Teacher Comparative Test Overview
Overview
- Ten teachers (Six science teachers and four math teachers) participated in an extensive survey that walked them through the following tutors: Flexi, Khanmigo, and Q-Chat.
- Teachers used each of the three AI tutors and provided their feedback and preferences.
Purpose
The purpose of this study is to provide comparative data with Flexi and similar AI tutors. We want to determine which tutor teachers would prefer for their students to use, the teachers' attitudes on the functionality of each tutor, and the conditions that are optimal for the use of each tutor.
Methodology
Teachers were instructed to enter 4 pre-determined questions into Flexi, Khanmigo, and Q-Chat. These questions were designed to showcase the capabilities of all four tutors. Math teachers asked the math questions, and science teachers asked the science questions. After exploring the given questions, the teachers were instructed to explore each of the AI tutors on their own (20 minutes each).
Key Takeaways
- Students showed a clear preference for answers with imagery preferred responses with imagery in 81% of opportunities.
- For non-image-based questions, students preferred Flexi in 60.1% of opportunities.
- Students did not have a strong preference between the two options they were given over the 12 questions.
- Students didn’t like responses that were too long or too short; 200-300 characters seems like the ideal response range for minimizing cognitive load.
- Students had stronger negative responses to receiving shorter and more direct answers in science vs. math.
CK-12 Foundation is a non-profit organization that provides free educational materials and resources.
