Pedagogy Reliability Memory Cognitive Load And Robust Effects

---
name: pedagogy-reliability-memory-cognitive-load-and-robust-effects
description: "Use when you need what is actually well established about learning: the reliability gradient separating replicated findings from failed ones and from claims that were only ever marketed, memory architecture and working memory limits, cognitive load theory and the intrinsic, extraneous and germane distinction, the robust effects of spacing, retrieval practice and interleaving, prior knowledge, schemas and expertise, motivation, and metacognition and the illusion of fluency. Includes the router for the whole pedagogy reference."
---

# Pedagogy and Learning Science: The Reliability Gradient, Memory Architecture, Cognitive Load, the Robust Effects, Prior Knowledge, Motivation, and Metacognition

> **Part 1 of 5** of the *Pedagogy and the Study of Learning and Teaching* reference (plugin `pedagogy-and-the-study-of-learning`), covering §0–§7. Sibling skills: `pedagogy-explicit-instruction-worked-examples-and-feedback` (§8–§13), `pedagogy-subject-instruction-curriculum-and-classroom-practice` (§14–§21), `pedagogy-reading-research-and-zombie-ideas` (§22–§24), `pedagogy-reference` (§25–§30). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.
>
> **Currency:** The cognitive science is stable and well-replicated. Two areas moved. See §25 → `pedagogy-reference` for AI in education, and post-pandemic achievement data.

> **⚠️ This field has an unusually wide gap between what is well-established and what is
> widely believed and practised.** ⚠️ **Some findings here are among the most robust in
> all of psychology; others in the same textbooks failed replication entirely; and a third
> group were never research findings at all — they were marketed.**
>
> **Complements a psychology reference (§1's replication context) and a family/community
> reference (behaviour genetics, which constrains how much any intervention can do).**
>
> **⚠️ GOTCHA** boxes mark widely-taught ideas that the evidence contradicts.
>
> **⚠️ A disclosure about the political layer:** ⚠️ **education research is unusually
> ideologically loaded, because it touches how children are raised and how resources are
> allocated.** **⚠️ Where a genuine scholarly dispute exists (§8 → `pedagogy-explicit-instruction-worked-examples-and-feedback`, §14 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`) I set out both
> positions; where the evidence is one-sided I say so and show the evidence.**
>
> **The three ideas that organize this document:**
> 1. **⚠️ Learning is a change in LONG-TERM MEMORY; performance during a lesson is not
>    learning** (§2, §7). **The two come apart routinely, and confusing them is the single
>    most consequential error in teaching.**
> 2. **⚠️ Working memory is tiny and prior knowledge is the great multiplier** (§2, §3).
>    **Almost every instructional design principle that works is managing that limit.**
> 3. **⚠️ Things that FEEL like good learning usually aren't** (§4, §7). **Fluency,
>    familiarity and ease are systematically misleading — for students AND teachers.**

---

## §0. Routing

| You want... | Go to |
|---|---|
| **⚠️ What's actually established** | **§1** |
| Memory architecture | §2 |
| **⚠️ Cognitive load** | **§3** |
| **⚠️ The robust effects** | **§4** |
| Prior knowledge and expertise | §5 |
| Motivation | §6 |
| **⚠️ Metacognition and fluency** | **§7** |
| **⚠️ Explicit vs discovery** | **§8 → `pedagogy-explicit-instruction-worked-examples-and-feedback`** |
| Worked examples | §9 → `pedagogy-explicit-instruction-worked-examples-and-feedback` |
| Explaining and modelling | §10 → `pedagogy-explicit-instruction-worked-examples-and-feedback` |
| Questioning | §11 → `pedagogy-explicit-instruction-worked-examples-and-feedback` |
| **Feedback** | **§12 → `pedagogy-explicit-instruction-worked-examples-and-feedback`** |
| Assessment | §13 → `pedagogy-explicit-instruction-worked-examples-and-feedback` |
| **⚠️ Reading instruction** | **§14 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`** |
| Mathematics | §15 → `pedagogy-subject-instruction-curriculum-and-classroom-practice` |
| Curriculum and sequencing | §16 → `pedagogy-subject-instruction-curriculum-and-classroom-practice` |
| Classroom practice | §17 → `pedagogy-subject-instruction-curriculum-and-classroom-practice` |
| **⚠️ Differentiation** | **§18 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`** |
| Higher and adult education | §19–§20 → `pedagogy-subject-instruction-curriculum-and-classroom-practice` |
| Educational technology | §21 → `pedagogy-subject-instruction-curriculum-and-classroom-practice` |
| **⚠️ Reading the research** | **§22 → `pedagogy-reading-research-and-zombie-ideas`** |
| **⚠️ Zombie ideas** | **§23 → `pedagogy-reading-research-and-zombie-ideas`** |
| Teacher expertise | §24 → `pedagogy-reading-research-and-zombie-ideas` |
| **What's live** | **§25 → `pedagogy-reference`** |
| Misconceptions, numbers | §26–§27 → `pedagogy-reference` |
| Books, quick ref, method | §28–§30 → `pedagogy-reference` |

---

## §1. ⚠️ The Reliability Gradient

```
⚠️ ROBUST — large, replicated, mechanistically understood
   ⚠️ Spacing/distributed practice · ⚠️ retrieval practice (testing
   effect) · working memory limits · ⚠️ the expertise reversal effect ·
   worked example effect · ⚠️ prior knowledge as the dominant predictor
   of new learning · ⚠️ phonics for early decoding (§14)

⚠️ REASONABLY SUPPORTED — real but context-dependent, smaller than claimed
   Interleaving · elaborative interrogation · self-explanation ·
   formative assessment · feedback (⚠️ highly variable, §12) ·
   explicit instruction for novices (§8)

⚠️ CONTESTED — genuine scholarly disagreement
   ⚠️ The role of inquiry for intermediate learners · how much
   knowledge transfers · grit and mindset effect sizes ·
   ⚠️ optimal balance of explicit and constructivist approaches

⚠️ FAILED OR NEVER SUPPORTED
   ⚠️ LEARNING STYLES (matching instruction to a "style") ·
   ⚠️ the Learning Pyramid / "we remember 10% of what we read" ·
   ⚠️ left-brain/right-brain teaching · ⚠️ Brain Gym ·
   ⚠️ multiple intelligences as an instructional prescription ·
   ⚠️ "digital natives" · ⚠️ teaching 21st-century skills
   content-free (§23)
```
> **⚠️ GOTCHA — why bad ideas persist here specifically.** ⚠️ **Education has weak
> gatekeeping between research and practice, a large commercial training industry, and
> ideas that FEEL true and flatter both teachers and students.** **⚠️ Learning styles
> survives despite comprehensive refutation because it is intuitive, kind-sounding, and
> sells professional development.** **⚠️ Treat "research shows" in an education context as
> a claim to be checked, not a citation** (§22 → `pedagogy-reading-research-and-zombie-ideas`).

**⚠️ The replication context**: ⚠️ **education research inherits psychology's replication
problems and adds its own — small samples, unblinded outcomes, researcher-developed
measures, publication bias, and interventions delivered by their own inventors.**
**⚠️ The Bakker/EEF-style finding is worth holding: effects shrink dramatically when moved
from small researcher-run trials to large independent ones** (§22 → `pedagogy-reading-research-and-zombie-ideas`).

---

# PART I — HOW LEARNING WORKS

## §2. Memory Architecture

```
⚠️ SENSORY → WORKING MEMORY → LONG-TERM MEMORY
⚠️ WORKING MEMORY  ⚠️ tiny (the classic "7±2" is now generally
   revised DOWNWARD toward ~4 chunks for novel material) and brief
   without rehearsal. ⚠️ THE bottleneck in all instruction
⚠️ LONG-TERM MEMORY  effectively unlimited; ⚠️ the CONSTRAINT is
   RETRIEVAL, not storage
⚠️ CHUNKING  ⚠️ the escape hatch: expertise turns many items into one.
   A chess master sees a position, not 25 pieces
⚠️ SCHEMAS  organized knowledge structures in long-term memory that
   working memory can treat as SINGLE items (§5)
```
> **⚠️ GOTCHA — "learning is a change in long-term memory," and PERFORMANCE IS NOT
> LEARNING.** ⚠️ **A student can perform perfectly during a lesson and retain nothing;
> conditions that depress performance during practice often IMPROVE long-term retention
> (§4's desirable difficulties).** **⚠️ This is Soderstrom and Bjork's central point and it
> should reorganize how anyone judges a lesson.** **A lesson that looks smooth and
> successful may be teaching nothing durable.**

**⚠️ Forgetting is not simply decay** — ⚠️ **the storage/retrieval distinction matters:
things become inaccessible rather than erased**, **which is why relearning is faster than
learning and why cues matter so much.**

---

## §3. ⚠️ Cognitive Load Theory

**⚠️ The most practically useful theory in instructional design, and it follows directly
from §2.**
```
⚠️ INTRINSIC LOAD    inherent difficulty and ELEMENT INTERACTIVITY —
   how many things must be held simultaneously. ⚠️ Reduce by
   SEQUENCING and by building prior knowledge, not by dumbing down
⚠️ EXTRANEOUS LOAD   imposed by poor design. ⚠️ ELIMINATE THIS
⚠️ GERMANE LOAD      effort that builds schemas. ⚠️ Note this construct
   is contested and has been reformulated by CLT researchers themselves
```
**⚠️ The design effects that follow, and they're all testable:**
```
⚠️ SPLIT-ATTENTION  ⚠️ don't force learners to integrate separated
   sources (diagram here, labels there). INTEGRATE them physically
⚠️ REDUNDANCY  ⚠️ identical information in two forms HURTS —
   reading slides aloud verbatim is worse than either alone
⚠️ MODALITY  narration + visual beats text + visual (uses both channels)
⚠️ WORKED EXAMPLE EFFECT  (§9)
⚠️ EXPERTISE REVERSAL  ⚠️ THE crucial one: support that HELPS novices
   HARMS experts. There is no universally optimal method — only
   optimal-for-this-learner-at-this-point
```
**⚠️ Expertise reversal deserves emphasis because it dissolves most method wars** (§8 → `pedagogy-explicit-instruction-worked-examples-and-feedback`):
⚠️ **the question "is explicit instruction or discovery better?" is malformed.** **The
answer depends on the learner's prior knowledge, and it CHANGES as they learn.**

---

## §4. ⚠️ The Robust Effects

> **⚠️ If you take one thing from this document, take this section. These are among the
> best-replicated findings in cognitive psychology and are massively underused.**
```
⚠️ SPACING / DISTRIBUTED PRACTICE
   ⚠️ The same total study time spread out beats massed practice,
   substantially, for long-term retention. ⚠️ Massing produces better
   IMMEDIATE performance — which is exactly why students choose it
   ⚠️ Optimal gap scales with the retention interval you want

⚠️ RETRIEVAL PRACTICE (the TESTING EFFECT)
   ⚠️ RETRIEVING information strengthens memory far more than
   RE-READING it. ⚠️ The act of pulling it out IS the learning event
   ⚠️ Works even when retrieval fails, if feedback follows
   ⚠️ Low-stakes and self-testing count — this is not about exams

⚠️ INTERLEAVING
   ⚠️ Mixing problem TYPES beats blocking them. ⚠️ Blocked practice
   lets you apply a method without SELECTING it — and selection is
   usually the actual skill being tested later
   ⚠️ Feels worse and produces worse practice performance. Reliably

⚠️ ELABORATION and SELF-EXPLANATION
   ⚠️ Asking "why is this so?" and explaining to oneself

⚠️ DUAL CODING
   ⚠️ Words plus relevant visuals. ⚠️ NOTE: this is NOT learning styles
   — it applies to EVERYONE, not to "visual learners" (§23)
```
**⚠️ DESIRABLE DIFFICULTIES is the unifying frame** (Bjork): ⚠️ **conditions that slow
acquisition and feel harder frequently improve long-term retention and transfer.**
**⚠️ The corollary is uncomfortable: student preference and student satisfaction are
NEGATIVELY correlated with some of the most effective methods.**

---

## §5. Prior Knowledge, Schemas and Expertise

**⚠️ Prior knowledge is the single strongest predictor of new learning, and this has an
enormous curricular implication** (§16 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`).
**⚠️ The Matthew effect**: ⚠️ **knowledge compounds — those who know more learn more
easily, so gaps widen without intervention.**
> **⚠️ GOTCHA — the "baseball study" is the one to know, because it reframes what reading
> comprehension IS.** ⚠️ **Poor readers who knew baseball outperformed good readers who
> didn't on a passage about baseball.** **⚠️ Comprehension is not a transferable skill
> applied to any text — it is heavily dependent on DOMAIN KNOWLEDGE** (§14 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`, §16 → `pedagogy-subject-instruction-curriculum-and-classroom-practice`).
> **⚠️ Which means "teaching comprehension strategies" has sharply limited returns compared
> to building knowledge, and this is one of the more consequential findings in the field.**

**⚠️ Novice-expert differences**: ⚠️ **experts chunk, perceive deep structure rather than
surface features, and have automated basics — freeing working memory.** **⚠️ Novices sort
physics problems by surface features (inclined plane); experts sort by principle
(conservation of energy).**
**⚠️ The instructional consequence**: ⚠️ **you cannot teach "thinking like an expert" as a
generic skill, because expert thinking is largely a product of domain knowledge
organization.**

---

## §6. Motivation

```
⚠️ SELF-DETERMINATION THEORY  autonomy, competence, relatedness.
   ⚠️ Well-supported as a framework; ⚠️ intervention effects are more modest
EXPECTANCY-VALUE  ⚠️ effort follows expectation of success × value
⚠️ INTEREST  situational (triggered) → individual (durable).
   ⚠️ Interest FOLLOWS competence at least as often as it precedes it —
   which inverts the usual "make it interesting first" advice
⚠️ ATTRIBUTION  ability vs effort vs strategy. ⚠️ Attributing failure
   to STRATEGY is more useful than either
GOAL ORIENTATION  mastery vs performance
```
> **⚠️ GOTCHA — growth mindset is a genuine and contested case, and worth stating
> carefully.** ⚠️ **The theory is plausible and the original claims were large.** **⚠️ Large
> replications and meta-analyses found effects far smaller than the popular presentation —
> often near zero on average, with some evidence of small effects concentrated in
> low-achieving or at-risk students.** **⚠️ Meanwhile, the classroom version frequently
> degenerated into praising effort regardless of outcome, which Dweck herself has
> criticized as a distortion.**
> **⚠️ The honest statement: not nothing, much smaller than sold, and NOT a substitute for
> teaching.** **The same shape applies to grit.**

---

## §7. ⚠️ Metacognition and the Illusion of Fluency

**⚠️ Learners are systematically bad at judging their own learning, and this explains why
they choose ineffective methods.**
```
⚠️ THE ILLUSION OF FLUENCY  ⚠️ re-reading and highlighting FEEL
   productive because the material becomes familiar. ⚠️ FAMILIARITY IS
   NOT RETRIEVABILITY, and students mistake one for the other
⚠️ JUDGEMENTS OF LEARNING  ⚠️ made immediately after study are
   badly miscalibrated; delayed judgements are far more accurate
⚠️ SURVEY EVIDENCE  ⚠️ students overwhelmingly report re-reading and
   highlighting as their main strategies — among the LEAST effective
⚠️ THE TEACHER VERSION  ⚠️ smooth lessons where everyone succeeds feel
   like good teaching (§2). They may indicate the work was too easy
```
**⚠️ Dunlosky's utility ratings are worth memorizing**: ⚠️ **HIGH utility — practice
testing, distributed practice.** ⚠️ **LOW utility — highlighting, re-reading, summarization
(as typically done), keyword mnemonics, imagery for text.**
**⚠️ Teaching metacognition works best embedded in a domain** — ⚠️ **generic "learn how to
learn" courses transfer poorly, for §5's reasons.**

---

# PART II — TEACHING