Data Needed for Learning Analytics

[jo-elimu]

To perform learning analytics (and enable personalized learning per student), we need some clarifications regarding the data and how it fits into the platform's software architecture:

Gradual Unlocking of Content

• Initially, a student is only able to access a small part of the curriculum. And over time, more advanced content will be unlocked.
• More advanced content will only be unlocked if the student demonstrates mastery of prerequisite skills.

Learning Events & Assessment Events

• As a student interacts with the apps and content, we collect two types of events; LearningEvent and AssessmentEvent.
• A LearningEvent is recorded whenever a student views an educational resource, while an AssessmentEvent is recorded whenever a student's knowledge level is assessed for an educational resource.

Definition of Mastery

• We could define mastery as "At least 80% of the 10 most recent assessments were answered correctly," for example. Or "The 6 most recent assessments were all answered correctly."

Mastery at Multiple Levels

At the highest level, we have mastery of the literacy skills and numeracy skills. And within some of those, there are educational resources to be mastered. For example; Achieving 100% mastery of the literacy skill LETTER_IDENTIFICATION would require a student to learn the sound-letter correspondences for each sound in the language.

Literacy Skill Mastery

• 11 skills: LiteracySkill

Numeracy Skill Mastery

• 11 skills: NumeracySkill

Two Types of Mastery

• We can stick with the definition of mastery as having correctly answered a sequence of assessments in the past, but what if the student took a break from studying for an extended period?
• Besides a mastery score, we can introduce a second type of score—memory score. A memory score will also be calculated based on assessment events, but its score will drop over time.