Research foundation ยท Observability
Modality Effectiveness Lab
Which modality helps which concept โ and which grade-band. This is the data layer for future modality-vs-learning-speed studies.
Observational, not causal. The modality shown here is learner-chosen โ it records which study material a learner opened, not a modality the bandit assigned. It never feeds the MAB and never touches the sealed evaluation (headline AUC 0.6051 is untouched). Read these charts as "learners who chose X gained Y", a foundation for a future controlled study of modality โ learning speed, not as a causal claim today.
1 ยท Average mastery gain by modality
Mean per-learner mastery gain on the concepts each modality was opened for. Bars are the observed average; the caption below each is the sample size.
๐ญ
0 engagements so far โ no modality has enough data yet.
The chart fills in as learners open study materials on /learn.
2 ยท Best modality per concept
For each concept, which modality shows the highest observed mastery gain. Winning modality is starred in the table.
๐ญ
0 concepts with engagements โ need materials opened across concepts.
The chart fills in as learners open study materials on /learn.
3 ยท Grade-band ร modality
Observed mastery gain by grade-band (K-2 / K-5 / K-8 / K-12) and modality โ which modality suits which age band.
๐ญ
0 grade-band-tagged engagements โ fills as learners across grades use materials.
The chart fills in as learners open study materials on /learn.
The point of this page
Research questions this answers
Each question below maps to a chart above. This is the observability foundation the thesis reviewer asked for โ it makes modality a first-class research variable.
Chart 1How does mastery gain differ per modality โ is reading, interactive, or audio associated with more learning on average?
Chart 2Which modality suits which concept โ does a given concept learn best via diagram vs reading vs interactive?
Chart 3Which grade-band suits which modality โ do younger (K-2) learners benefit more from audio/interactive than older (K-12)?
Chart 1 + attemptsHow fast per modality โ average attempts-to-mastery alongside gain hints at learning speed per modality (foundation for a timed study).
Next step for a causal answer: randomise the offered modality (A/B) instead of letting the learner choose, then re-run these three cuts.
No engagements recorded yet. As learners open reading / diagram / audio / interactive / video materials on /learn, a beacon records the modality and this lab fills in.