The research library.
The state of the field — memory science, assessment, AI tutoring, and the findings that didn’t survive replication. Every article is citation-first and peer-review-anchored; Future Proof™ appears only where the evidence earns it.
Memory & practice
The Forgetting Curve, Replicated
Ebbinghaus’s 1885 curve, its 2015 replication, and the most expensive ignored fact in training.
Read →The Testing Effect
Why quizzing beats re-reading — the highest-yield intervention in the learning literature.
Read →The Spacing Effect at Industrial Scale
Distributed-practice meta-analyses and what optimal gaps mean for a workforce.
Read →Desirable Difficulties
Bjork’s programme: why easy training feels good and fails, and struggle predicts retention.
Read →Interleaving
Mixing related-but-distinct topics feels worse and works better. The evidence.
Read →Calibration: Knowing What You Don’t Know
Confidence-accuracy gaps in adults, and why calibration is trainable.
Read →Cognitive Load: The Bottleneck in Every Lesson
Working memory holds about four items. Sweller’s programme on designing instruction that fits.
Read →The Pretesting Effect
Why answering before you’ve learned — and getting it wrong — accelerates the learning that follows.
Read →Successive Relearning
Retrieval to criterion, repeated across spaced sessions — the technique that combines the field’s two best effects.
Read →Sleep: The Second Half of Learning
Memory consolidates offline — the century of evidence that schedules, not just lessons, decide retention.
Read →Background Knowledge: The Skill Nobody Lists
Comprehension and critical thinking run on domain knowledge — the baseball study and its descendants.
Read →Microlearning: The Evidence Behind the Buzzword
Short is not a mechanism. What segmenting, spacing, and retrieval actually contribute — and what “micro” alone buys.
Read →Does Cramming Actually Work?
Yes — for tomorrow. The spacing research on what the exam-day peak costs a month later, and when massing is rational.
Read →Laptop vs. Longhand: What Replication Left Standing
The famous 2014 study, the 2021 replication that found no reliable advantage, and what actually predicts learning from notes.
Read →Assessment science
How 24 Questions Can Map a Mind
Item Response Theory and computerized adaptive testing, from ETS to today.
Read →Bloom’s Taxonomy, Seventy Years On
The 1956 original, the 2001 revision, and why most LMS assessment never leaves recall.
Read →What Actually Predicts Job Performance
A century of selection meta-analysis, including the 2022 re-ranking.
Read →Structured Interviews and BARS
Why anchored scoring beats gut feel in validity and legal defensibility.
Read →The Big Five at Work
What the meta-analyses show — and the honest limits vendors gloss over.
Read →Skills-Based Hiring: Evidence vs. Hype
Degree inflation research and what happens when employers drop the proxy.
Read →Feedback: The Intervention That Can Backfire
Kluger & DeNisi’s uncomfortable meta-analysis — over a third of feedback interventions made performance worse.
Read →Situational Judgment Tests
Scenario-based assessment: ninety years of validity evidence, and the honest boundary conditions.
Read →One Score Is Not a Fact
Error bands, regression to the mean, and why hairline rankings are coin flips wearing arithmetic.
Read →Performance Ratings Measure the Rater
The idiosyncratic-rater evidence — and the structure that actually fixes evaluation.
Read →When Formulas Beat Experts
Seventy years of mechanical-vs-clinical evidence — and the last mile of hiring it indicts.
Read →Formative Assessment: The 0.4 That Became 0.2
Inside the Black Box made it a movement; the meta-analysis cut the number in half. What separates the versions that work.
Read →AI & tutoring
Bloom’s 2-Sigma Problem, Forty Years Later
The 1984 benchmark and how close intelligent tutoring systems actually got.
Read →LLM Tutors: The First Real RCTs
Harvard, Nigeria, Turkey — where AI tutoring genuinely moves outcomes and where it backfires.
Read →Can AI Write Exam-Quality Questions?
Automatic item generation research, and why human review remains a hard gate.
Read →The Socratic Constraint
Worked examples, guidance fading, and when withholding the answer drives learning.
Read →Early-Warning Systems That Actually Warn Early
From Purdue Course Signals to modern at-risk prediction — achievements and failure modes.
Read →Productive Failure: When Struggling First Wins
Problem-solving before instruction loses the session and wins the transfer test — under specific conditions.
Read →What Makes Training Video Work
Mayer’s principles, the six-minute engagement cliff, and why watching alone isn’t learning.
Read →Simulation: Rehearsing the Unrehearsable
Large effects, overrated fidelity — the medical and aviation evidence on practicing rare-critical skills.
Read →Can AI Grade Writing?
Machine scoring matches human raters on-distribution and fails off it — the benchmarks, the critics, the governance.
Read →The Self-Explanation Effect
Learners who explain the material to themselves learn far more — and the behavior turns on with a prompt.
Read →What Makes Training Video Work
The six-minute engagement knee, Mayer’s principles on screen, and why watching is not learning until retrieval interrupts it.
Read →Peer Learning: The Cheapest Large Effect
Peer instruction’s decade of doubled gains, learning-by-teaching — and why unstructured group work is not the same thing.
Read →Cognitive Offloading: Memory in the Machine Age
The Google effect and its replication trouble, GPS and photos, and how to decide what still belongs in a human head.
Read →The uncomfortable evidence
Learning Styles Is Dead. Here’s What Survived.
The meshing-hypothesis takedown and the replication reckoning in learning research.
Read →Why Compliance Training Doesn’t Transfer
The completion-vs-behavior gap in the transfer-of-training meta-analyses.
Read →Gamification: What the Meta-Analyses Say
Streaks and leaderboards work under specific conditions — and backfire under others.
Read →The 10,000-Hour Rule: What the Data Says
Deliberate practice explains 26% of variance in games — and under 1% in professions.
Read →70:20:10: The Ratio That Isn’t Research
The most quoted number in L&D traces to a memory survey of 191 executives. What the evidence actually shows.
Read →Growth Mindset: What the Big Trials Found
From lab promise to d = 0.08 — where mindset interventions still earn a place, and where they don’t.
Read →The Learning Pyramid Is Made Up
10%-of-what-you-read has no study behind it — the citation archaeology of training’s favorite chart.
Read →Brain Training: The Transfer That Never Came
The n-back saga, the 11,430-person test, the FTC case — and the specificity law underneath.
Read →Neuromyths: The Beliefs That Won’t Die
Half of educators endorse refuted brain claims — the prevalence data, and what actually reduces belief.
Read →Smile Sheets: Why Happy Isn’t Learned
Satisfaction predicts learning at the level of noise — the Kirkpatrick critique and the honest stack.
Read →Compliance Training: What Billions of Hours Buy
Ethics meta-analyses, the phishing RCT that backfired, and the formats the learning literature says would actually change behavior.
Read →Skills & the future of work
The Half-Life of Skills
Decay through disuse, obsolescence through change — two measured clocks, and the keynote numbers that measure neither.
Read →Reskilling at Scale: What the Trials Show
The J-curve, the sectoral exception, and the demand-linkage that separates programs from rituals.
Read →AI at Work: The First Field Experiments
Novices gain most, experts barely, and outside the jagged frontier assistance turns negative.
Read →Onboarding, Measured
Role clarity, self-efficacy, social acceptance — the three states that carry retention and ramp speed.
Read →Remote and Hybrid Work: What the Trials Show
Ctrip’s +13%, the hybrid RCT’s free retention win, and where full remote sends the bill.
Read →AI Task Exposure: Which Jobs Actually Change
From 47% panic to task-level precision — what LLM exposure maps measure and how to plan with them.
Read →What Automation Did Last Time
Rising ATMs, rising tellers, and the forty-year dynamo delay — base rates for every AI forecast.
Read →Does Workplace Coaching Work?
The meta-analyses finally reported: moderate, real, and driven by goals, alliance, and dosage.
Read →Leadership Training: Better Than Its Reputation
The meta-analysis says it works — and the design moderators decide how much.
Read →Internal vs External Hiring: What the Data Shows
Paying more to get less — the external premium, the posting effect, and the Peter Principle, measured.
Read →Apprenticeships: The Earn-and-Learn Evidence
Strong returns, firms that break even — and the lifecycle crossover nobody quotes.
Read →Microcredentials: What Employers Actually Value
Signal theory, employer surveys, and the platform test — where badges work and where they’re noise.
Read →What Makes Teams Smart
The collective-intelligence factor, its critics, and the psychological-safety evidence that holds.
Read →Older Workers: The Evidence vs. the Assumption
Age barely predicts performance — the meta-analyses, the assembly line, and the one stereotype that survived.
Read →Motivation & behavior
Goal Setting: The Theory That Kept Its Promises
Specific-difficult beats do-your-best across a thousand studies — plus the learning-goal exception and the dark side.
Read →Habit Formation: The 66-Day Question
Context-stable repetition makes practice automatic — the real curve, the harmless missed day, the cue design.
Read →Motivation: What Rewards Can and Can’t Buy
Incentives buy quantity and corrode interest; autonomy, competence, and relatedness sustain quality.
Read →Test Anxiety: The Measurable Tax
Evaluation itself depresses scores below ability — the working-memory mechanism and the fixes that replicate.
Read →Burnout Interventions: Fix the Job, Not Just the Person
The meta-analytic verdict: modest effects overall, and organization-directed redesign beats individual resilience training.
Read →Curiosity: The Information-Gap Engine
Why the right-sized gap between what you know and what you almost know is the strongest free motivator instruction has.
Read →The science pillars
Memory Science & Spaced Repetition
Why scheduling review before forgetting lifts long-term retention 30–40%.
Read the deep dive →Adaptive Diagnostic & Item Response Theory
How 24 adaptive questions pinpoint a learner as precisely as a 100-item test.
Read the deep dive →Confusion Pairs & Interleaving
The retention boost from mixing related-but-distinct concepts.
Read the deep dive →All seven pillars
Memory, diagnostics, interleaving, misconception repair, calibration, tutoring, knowledge graphs.
Open the library →Reviewing us for a pilot?
We’re happy to share methodology, raw effect sizes, and references with your research team.