The Science of Learning
Algo School is built on decades of learning science research. Here's the evidence-based foundation behind every feature.
Bloom's Taxonomy
Benjamin Bloom's framework categorizes learning objectives by cognitive complexity. Algo School ensures assessments cover all levels—not just basic recall.
Cognitive Levels (Higher to Lower)
Why It Matters
Most assessments cluster at "Remember" and "Understand." Higher-order thinking (Apply, Analyze, Evaluate, Create) is often under-assessed, leaving gaps in true learning.
How Algo School Uses It
Every learning objective and question is tagged with its Bloom's level. Teachers can request balanced coverage, and leaders can verify cognitive rigor across the curriculum.
Per-Level Mastery
Track student mastery at each cognitive level. A student may remember facts but struggle to apply them—Algo School reveals these nuances.
Spaced Repetition & The Forgetting Curve
Hermann Ebbinghaus discovered that memory decays exponentially—but strategic review intervals can dramatically improve retention.
The Forgetting Curve
Without review, students forget approximately 50% of new information within an hour, 70% within 24 hours, and 90% within a week. But each review resets and flattens the curve, building durable memory.
Optimal Spacing
Review intervals expand as mastery grows: 1 day → 3 days → 1 week → 2 weeks → 1 month. This maximizes retention with minimal review time.
Automatic Scheduling
Algo School tracks when each concept was last practiced and schedules reviews at optimal intervals—no manual tracking required.
Personalized to Each Student
Review schedules adapt to individual mastery levels. Struggling concepts get more frequent review; mastered concepts are spaced further apart.
Multi-Dimensional Difficulty
"Easy/Medium/Hard" is too simplistic. Real difficulty has multiple dimensions that affect different students differently.
Cognitive Load
How much working memory does the question require? Multi-step problems have higher cognitive load than single-step ones.
Abstraction Level
Is the question concrete or abstract? Abstract concepts require more mental effort to process.
Prerequisite Depth
How many prerequisite concepts must be understood first? Deeper prerequisite chains increase difficulty.
Context Familiarity
Is the problem set in a familiar context? Unfamiliar contexts add difficulty beyond the core concept.
Response Format
Multiple choice vs. open response vs. constructed response—each format has different difficulty characteristics.
Time Pressure
Speed requirements add difficulty. Algo School considers whether questions are timed and how that affects challenge.
Why This Matters
A question can be "easy" on one dimension and "hard" on another. Multi-dimensional analysis provides nuanced difficulty calibration that simple labels cannot.
Mastery Levels & Bayesian Knowledge Tracing
A percentage score doesn't tell you if a student truly understands a concept. Bayesian Knowledge Tracing (BKT) provides a probabilistic model that's more accurate.
P(L) - Learn Rate
The probability that a student learns a concept after practicing it. Varies by concept difficulty and student.
P(G) - Guess Rate
The probability of a correct answer by guessing. Accounts for lucky correct answers that don't indicate true mastery.
P(S) - Slip Rate
The probability of an incorrect answer despite knowing the concept. Everyone makes mistakes sometimes.
P(K) - Prior Knowledge
The initial probability that a student knows a concept before any practice. Based on prerequisites and prior performance.
Mastery Confidence
BKT provides not just a mastery estimate but a confidence level. Low data means low confidence—the system distinguishes between "unknown" and "struggling."
Dynamic Updates
Mastery estimates update in real-time with each student response. The model continuously refines its understanding as more evidence accumulates.
Mastery Progression
As students demonstrate understanding through correct responses, their mastery probability increases. But guessing won't fool the system—sustained performance is required to reach true mastery.
- Novice (0-25%): Just starting, limited exposure
- Developing (25-50%): Some understanding, inconsistent
- Proficient (50-75%): Solid grasp, occasional errors
- Master (75-100%): Deep understanding, reliable performance
Knowledge Graphs in Education
Knowledge graphs model the structure of knowledge itself—showing how concepts relate and depend on each other.
Concepts, Not Lessons
Lessons are delivery mechanisms. Concepts are the actual knowledge units that students learn. Knowledge graphs focus on what matters: understanding.
Prerequisite Relationships
Some concepts must be learned before others. Knowledge graphs make these dependencies explicit, enabling truly personalized learning paths.
Cross-Curriculum Connections
The same concept appears in multiple subjects. "Ratio" connects math and science. Knowledge graphs reveal these connections.
Gap Diagnosis
When a student struggles, the knowledge graph traces back to find prerequisite gaps—the root cause, not just the symptom.
Adaptive Sequencing
Instead of linear curriculum progression, the graph enables adaptive paths that respond to individual mastery patterns.
Unified Understanding
The knowledge graph provides a single source of truth for what's being taught, assessed, and mastered across the entire organization.
Experience Learning Science in Action
See how these research-backed methods work together in Algo School.
Request Demo