Watch it fade.
Bend a forgetting curve with spaced reviews, push nine items through the memory stores and watch where they die, then retell a story until it is no longer true.
The forgetting curve
FIG. 01 · MEMORY · 3.1.1Each review pulls you back to 100% — and makes the next fall slower. That second part is the whole trick.
Leave it too short and you revise things you already know. Leave it too long and you are learning them again from scratch.
Encoding by meaning (semantic) outlasts encoding by sound or looks. Same material, different starting strength.
Why this is worth marks. Memory (3.1.1) asks you to explain encoding, storage and retrieval, and to evaluate research on forgetting. The curve is the evidence: retention decays fastest immediately after learning, which is why re-reading your notes on the night before an exam feels productive and does almost nothing a fortnight later. It is also, in the plainest possible terms, why Vofti reschedules a question you got right instead of showing it to you again tomorrow.
Push it through the stores
FIG. 02 · STRUCTURES OF MEMORY · 3.1.1Nine items go in. The two models disagree about what the middle box actually is.
In the multi-store model, rehearsal is the only bridge from short-term to long-term memory.
Capacity is about 7 ± 2 chunks, not 7 things. Group them and the same box holds more.
Two jobs for the same subsystem jam each other. Two jobs for different subsystems barely interfere — the dual-task finding the multi-store model cannot explain.
Why this is worth marks. "Structures of memory" wants the stores described by capacity, duration and encoding, and then evaluated. Run the multi-store model with attention off and it behaves sensibly; switch to working memory and give it two verbal tasks at once and you have the dual-task evidence that the single short-term store cannot account for. That contrast — a model that works until one experiment breaks it — is the evaluation an eight-mark question is asking for.
Tell it again
FIG. 03 · MEMORY AS AN ACTIVE PROCESS · 3.1.1
One witness, told four times: straight after, that evening, a week later, a month later.
Same event, same witness, different schema. Drag the retelling slider under each and compare the last version.
Why this is worth marks. "Memory as an active process" is examined through Bartlett's War of the Ghosts and the idea that we store meaning, not recordings. The three processes have names the mark scheme wants: levelling (detail dropped), sharpening (a few details exaggerated) and assimilation (the story bent to fit what the rememberer already believes). Run both schemas to the fourth retelling and you have the eyewitness-testimony evaluation point in one screen: neither witness is lying.