learn-course-design-with-phoebe / Designer session 3 of 6
Learn Course Design with Phoebe · Designer track, session 3 of 6

The arc and cognitive load

When a room goes quiet and stops asking questions, the instinct is to reach for engagement - more energy, a story, an icebreaker. Usually the room is not bored, it is full. Tonight: how little anybody can hold at once, why a worked example teaches a beginner more than letting them struggle, when to take the support away, and the sequencing decision that makes the difference between a session that builds and one that just accumulates.

🟡 Medium Anyone whose session loses the room in the second half Bring your a2 content list 45 min
0-3 · Welcome 3-18 · Load, worked examples, fading 18-40 · Build-along: sequence the session 40-45 · Q&A
Part 0

Where we are

You have an objective, a proof and a short content list. Tonight decides the order and the support, which is what determines whether the content is learnable in the time you have.

🎫 Clear all three quiz questions to stamp your passport Live = presented in the 45 minutes Self-study = the full depth, after class

★ What you walk out with tonight. Your session sequenced with load in mind, two worked examples written, a fading plan, and a diagnosis for the moment your room usually goes quiet.

LiveThe reframe this session is built on2 min▶

Working memory - the part of your attention holding things right now - is small and does not get bigger with motivation. An expert can hold a schema change as one thing. A novice is holding the column name, the type, the consumer list, what "breaking" means and what you just said about notice periods, all separately, and there is no room left to think.

That produces the single most useful diagnostic question in course design: when your room goes quiet, ask whether they are bored or full. They look identical from the front and they need opposite responses. Bored needs more stakes. Full needs less content and more time.

Almost every time somebody adds energy to a full room, it gets worse.

Part 1 · covers load

Three kinds of load, and only one is worth having

6 min live
Wasted load caused by how you presented it A slide they read while you talk over it A diagram and its key on separate pages Remove all of it Necessary load the subject is genuinely hard Six change types, and which of them break You cannot remove this, only sequence it Break it into pieces Useful load the effort that builds the skill Actually classifying a change and being wrong This is the part that feels hard and works Protect the time for it The design job: delete the first, sequence the second, and spend everything you saved on the third. Adding content takes minutes from the third column, which is the only one that produces an ability.
🔍 Click to zoom - most "engagement" fixes are attempts to solve a first-column problem with energy
LiveFour presentation habits that waste load for free4 min▶
  • Talking over text they are reading. They cannot do both. Either the slide has words and you are quiet, or you are talking and the slide has a picture. This one is free to fix and it is everywhere.
  • Splitting a thing from its label. A diagram with the key underneath, code with the explanation on the next slide, a table whose column meanings were said out loud two minutes ago. Put the explanation where the thing is.
  • Saying it and also writing it and also having them read it. Redundancy feels thorough and is a load cost, because they process all three before discovering they were the same.
  • Introducing vocabulary before the thing it names. Show the schema change, let them see the problem, then name it. The reverse asks them to hold an empty label for two minutes.

None of these are about being a better presenter. They are design defects that can be fixed in the file, at your desk, with nobody watching - which makes them the cheapest improvements available in this entire course.

Part 2 · covers worked examples

Show them a solved one, then take the support away

6 min live

The instinct is that people learn by struggling. Beginners mostly do not - they learn by studying a complete solution and then trying one. The struggle becomes valuable later, and knowing when to switch is the actual skill.

1 · Fully worked You classify a change out loud, showing every step you take they study, not watch 2 · Half done You do the first two steps, they finish it the load stays low 3 · Prompted They do it all, with the checklist visible support is now external 4 · Alone Checklist away. This is the proof now the struggle helps The failure at each end Start at 4 and beginners flounder, guess, and learn the wrong lesson from being wrong. Stay at 1 too long and competent people are bored, and they stop paying attention before the part that would have taught them something. Fading is the whole technique.
🔍 Click to zoom - four rungs, and most sessions use exactly one of them
Self-studySequencing decisions worth making on purpose4 min▶
  • Whole task before parts. Show the complete thing first, badly and quickly, so the parts have somewhere to attach. "We will cover the theory and it will make sense later" asks people to hold unattached pieces for forty minutes.
  • Simple to complex, on real cases. Not simplified toy cases - genuinely simple real ones. A simplified case teaches a method that does not survive the real thing.
  • Segment it. Learner-paced chunks with a stopping point beat one continuous flow, because a continuous flow gives nobody a moment to consolidate and everybody a moment to fall behind permanently.
  • Put the hardest thing about 60 percent through. Not at the start, when they have no hooks, and not at the end, when they are tired and there is no time to recover from getting it wrong.
Daybreak, sequenced

One incident (whole task, 5 min) → the six change types with two worked examples (10 min) → three changes half-done together (8 min) → ten changes alone, checklist visible (17 min) → two ambiguous ones argued out (10 min, and this is the hardest part, at 60 percent through) → the job aid (5 min).

Build-along

Sequence yours, and write two worked examples

★ 22 min · everyone builds

Whole task first. What is the fastest possible version of the complete thing? Five minutes, done badly, so the parts have somewhere to attach.

Load audit. Go through your materials for the four wasted-load habits. Talking over text, split labels, triple redundancy, vocabulary before the thing. Fix them in the file.

Two worked examples. Not one. One looks like a special case; two lets them see what varies and what does not, which is where the pattern actually comes from.

Fade. Worked, half-done, prompted, alone. Write which of your activities is at which rung, and check you are not using the same rung four times.

Place the hard part. About 60 percent through. Then check what happens if it goes badly - is there time left to recover, or does the session end on a failure?

Write your whole-task opener. Five minutes, complete, deliberately rough. For Daybreak it is one incident: the rename, the silent load, the wrong dashboard at 08:40.

Do the load audit on real materials. Open the last deck you used and count the four habits. Most decks have all four, and every fix is free.

Write worked example one, out loud. Narrate every step including the ones you would normally skip because they are obvious to you. Those are the steps beginners are missing.

Write worked example two, deliberately different. Different in the surface features, same in the method. That contrast is what turns two examples into a pattern.

Label each activity with its rung. If everything is at rung 4, beginners will flounder. If everything is at rung 1, competent people will leave. Most sessions in the wild are all one rung.

Place your hardest moment and check the recovery. If somebody gets it wrong at 60 percent, they have time to try again. At 95 percent, they go home having failed, which is the last thing they remember.

Ask the diagnostic out loud once. Next time a room goes quiet: bored or full? Then respond accordingly, rather than reaching for energy by reflex.

If you take one thing from tonight. Two worked examples, not one. It is twenty minutes of preparation, it costs nothing in the room, and it is the difference between a group that has seen a method and a group that can apply it.

Homework

Before session a4

25 min
  1. Load-audit one real deck. Count the four habits and fix them. Free improvement, no room required.
  2. Write two worked examples for the hardest thing in your session, narrating the obvious steps.
  3. Label your activities with their rung. Notice if they are all the same one.
  4. Find a course you attended that started with theory. Ask what the whole-task version would have been, and how much of the first twenty minutes it would have replaced.
  5. Bring your sequence to a4, which is about making the practice actually produce the behaviour rather than exposure to it.
Check yourself

Three questions

Clear all three and this session gets stamped in your passport.

1 · Your room goes quiet halfway through. What should you check first?

Bored needs more stakes; full needs less content and more time. Adding energy to a full room makes it worse, and that is the default reflex - which is why the diagnostic question is worth having ready before you need it.

2 · Why give beginners a fully worked example instead of letting them try first?

Struggle becomes valuable once there is something to struggle with. Before that it produces guessing and the wrong lesson from being wrong. The skill is knowing when to fade the support, which is why there are four rungs rather than a rule.

3 · Why two worked examples rather than one?

A single example is indistinguishable from an anecdote. The contrast between two - different surface, same method - is what makes the method visible as a method, and it costs twenty minutes of preparation rather than any time in the room.

Sources covered

What this session teaches, and where it came from

✓Working memory as the binding constraintThree kinds of load - wasted, necessary, useful - with what to do about each, and the bored-or-full diagnostic
✓Four wasted-load habitsTalking over text, split labels, triple redundancy, vocabulary before the thing. All fixable in the file at your desk
✓Worked examples and fadingFour rungs from fully worked to alone, with the failure at each end named
✓Two examples, not oneThe contrast is where the pattern becomes visible
✓SequencingWhole task before parts, simple real cases over simplified ones, segmenting, and the hardest thing at 60 percent
◐Chunking, schema formation, expertise reversalNamed. The cognitive-science literature is pointed at rather than taught, and this session stays at what to do on Tuesday

Every fact on this page, and its verification tier, is recorded in the course's source map.

Designer a3 cheat sheet · pin this

Bored or full?They look identical from the front and need opposite responses.
Three loadsDelete the wasted, sequence the necessary, protect time for the useful.
Four free fixesDo not talk over text, keep labels with things, cut redundancy, name it after showing it.
Four rungsWorked · half-done · prompted · alone. Most sessions use exactly one.
Two examplesOne is an anecdote. Two is a pattern.
Hard part at 60 percentLate enough to have hooks, early enough to recover from failing.