Sensory, Working, and Long-Term Memory
Memory isn't a single system. It's a series of stages that filter, hold, and store information. Sensory input enters the system, gets processed in short-term or working memory, and can then be encoded into long-term storage. Each stage has different rules for capacity, duration, and how forgetting happens.
Understanding these stages matters for Educational Psychology because they directly shape how students learn and retain material. If you know where information tends to get lost, you can design better study strategies and teaching methods.
Sensory and Short-Term Memory
Sensory Memory and Short-Term Memory
Sensory memory is the first stop. It holds raw sensory information for a very brief time after the original stimulus disappears. You can think of it as a buffer that captures everything your senses take in before most of it fades away.
There are two main types:
- Iconic memory stores visual information for roughly 0.5 seconds.
- Echoic memory stores auditory information for about 3–4 seconds (which is why you can "replay" something someone just said, even if you weren't fully paying attention).
Only the information you pay attention to moves forward into short-term memory (STM). STM holds a small amount of information in an active, readily available state. Its key limitations:
- Capacity: roughly 7 ± 2 items (this comes from George Miller's classic research)
- Duration: about 15–30 seconds without rehearsal
If you don't actively do something with the information in STM, it disappears.
Working Memory and Capacity Limitations
Working memory goes beyond simple short-term storage. It's a multi-component system that actively manipulates information and connects it to long-term memory. The most widely used model comes from Baddeley and Hitch, which breaks working memory into several parts:
- Central executive – the "boss" that directs attention and coordinates the other components. It decides what to focus on and how to allocate mental resources.
- Phonological loop – handles speech-based and verbal information. When you repeat a phone number in your head to remember it, that's the phonological loop at work.
- Visuospatial sketchpad – handles visual and spatial information. When you mentally rotate an object or picture a map, you're using this component.
Working memory capacity varies from person to person and depends on task complexity. Tasks like reading comprehension and mental arithmetic place heavy demands on working memory, which is why they feel effortful.
Decay and Forgetting in Short-Term Memory
Information in STM is fragile. There are several explanations for why it gets lost:
- Decay theory proposes that memory traces simply fade over time if they aren't rehearsed. The neural representation weakens and eventually disappears.
- Interference theory proposes that other information disrupts the target memory. Proactive interference is when old information blocks new learning; retroactive interference is when new information pushes out older material.
- Retrieval failure occurs when information is technically still in memory but can't be accessed, often because the right cues aren't present.
The practical takeaway: if you want to move information out of STM before it's lost, you need to either rehearse it or encode it into long-term memory through deeper processing.
Long-Term Memory Types
Episodic and Semantic Memory
Long-term memory isn't one uniform warehouse. It contains distinct subtypes, and the two most important for declarative (conscious) knowledge are episodic and semantic memory.
Episodic memory stores personal experiences tied to a specific time and place. Your memory of your first day of college or a particular birthday party is episodic. These memories have a "mental time travel" quality, where you can re-experience the event.
Semantic memory stores general facts, concepts, and knowledge about the world. Knowing that Paris is the capital of France or that a hammer is used to drive nails is semantic memory. Unlike episodic memory, semantic knowledge is independent of when or where you learned it. It's shared across members of a culture.
The distinction matters in education because new information often enters as episodic memory (you remember the lesson where you learned it) but gradually becomes semantic knowledge (you just know the fact without remembering the specific learning event).
Procedural Memory
Procedural memory is the unconscious memory for skills and how to perform actions. Riding a bicycle, tying shoelaces, and typing on a keyboard are all procedural memories.
Key characteristics:
- Acquired through repetition and practice, not through a single exposure
- Largely automatic and unconscious once learned, which is why people often call it "muscle memory"
- Difficult to verbalize. Try explaining exactly how you balance on a bike. You know how to do it, but putting it into words is surprisingly hard.
Procedural memory is important in education because it underlies skill-based learning, from handwriting to solving math problems fluently.
Memory Models and Processes
Atkinson-Shiffrin Model and Long-Term Memory
The Atkinson-Shiffrin model (1968) is the foundational framework for understanding memory stages. It proposes three sequential stores:
- Sensory memory captures incoming stimuli.
- Short-term memory holds information selected by attention.
- Long-term memory stores information transferred through rehearsal.
The flow works like this: sensory input enters sensory memory, attention moves selected information into STM, and rehearsal (especially elaborative rehearsal) encodes it into LTM.
Long-term memory has two defining features:
- Virtually unlimited capacity – there's no known upper limit to how much LTM can hold.
- Potentially permanent duration – information in LTM can last a lifetime, though accessing it depends on how well it was encoded and organized.
Information in LTM is organized into schemas, which are mental frameworks that help you organize and interpret new information. For example, your "restaurant schema" includes expectations about menus, ordering, and paying. Schemas speed up processing but can also lead to errors when new information doesn't fit the existing framework.
Chunking and Memory Improvement
Since STM has a limited capacity, chunking is one of the most effective strategies for working within that limit. Chunking groups individual pieces of information into larger, meaningful units.
For example, the sequence 2-0-2-5-5-5-1-2-3-4 is ten separate digits, which exceeds typical STM capacity. But chunked as (202) 555-1234, it becomes three manageable units.
Other strategies that improve memory:
- Mnemonic devices aid encoding and retrieval. Acronyms (like NASA) and acrostics (like "Every Good Boy Does Fine" for musical notes E, G, B, D, F) create memorable structures for otherwise arbitrary information.
- Elaborative rehearsal involves linking new information to existing knowledge in LTM. Instead of just repeating a definition over and over (maintenance rehearsal), you connect it to something you already understand. Research consistently shows that elaborative rehearsal produces stronger, longer-lasting memories than simple repetition.
The core principle across all these strategies: the more meaningfully you process information, the better you'll remember it. This idea connects directly to levels of processing theory, which holds that deeper processing leads to more durable memory traces.