Lynnette is an online tutoring software created at CMU’s Human Computer Interaction Institute which middle school students use to solve simple algebraic equations. I was tasked with redesigning the interface with special attention towards making the software more engaging. I worked under the direction of Vincent Aleven and Jonathan Sewall.
User Flow Demo
A high priority was designing not only engaging, but informative interactions which helped demonstrate to students how Lynnette worked. Providing immediate logical feedback for every student input was key, as research showed that students cared about the tutor’s assessment of their learning and that it motivated them to learn.
A major component of engagement is to set concrete, short term goals to motivate students. I wanted to design a more elaborate interaction for the mastery of a skill as a greater reward.
A level is completed by mastering every skill, rewarding the student with a badge, in this case, a planet. The same rationale for creating a significant interaction behind the mastery of a skill applies even more for the mastery of a level.
Hints are a necessary part of online tutors to help nudge students through a problem. A priority with hints was to point them in-between the previous and current step, which made for an interesting visual real estate challenge.
Sketching and creating lo-fi animations in After Effects were the primary methods I used.
I started with documenting all of the student inputs for a problem, seeing what current feedback there already was, and looking into ways that feedback could be extended and improved. I wanted interaction feedback to be immediate, informative, and fun.
While demonstrating how Lynnette works through user feedback, I explored varying levels of explanation fidelity. Creating "juicy interactions" was always a priority in order to increase student engagement.
Showing error history helps remind students how successful they were in solving the problem and allows teachers to easily see who might need their help. I played with different ways of showing the history of a step being completed correctly on the third try (top left), second try (top right), and first try (bottom left).
Skill Bar Updates
Skill bars determined whether a student progressed through a problem set or not, so designing how their inner workings were displayed took a significant amount of time.
I experimented with how explicit to make skill bar updates. In the following two prototypes I realized I had gone too far, as the problem solving process itself was being slowed down due to displaying the changes in a student’s skill bars.
In creating an extra special interaction for mastery of a skill or a level, I wanted to make sure that the initial part of the animation looked just like a regular correct answer to avoid confusion.
These papers provided key insights for how to design Lynnette.
Students' Understanding of their Student Model
Yanjin Long and Vincent Aleven
"We found that:
students pay close attention to the OLM and report that seeing it change encourages them to learn
there is a significant discrepancy between the students’ self-assessment and the system’s assessment
students generally rely on the OLM to make judgments of their learning progress without much active reflection"
Top 20 Principles from Psychology for PreK-12 Teaching and
American Psychology Association
Principle 10: "Students persist in the face of challenging tasks and process information more deeply when they adopt mastery goals rather than performance goals."
Principle 12: "Setting goals that are short term (proximal), specific, and moderately challenging enhances motivation more than establishing goals that are long term (distal), general, and overly challenging.”
Active Learners: Redesigning an Intelligent Tutoring System
to Support Self-Regulated Learning
Yanjin Long and Vincent Aleven
“self-assessment support facilitates students’ learning processes, and enhances their learning outcomes significantly”
“students inspected the OLM frequently to see their progress, but thought that the design of the OLM was too simple to convey much progress information”