The science behindsmarter exam prep.
Every algorithm in our engine traces back to peer-reviewed cognitive science. We built what the research said should work — then proved it did.
MERIDIAN LABS is an applied cognitive science company. While most exam prep apps rely on basic flashcard logic, we engineered a learning engine grounded in decades of research on memory, metacognition, and testing psychology. The result: a system that doesn't just quiz you — it learns how you learn.
6 algorithms. One integrated system.
Each algorithm solves a specific learning problem identified in cognitive science research.
Confidence Calibration
The ProblemStudents consistently overestimate what they know — the most dangerous gap on exam day.
Our ApproachEvery answer is scored against your self-reported confidence, creating a 4-outcome weighting system that detects overconfidence before it becomes a failed exam.
The EvidenceDecades of metacognition research show learners systematically misjudge what they know, and that the gap between judged and actual knowledge predicts exam performance (Koriat & Goldsmith, 1996; Dunning & Kruger, 1999).
Adaptive Spaced Repetition
The ProblemWithout strategic review scheduling, study progress decays within days.
Our ApproachA 6-level mastery system with dynamically expanding intervals. Overconfident errors are penalized more harshly than humble mistakes — because knowing you don't know is safer than thinking you do.
The EvidenceA quantitative synthesis of 317 experiments across 184 published studies found that spacing study sessions apart reliably improves later recall over massed cramming (Cepeda et al., 2006).
Critical Misconception Detection
The ProblemConfident wrong answers are invisible to the student and devastating on exam day.
Our ApproachA dedicated detection system flags answers where confidence is high but accuracy is low, then elevates those items to maximum review priority until resolved through consecutive correct answers.
Multi-Factor Exam Readiness
The ProblemA student at 80% accuracy is not 80% ready for their exam.
Our ApproachA 4-factor readiness algorithm with non-linear scaling that weighs accuracy, consistency, confidence calibration, and domain coverage — because readiness is multidimensional.
Bloom's Taxonomy Integration
The ProblemMost prep apps test only recall and recognition — the two lowest cognitive levels.
Our ApproachQuestions are mapped across all 6 cognitive levels of Bloom's revised taxonomy, with progressive difficulty across 4 practice test levels that mirror real exam complexity.
Stress Inoculation Training
The ProblemTest anxiety can drop scores by 10–15% even when knowledge is strong.
Our ApproachTimed challenges with progressive pressure thresholds build exam-day resilience. By the time you sit for the real exam, the pressure feels familiar.
Every answer feeds the engine
Answer
You respond and indicate your confidence level.
Score
4-outcome confidence weighting applied.
Schedule
Spaced repetition interval assigned based on mastery.
Flag
Misconceptions detected and elevated to priority review.
Readiness
Multi-factor exam readiness score recalculated.
Certificate
All factors converge — Exam Ready certification earned.
What the engine runs on
Built on decades of peer-reviewed science
Our algorithms draw from four major research domains.
Cognitive Science
Memory encoding, retrieval practice, spacing effects, and desirable difficulties.
Metacognition
Confidence calibration, knowledge monitoring, and self-assessment accuracy.
Habit Formation
Consistency triggers, streak psychology, and behavioral automaticity.
Test Anxiety
Stress inoculation, arousal reappraisal, and performance under pressure.
Every Pro plan is backed by a Pass Guarantee
We've built our revenue model on the confidence that our methodology works. If you earn your Exam Ready Certificate and don't pass, we refund your subscription.
See Pass Guarantee Details →Research-validated methodology
Our engine has been analyzed across all four deployment domains using 16,450 production items. The results confirm what we built: confidence calibration works.
Perry, A. (2026). Confidence-Calibrated Adaptive Learning: An Integrated Adaptive Engine for Professional Exam Preparation. Zenodo.
Read Full Paper →Perry, A. (2026). Cross-Domain Analysis of a Confidence-Calibrated Adaptive Learning Engine. Zenodo.
Read Full Paper →Research references
33 peer-reviewed sources across cognitive science, metacognition, habit formation, and test anxiety.