Skip to main content Skip to navigation

Incentivizing supplemental math assignments and using AI-generated hints is associated with improved exam performance

Project Overview

The document explores the application of generative AI in education, specifically within an undergraduate physics course aimed at improving student performance. It details a dual intervention strategy that includes incentivized supplemental math assignments and AI-generated hints embedded in homework tasks. The study's findings indicate that these interventions not only enhance exam performance but also help address disparities in student preparedness, particularly benefiting historically underrepresented groups. The results suggest that incorporating AI-generated hints and offering scaled extra credit can significantly boost student engagement and success rates in challenging academic environments, highlighting the potential of generative AI to foster equity and improve educational outcomes.

Key Applications

AI-generated hints and incentivized supplemental assignments

Context: Undergraduate Introductory Physics and Mathematics courses at a public university. The implementations involve integrating AI technology to provide personalized support and incentives for students to engage with the coursework, addressing academic challenges across different sub-domains.

Implementation: AI-generated hints are embedded within homework assignments on an online platform (Kudu) to assist students with problem-solving. Additionally, supplemental assignments are offered with incentivized extra credit points to motivate student participation, particularly for those with lower exam scores.

Outcomes: The use of AI-generated hints resulted in increased exam performance, particularly for students from underrepresented backgrounds, with 88% of students utilizing the hints. The incentivized assignments led to higher completion rates, especially among students needing more support, and reduced disparities in completion rates across racial groups.

Challenges: The effectiveness of AI hints is still relatively unknown, and access to AI technology may vary among students. Additionally, while completion rates improved, they remained below 60%, and motivating students to participate in optional assignments continues to be challenging.

Implementation Barriers

Equity Barrier

Disparities in access to advanced mathematics courses based on racial and socioeconomic privilege.

Proposed Solutions: Incentivizing supplemental assignments and integrating AI-generated hints to support underprepared students.

Participation Barrier

Low completion rates for optional supplemental math materials.

Proposed Solutions: Using scaled extra credit to motivate students and considering integrating materials into formal class time.

Project Team

Yifan Lu

Researcher

K. Supriya

Researcher

Shanna Shaked

Researcher

Elizabeth H. Simmons

Researcher

Alexander Kusenko

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Yifan Lu, K. Supriya, Shanna Shaked, Elizabeth H. Simmons, Alexander Kusenko

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies