Students forget nearly 40 percent what they study within a day. To retain longer, it requires revision at specific intervals, based on a student’s forgetting curve.

To address this, schools in India follow an academic revision cycle which starts about 30- 45 days before the final exams. However, in reality, when a student revises a chapter in December or say Jan, it is equivalent of studying the chapter for the first time. See a problem?

Students forget 40% of what they study with 24 hours. Repeating information within 24 hours boosts recall.

Further, when we looked at ‘quiz attempted’ data on QuizNext, it was evident that the pattern in which students picked chapters for practice did not help them to retain learning longer. Over 36% students, hardly come back to practice chapters where they performed well or performed badly.

This left us with two insights:

  • Students do not know what chapters to practice
  • Students don’t know how often they should practice a chapter

As a learning app, we knew we wanted to change this ineffective way for revising. So we started to deep- dive and find solutions to this problem. 

Luckily, we had already solved a part of this problem i.e. we knew how we could get our users to take up a quiz on a chapter. Gamification elements in the app such as live challenges and school squads have consistently engaged users ( more on this in the next post).

The challenging part was to identify intervals for suggesting revision to students for reducing effects of forgetting curve, which is unique to all students. Further, the rate for recall varies by subjects too.

For example, a CBSE Class 10 student Riya, might remember a chapter ‘Magnetic Effects of Electric Current’ better after 15 days, compared to a chapter ‘Acids Bases and Salts’ in Chemistry.

So a predefined ‘spaced repetition’ ( Ex: 24 hours, 72 hours, 15 days) would do only half good and would not have delivered what we intended to.

“Forgetting Curve for students vary, so the time intervals between revision varies”

To understand this better, we carried out an intervention. We started sequencing and scheduling quizzes for our ‘active user set’, at cohort level. We wanted to determine primarily two things:

  • Relationship between quiz performances and retention index
  • Impact on ‘Quiz Performance’ with delayed revision cycle

The experiment proved successful when a relationship emerged between ‘quiz performance’ at any point of time and the ‘Retention Index’ with acceptable degree of confidence, at cohort level. With more quiz attempts, Retention Index only got better, giving us confidence to roll it out for all users. Thus Revise+ was created.

What is Revise+?

Revise+ is an AI based revision scheduling system that considers quiz performances of students for a chapter to decide the right day for revision for that chapter. It determines forgetting curve of students based on subjects, chapter quiz performance, questions exposed and other parameters to schedule revision recommendations. 

What’s next?

At the moment, we have been able to recommend chapters at the cohort level, where cohorts are based on ‘chapters practiced’ and student’s overall performance index. It would be interesting to completely personalize these recommendations for every user and also understand the way to achieve this even for new users, from the first attempt.

Get the QuizNext app here