From the Student Success edition of October 24, 2026
Michigan State study: assigning Khanmigo activities raised AI-tutor use in undergraduate algebra, and use predicted grades after controlling for prior GPA (n=58)
Michigan State University (study authors); Khan Academy (Khanmigo tool) · Higher education AI tutoring in gateway mathematics · East Lansing, Michigan, United States
- Publisher
- International Journal of Education in Mathematics, Science and Technology
- Original publication
- International Journal of Education in Mathematics, Science and Technology, Early View, published 22 October 2026
- Source retrieved
- 2026-10-25
What happened
Michigan State University researchers evaluated undergraduate algebra students' use of Khanmigo, a Socratic generative AI tutor, in a quasi-experimental mixed-methods study (n=58) using survey, registrar, grade and software-usage data. One section was assigned four Khanmigo activities across the semester; the comparison section had access and only an introductory activity. Students in the assigned section held significantly more chats (Wilcoxon rank-sum, p<0.05). Instrumental help seeking correlated positively with use, while intentions to seek help from technology sources correlated negatively. Khanmigo use significantly predicted mathematics grades after controlling for prior GPA and course section. Most students used it for homework help and exam preparation; lack of visuals and the Socratic design limited perceived usefulness for some. The authors conclude that integrating the tutor into class activities and setting expectations may improve effective use.
Why it matters
Directly relevant to gateway-math student success: it suggests that uptake of an institution-provided AI tutor depends on assigning it within the course, not just licensing it.
Evidence and measured results
Small two-section quasi-experiment at one US public university, combining usage logs, registrar data, grades and surveys. Association with grades, not a randomized effect.
Limitations and uncertainty
n=58 across two sections; non-random assignment; the grade association could reflect motivation. Only the abstract was inspected.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-25; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Licensed AI tutors in gateway math courses go unused unless faculty build them into coursework.
- Stakeholders
- Mathematics department chair
- Gateway-course instructors
- Center for teaching and learning
- Institutional research
- Discovery questions
- Which gateway courses have AI tutor access, and how many students use it?
- Do instructors assign tutor activities or only announce access?
- Can usage be linked to grades for evaluation?
- Value hypothesis
- Interpretation: assigning a few tutor activities can raise use, and use may be associated with better grades in gateway math.
- Potential engagement
- A one-term design-and-evaluation pilot of assigned AI tutor activities in a single gateway math course.
- Unsupported claims
- That Khanmigo causes higher grades
- That results generalize beyond one university and 58 students
- Any retention or completion effect
Pre-sales engineering
Role takeaway
- Fit
- Institutions with or considering licensed AI tutors in mathematics.
- Architecture and integration
- Vendor tutor used alongside the LMS, with usage logs linked to registrar and grade data for analysis.
- Prerequisites
- Per-student usage export from the tutor
- Identity matching to the student information system
- Research or data-use approval
- Constraints
- Small single-institution sample
- Text-only Socratic design
- Vendor data terms not described
- Security
- Minimize linked fields, restrict access to the analysis dataset and confirm vendor retention and deletion terms.
- Proof of value
- Compare usage between assigned and access-only sections
- Model grades controlling for prior GPA
- Survey usefulness, including need for visuals
Delivery
Role takeaway
- Work
- Embed a small set of required AI tutor activities in a gateway math course and evaluate use and outcomes.
- Dependencies
- Instructor design time
- Tutor licence and log access
- Institutional research support
- Ownership
- Mathematics department with the center for teaching and learning.
- Skills and adoption
- Instructors learn to write tutor activities and set expectations; students are shown how to use the tutor for practice rather than answers.
- Governance checkpoints
- Data-use approval before linking records
- Mid-term usage review
- End-of-term outcome review before expansion
- Acceptance criteria
- Proposed: assigned sections show higher tutor use than access-only sections
- Proposed: grade comparison controls for prior GPA
- Proposed: usefulness feedback collected from at least half of students
- Risks
- Low uptake if activities are optional
- Misreading association as causation
- Design not suited to students who need visuals
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Uses a commercial tutor alongside the course; usage logs were linked with registrar and grade data for analysis. Interpretation: evaluating similar pilots needs log export and record linkage agreed with the vendor in advance.
Governance
Who approves, reviews and stays accountable for outcomes?
Not described in the abstract: consent, data sharing with the vendor or policy on use in assessed work.
Security and privacy
What data, permissions and controls need testing?
Linking tutor logs to registrar and grade data requires research approval and data-handling controls; specifics are not in the abstract.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Some students found the Socratic, text-only design less useful (lack of visuals), which matters for learners who need worked examples or visual explanation.
Procurement
What should contracts, pricing and exit terms secure?
A tutor licence alone may not drive use; contracts should include usage-data access for evaluation.
Operating model
Which teams own the service once it runs?
Instructors embedding a small number of required activities was associated with greater use, implying faculty design time is part of the cost.
Publication history
- 2026-10-24Student Success · Issue 455 resources
Stable resource ID: ijemst-msu-khanmigo-socratic-ai-tutor-undergraduate-algebra-quasi-experiment-20261022