodology
Data was collected using the Pen-and-Paper Personal Interview (PAPI) method, with all materials in English.
- Quantitative Phase
10,000 surveys with Grade 9 students across Delhi, Uttar Pradesh, Haryana, Telangana, West Bengal, Maharashtra, and Rajasthan.
