PANELA government or expert panel synthesized the literature and issued a recommendation.
Retrieval practice
Make the learner pull the answer out of their own head before showing it to them. Recalling something is what strengthens it; rereading it mostly builds a feeling of familiarity that gets mistaken for knowing.
What it is not: Not high-stakes testing. The value is in the act of recall, which means the score must not matter.
At every level
- PreK 2:
- Ask the child to tell it back before you re-read it. 'What happened first?' before turning the page.
- grades 3 5:
- Two or three no-stakes questions at the start of a lesson, on last week's material, not today's.
- grades 6 8:
- Blank-page recall: close everything, write what you remember, then check.
- grades 9 12:
- Self-quizzing from a question bank the learner wrote themselves.
- adult:
- Every Academy unit ends with a reveal-answer check; every Arcade round is retrieval wearing a game's clothes.
IES/What Works Clearinghouse practice guide 'Organizing Instruction and Study to Improve Student Learning', Recommendation 5: use quizzing with active retrieval at all phases of learning. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Spaced practice
Spread the same material across days instead of massing it into one session. The forgetting between sessions is not wasted time — it is the mechanism.
What it is not: Not a daily-goal nag. Outpost never manufactures anxiety to produce a return visit.
At every level
- PreK 2:
- The same five sounds, four short times a week, not one long Friday.
- grades 3 5:
- Weekly review of a topic from three weeks ago, permanently on the calendar.
- grades 6 8:
- Cumulative homework: a third of every set comes from earlier units.
- grades 9 12:
- A revision schedule built backwards from the exam, not crammed forwards from today.
- adult:
- The Arcade day streak exists for this and nothing else. It rewards returning, never duration.
Same WWC guide, Recommendation 1: space learning over time. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Interleaving and worked examples
Alternate studying a solved example with solving one yourself, and mix problem types rather than blocking them. Blocked practice feels smoother and produces less durable learning.
What it is not: Not random. The mix is designed so the learner has to identify the type, which is the skill.
At every level
- PreK 2:
- Watch me do one, then you do one, then me, then you.
- grades 3 5:
- Mixed problem sets where you must first decide which kind of problem it is.
- grades 6 8:
- Fade the worked example: full, then half, then blank.
- grades 9 12:
- Mixed review where the hard part is choosing the method.
- adult:
- Sorting games mix categories on purpose; the difficulty is classification, not recall.
Same WWC guide, Recommendation 2: interleave worked example solutions with problem-solving exercises. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Words and pictures together
Pair a verbal explanation with a graphic that carries the same structure. Two channels, one idea — not decoration.
What it is not: Not illustration for mood. A picture that does not carry information is cognitive load with no payoff.
At every level
- PreK 2:
- Say it, show it, act it out.
- grades 3 5:
- Timelines and diagrams the learner draws, not ones handed to them.
- grades 6 8:
- Flowcharts of a process before prose about the process.
- grades 9 12:
- Diagram first, then write the paragraph the diagram implies.
- adult:
- Every civic process in Outpost gets a structural graphic, not stock photography.
Same WWC guide, Recommendation 3: combine graphics with verbal descriptions. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Concrete → pictorial → abstract
Start with something you can hold, move to a drawing of it, then to the symbol. Singapore's mathematics syllabus calls this the model method and it is where the bar model comes from.
What it is not: Not manipulatives forever. The concrete stage is a bridge; leaving a learner on it is a failure too.
At every level
- PreK 2:
- Count real objects before counting pictures of objects before counting numerals.
- grades 3 5:
- Bar models for every word problem, before any equation is written.
- grades 6 8:
- Bars become variables. The drawing survives the transition; that is the point.
- grades 9 12:
- Return to a concrete case whenever an abstraction stops making sense.
- adult:
- A budget you actually allocate before a lecture about appropriations.
WWC guide Recommendation 4 (connect and integrate abstract and concrete representations); the CPA/bar-model lineage is documented in the mathematics-education literature. check it
REPLICATEDMultiple independent studies or a meta-analysis, same direction.
Explicit instruction (Rosenshine's principles)
Small steps, model each one, ask many questions, check every student's understanding, guide the first practice, then release. Ten principles drawn from what effective teachers, cognitive science, and cognitive-support research independently converged on.
What it is not: Not lecturing. Rosenshine's teachers ask more questions and check more often than anyone else — that is the finding.
At every level
- PreK 2:
- One idea per lesson, modeled out loud, practiced immediately.
- grades 3 5:
- Daily review, new material in small chunks, high question rate.
- grades 6 8:
- Model the thinking, not just the answer. Think aloud where students can hear the doubt.
- grades 9 12:
- Guided practice until the success rate is high, then independent work.
- adult:
- Every Academy unit states the objective, teaches in small pieces, and checks before moving.
Barak Rosenshine, 'Principles of Instruction', American Educator, Spring 2012 — a synthesis of three independent research bodies that agreed. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Mastery before moving on
Do not advance on a calendar. Advance on demonstrated understanding, with corrective work in between rather than a low grade and a new topic.
What it is not: Not holding a child back a year. It is holding a topic open until it is learned.
At every level
- PreK 2:
- A letter-sound is not done because the week is over.
- grades 3 5:
- Reteach and retest is normal, not remedial.
- grades 6 8:
- Prerequisite check before a unit that depends on one.
- grades 9 12:
- Revision to standard is the assignment, not extra credit.
- adult:
- Arcade mastery is bronze/silver/gold on a clean run — 'I finished it' and 'I own it' look different.
The mastery-learning literature descends from Bloom's work on group instruction with feedback and correctives; effects are consistently positive though smaller than one-to-one tutoring. check it
REPLICATEDMultiple independent studies or a meta-analysis, same direction.
Checking for understanding, and feedback that tells you what to do next
Frequent low-stakes checks that change what the teacher does next. Feedback that describes the gap and the next action beats a score every time.
What it is not: Not grading. A number tells a learner where they rank; it does not tell them what to do Monday.
At every level
- PreK 2:
- Thumbs up, show me, say it back.
- grades 3 5:
- Exit tickets read before the next lesson is planned.
- grades 6 8:
- Feedback in the form of one specific next move.
- grades 9 12:
- Comment-only marking on drafts; grades at the end, not throughout.
- adult:
- Games say what the right answer was and why, immediately, with no score kept anywhere public.
One of the most consistently supported findings in classroom research; it is also Rosenshine's principles 4, 6, and 8. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Structured literacy — the five pillars
Phonemic awareness, phonics, fluency, vocabulary, and comprehension, taught explicitly and in sequence rather than hoped for through exposure.
What it is not: Not phonics alone. The panel's finding was that phonics has to be integrated with the other four.
At every level
- PreK 2:
- This is the whole job in these grades. Sound, letter, blend, read, understand.
- grades 3 5:
- Morphology and fluency continue; comprehension leans on knowledge.
- grades 6 8:
- Explicit vocabulary and academic language; decoding support for anyone still missing it.
- grades 9 12:
- Discipline-specific reading — a historian and a scientist read differently.
- adult:
- Adult learners who struggle to decode are taught to decode, plainly and without shame.
National Reading Panel (NICHD, 2000) identified these five components; systematic phonics must be integrated with the other four, not taught alone. WWC's K–3 foundational-skills guide is the practice translation. check it
PANELA government or expert panel synthesized the literature and issued a recommendation.
Comprehension runs on background knowledge
Reading comprehension is not a transferable skill you can drill in the abstract. What you already know about the topic predicts how much you understand. So the curriculum has to deliberately build wide knowledge.
What it is not: Not trivia. Knowledge is organized and sequenced so later learning has something to attach to.
At every level
- PreK 2:
- Read aloud far above the child's own decoding level. Talk about the world.
- grades 3 5:
- Content-rich units — history, science, art — as the reading material itself.
- grades 6 8:
- Sequenced knowledge so later texts assume earlier ones.
- grades 9 12:
- Primary sources, because a document is a knowledge object and a reading task at once.
- adult:
- The Founders' Library and Founder Files exist for exactly this.
WWC practice guide 'Improving Reading Comprehension in Kindergarten Through 3rd Grade' pairs comprehension strategy instruction with knowledge and language development. check it
REPLICATEDMultiple independent studies or a meta-analysis, same direction.
High-dosage tutoring
One-to-one or very small group, several times a week, during the school day, by a consistent tutor. Among the largest reliable effects in education research.
What it is not: Not homework help once a month, and not an app. The effect comes from dosage, consistency, and a person.
At every level
- PreK 2:
- Highest payoff here. Small group, daily, on foundational reading.
- grades 3 5:
- Same, plus arithmetic fluency.
- grades 6 8:
- Prerequisite repair before the gap compounds.
- grades 9 12:
- Course-embedded, on the actual coursework.
- adult:
- Outpost cannot provide a human tutor. It says so rather than implying software is one.
Nickow, Oreopoulos & Quan, meta-analysis of PreK-12 tutoring RCTs across roughly forty years: consistently substantial positive effects, larger when tutors are teachers or paraprofessionals, in earlier grades, at least three days a week, and during school. check it
PRACTITIONERWidely used and documented, not established by controlled trial.
Multiple ways in, for everyone
Offer more than one representation of the same content — text, audio, image, worked example, hands-on — to ALL learners, and let each person use what helps. Adjustable text, captions, contrast, and pacing are part of this, not an add-on.
What it is not: Not lowering the standard for anyone. The destination is fixed; the road is not.
At every level
- PreK 2:
- Hear it, see it, do it — all three, for every child.
- grades 3 5:
- Read it or listen to it; both count as reading the assignment.
- grades 6 8:
- Choice of how to show what you know, not choice of whether to know it.
- grades 9 12:
- Same standard, multiple routes to demonstrating it.
- adult:
- Adjustable text size, high contrast, screen-reader labels, and three-channel feedback are already shipped.
This is the defensible version of 'different learners are different.' The support for multiple representations is real (dual coding, accessibility law and practice). What is NOT supported is sorting learners into styles and matching instruction to the label — see learning-styles-meshing below. check it
PRACTITIONERWidely used and documented, not established by controlled trial.
Lesson study — the improvement loop
Teachers plan one lesson together, one teaches it while others watch the students, then they revise it and do it again. The unit of improvement is the lesson, not the teacher.
What it is not: Not teacher evaluation. Watching a colleague to grade them destroys the thing that makes it work.
At every level
- PreK 2:
- Same loop, applied to a read-aloud.
- grades 3 5:
- Watch the students, not the teacher. That is the discipline.
- grades 6 8:
- Revise the lesson, keep the notes, run it again next year.
- grades 9 12:
- Departmental, on the lesson everyone finds hardest to teach.
- adult:
- Outpost's version: every unit carries a revision history and the gate re-runs on every change.
Long-standing Japanese professional-practice tradition (jugyou kenkyuu), extensively documented in the education literature. Adopted widely outside Japan; outcome evidence at scale is thinner than its documentation. check it
