The multi-store model (Atkinson-Shiffrin) proposes that memory works as three connected stores, sensory memory, short-term memory, and long-term memory, and that information must pass through each stage via automatic or effortful processing to be remembered long term.
The multi-store model is the classic 'assembly line' picture of memory. Information from your senses enters sensory memory first, a huge but extremely brief store that holds raw sights and sounds for a second or less. If you pay attention to it, the information moves into short-term memory, a small workspace that holds a limited amount of info for around 20-30 seconds unless you rehearse it. With enough rehearsal or meaningful processing, it gets encoded into long-term memory, which has essentially unlimited capacity and duration.
The model also distinguishes how information gets moved along. Automatic processing handles things you encode without trying (like where you sat in class today), while effortful processing requires attention and work (like memorizing vocab for this exam). The big idea is that memory isn't one thing. It's a sequence of separate stores with different capacities and durations, and information has to survive each stage to stick.
The multi-store model lives in Topic 2.3 (Introduction to Memory) in Unit 2: Cognition, supporting learning objective 2.3.A: explain how the types, structures, and processes of memory work. It's the structural backbone for everything else in the memory topics. Explicit vs. implicit memory, encoding strategies, working memory components, and forgetting all get explained in terms of which store is involved and what happens at each transfer point. If you can't say where information is in the system, you can't explain why it was remembered or lost, and that's exactly the kind of explanation Unit 2 questions ask for.
Keep studying AP® Psychology Unit 2
Levels of Processing Model (Unit 2)
This is the model's main rival in Topic 2.3. The multi-store model says memory depends on which store information reaches; levels of processing says it depends on how deeply you think about the information's meaning. AP questions love making you tell these two apart.
Phonological Loop & Visuospatial Sketchpad (Unit 2)
The working memory model is basically an upgrade to the multi-store model's 'short-term memory' box. Instead of one simple store, working memory splits it into a phonological loop for sounds and a visuospatial sketchpad for images, which explains why you can hum a song and read a map at the same time.
Explicit, Implicit, and Procedural Memory (Unit 2)
The multi-store model tells you how information gets INTO long-term memory; the explicit/implicit distinction tells you what long-term memory actually contains. Episodic, semantic, and procedural memories are all subdivisions of the model's final store.
Long-Term Potentiation (Unit 2)
LTP is the biological version of the model's last step. When information transfers into long-term memory, synaptic connections between neurons strengthen, so LTP is the neural mechanism behind the box-and-arrow diagram.
Multiple-choice questions usually test the model in two ways. First, scenario questions ask you to identify which store is involved or where the breakdown happened, like a stem asking which scenario 'best illustrates the limitations of sensory memory.' Second, questions probe automatic vs. effortful processing, such as asking which process would be LEAST affected by divided attention during initial exposure (automatic processing is the answer logic there, since it doesn't require attention). Research-based questions also pit the multi-store model against levels of processing, describing groups who encode words by appearance, sound, or meaning and asking you to predict recall. No released FRQ has used the term verbatim, but the Article Analysis and Evidence-Based questions on the revised exam often involve memory studies, and you'll need this model's vocabulary (sensory, short-term, long-term, rehearsal, encoding) to explain results accurately.
The multi-store model says memory is about location, meaning information moves through three separate stores and rehearsal pushes it forward. The levels of processing model says memory is about depth, meaning shallow encoding (how a word looks) produces weak memories while deep encoding (what a word means) produces strong ones, regardless of any 'stores.' If a question describes structures and stages, it's multi-store. If it describes how meaningfully something was encoded, it's levels of processing.
The multi-store model describes three connected memory systems, sensory memory, short-term memory, and long-term memory, that information must pass through to be remembered.
Sensory memory holds huge amounts of raw input for under a second, short-term memory holds a small amount for about 20-30 seconds, and long-term memory is essentially unlimited.
Attention moves information from sensory to short-term memory, and rehearsal or effortful processing moves it from short-term to long-term memory.
Automatic processing encodes information without attention, while effortful processing requires deliberate work, so dividing attention hurts effortful encoding much more than automatic encoding.
The working memory model (with the phonological loop and visuospatial sketchpad) is a more detailed replacement for the model's single short-term memory store.
On the exam, distinguish the multi-store model (memory depends on which store) from the levels of processing model (memory depends on depth of encoding).
It's the model proposing that memory consists of three interacting stores, sensory memory, short-term memory, and long-term memory, and that information must pass through each stage to be permanently remembered. It falls under Topic 2.3 and learning objective 2.3.A in Unit 2.
The multi-store model explains memory through structures (three stores plus rehearsal moving information between them), while levels of processing explains memory through encoding depth, where focusing on meaning beats focusing on appearance or sound. Exam scenarios with groups encoding words by looks, rhymes, or meaning are testing levels of processing, not the multi-store model.
No. Working memory is a later, more detailed take on the multi-store model's short-term memory store. It breaks that single box into components like the phonological loop (sound) and visuospatial sketchpad (images), explaining how you juggle different types of information at once.
Not always, and that's a real limitation of the model. Automatic processing encodes some information (like time, space, and frequency) into long-term memory without rehearsal or even attention, which is why an exam question can ask which process is least affected by divided attention.
Sensory memory is large in capacity but lasts under a second, short-term memory holds a limited amount (classically about 7 items) for roughly 20-30 seconds, and long-term memory has essentially unlimited capacity and duration.
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.