Evidence-Based Learning
Built on science,
not assumptions
Every algorithm, every design decision, every pedagogical choice in Econ Academy is grounded in published cognitive science and learning research. Here is exactly what we use and why.
Under the Hood
Five algorithms working together
Our adaptive learning system has five layers. Each is backed by published research and chosen for the balance of effectiveness, implementation quality, and data requirements.
Free Spaced Repetition Scheduler(FSRS)
Schedules reviews at the mathematically optimal moment — right before you'd forget. Each successful recall strengthens the memory and pushes the next review further out.
Produces 20-30% fewer reviews than SM-2 for the same retention. Default parameters trained on 20,000+ learner histories.
open-spaced-repetition/fsrs4anki — FSRS Algorithm WikiFractional Implicit Repetition(FIRe)
When you practice an advanced topic, you implicitly review its prerequisites. FIRe gives partial credit for that implicit practice, reducing total reviews needed.
Inspired by MathAcademy's hierarchical spaced repetition. Prerequisite review dates are pushed back proportionally to encompassing weights.
Skycak, J. — Individualized Spaced Repetition in Hierarchical Knowledge StructuresPerformance Factor Analysis(PFA)
Tracks your successes and failures per topic to produce a continuous mastery score. Predicts your probability of answering correctly before you even see the question.
Better predictive accuracy than Bayesian Knowledge Tracing for multi-skill items. Simpler to implement, handles overlapping concepts naturally.
Pavlik, Cen & Koedinger (2009) — Performance Factors Analysis, AIEDItem Response Theory(IRT)
Calibrates the true difficulty of every question using real student data. Replaces guesswork with statistical precision — after 100+ responses, we know exactly how hard each question is.
The gold standard for educational assessment (used by SAT, GRE, GMAT). Our diagnostic placement test uses IRT-style adaptive item selection.
De Ayala, R.J. — The Theory and Practice of Item Response TheoryZone of Proximal Development(ZPD)
Questions are selected to be challenging but achievable — targeting a 70-85% correct probability. Too easy and you're bored; too hard and you're frustrated.
Combines Vygotsky's ZPD with Csikszentmihalyi's flow state research. MathAcademy targets ~80% quiz accuracy for the same reason.
Vygotsky, L.S. (1978) — Mind in Society; Csikszentmihalyi (1990) — FlowCognitive Science
10 evidence-based learning principles
These principles are not theoretical aspirations — they are implemented in the platform's code. Each one is active every time you study.
Retrieval Practice
Answering questions is the primary learning activity, not re-reading. Even failed retrieval attempts improve memory — as long as feedback follows. We prioritize active problem-solving over passive consumption.
Research: Latimier et al. (2020) — meta-analysis, spacing + testing combinedSpaced Practice
Practice distributed over days and weeks dramatically outperforms cramming. Our FSRS algorithm automatically spaces your reviews at expanding intervals — 1 day, then 3, then 8, then 21 — precisely timed to the forgetting curve.
Research: Cepeda et al. (2008) — Spacing effects in learning, Psychological ScienceInterleaving
Review sessions mix questions from multiple topics rather than drilling one subject. This creates 'desirable difficulty' that forces your brain to discriminate between problem types — a critical skill for exams.
Research: Rohrer (2012) — Interleaving helps students distinguish among similar concepts, Educational Psychology ReviewGeneration Effect
Generating an answer is more effective than selecting from options. We use a mix of question types — numerical input and short answer (generation) alongside multiple choice (recognition) — because producing knowledge builds stronger memory traces than recognising it.
Research: Slamecka & Graf (1978) — The generation effect: Delineation of a phenomenonDesirable Difficulties
Spacing, interleaving, and retrieval practice all feel harder than re-reading. Students often prefer easier methods because of 'fluency illusions.' We embrace productive struggle — and explain to students why it works.
Research: Bjork & Bjork (2011) — Making things hard on yourself, but in a good wayScaffolding & Expertise Reversal
New topics start with full worked examples. As mastery grows, scaffolding fades — because what helps a novice actually hurts an expert. We adapt the level of support to your demonstrated skill.
Research: Kalyuga et al. (2003) — The expertise reversal effect, Educational PsychologistCognitive Load Management
Each lesson teaches one concept in 8-15 minutes. Text and diagrams are integrated spatially (not separated). We never present the same information in three formats simultaneously. Every design choice minimises unnecessary mental effort.
Research: Sweller (1988) — Cognitive load during problem solving, Cognitive ScienceSelf-Explanation Prompts
After incorrect answers, we prompt you to reflect on why the correct answer is right. Students who self-explain develop more accurate mental models than those who simply move on.
Research: Chi et al. (1994) — Eliciting self-explanations improves understanding, Cognitive ScienceNon-Interference
We never teach price elasticity of demand and income elasticity of demand in the same session. Related concepts are spaced apart to prevent 'associative interference' — the confusion that arises when similar material is learned too close together.
Research: MathAcademy pedagogy — Non-interference designMisconception Targeting
Research shows standard economics instruction barely reduces common misconceptions. We build 'misconception buster' questions that present the popular wrong answer as an attractive option — then explain exactly why it fails.
Research: Busom, Lopez-Mayan & Panadés (2017) — Students' persistent preconceptions and learning economic principlesEconomics Pedagogy
Designed for economics, not adapted from math
Economics differs from mathematics in important ways. Our content and assessment are designed specifically for how students learn economic reasoning.
Graph Literacy as a Skill
Economics is uniquely graph-dependent. Supply/demand diagrams, cost curves, and market structure graphs are the visual language of the discipline. We teach graph reading as an explicit skill with consistent visual conventions — demand curves are always blue, supply curves are always red, price is always on the Y-axis.
Source: Cohn et al. (2001) — The effect of economic literacy courses on student learning
Conceptual over Procedural
Economics requires understanding why, not just how. Our questions test reasoning ('Explain why a price ceiling causes a shortage') rather than rote calculation. This builds transferable economic intuition that survives beyond exam day.
Source: Simkins & Allen (2000) — Active learning in introductory economics
TUCE-Aligned Assessment
Our diagnostic and unit assessments are modeled on the Test of Understanding in College Economics (TUCE) — the validated standard for measuring economics comprehension. These items specifically target conceptual understanding, not surface-level recall.
Source: Walstad, Watts & Rebeck (2007) — Test of Understanding in College Economics (4th ed.)
The learning loop
Every session reinforces long-term retention through a cycle of retrieval, feedback, and optimally-timed review.
Learn
Read the explanation and worked example
Practice
Answer questions with immediate feedback
Master
Get 2 consecutive correct to advance
Review
FSRS schedules spaced reviews automatically
Science-backed learning, free to start
No signup fees. No gimmicks. Just evidence-based economics education.
Start Learning Now