Information Processing Theory and Memory
Information Processing Theory
Three memory types
- Sensory register (short-term): brief reception of input
- Working memory: conscious memory where active thinking happens
- Long-term memory: permanent storage
Core claim
- Nothing is learned unless it is in long-term memory
- Automation (doing without thinking) is the ultimate learning
Working memory facts
- Holds only a few unrelated pieces at once (Miller’s classic figure was 7 ± 2; newer estimates are closer to 3-5)
- If filled with low-level skills, no space for higher-order thinking
- Organizing into related chunks expands effective capacity
- Images take less space than words
A teacher who does not understand how memory works teaches into a black box. They deliver content, students pass tests, but the learning fades quickly. Understanding information processing theory shows why this happens and what to do about it.
Three memory types
Information processing theory describes human memory as three connected systems.
Sensory register (short-term memory). The first place information arrives. Sights, sounds, touches, and other inputs hit the sensory register briefly. Most of this information disappears within seconds unless attention pulls it forward.
A student in a classroom hears the teacher’s voice, the air conditioner humming, a classmate whispering, and a distant truck. All of these enter the sensory register. The student’s attention selects the teacher’s voice; the rest fades.
Working memory (conscious memory). What the student is consciously thinking about right now. The teacher’s words, the math problem on the board, the student’s own thoughts about it.
Working memory is where active thinking happens. The student manipulates information here. Solving a problem, comparing ideas, planning what to write. All of these use working memory.
Long-term memory. Permanent storage. Things the student knows even when not consciously thinking about them. Their name, their address, the multiplication table they learned years ago.
Long-term memory has effectively unlimited capacity. The challenge is getting information into it.
The core claim
Information processing theory makes a sharp claim: nothing is learned unless it is in long-term memory.
A student who has information in their working memory has not learned it yet. They are using it temporarily. As soon as they stop attending to it, it fades from working memory and is gone.
A student who has information in their long-term memory has learned it. They can retrieve it later, even years later, when they need it.
The teacher’s job, in this framework, is to move information from working memory to long-term memory. A teacher who delivers content that students hold briefly in working memory and then forget has not really taught.
Automation is the ultimate goal. A skill becomes automatic when the student can do it without thinking about it. A child who learned multiplication years ago can multiply two-digit numbers without conscious effort. The skill is in long-term memory and runs without using working memory space.
A teacher whose students cannot use the content automatically has not finished teaching. The students may know the rules consciously, but the rules are not yet in long-term memory.
Working memory has strict limits
There is a striking research finding: humans can hold only a small number of unrelated pieces of information in working memory at one time. The classic figure from George Miller’s 1956 paper is “seven plus or minus two”, which is where the 5 to 9 number comes from. More recent reviews (Cowan and others) suggest the practical capacity for unrelated items is closer to three to five chunks, especially for children.
The exact number matters less than the underlying truth: the limit is small. Anyone facing 10 unrelated items at once will lose track of some. Children manage fewer than adults. The teaching implications below do not change with the updated number.
Three implications for teaching:
Implication 1: Teachers cannot stuff lessons with unrelated content. A lesson that introduces 12 new vocabulary words at once exceeds working memory. Students cannot hold all 12. They will lose some. The lesson should introduce no more than 7 to 9 unrelated words at once.
Implication 2: Lower-level skills consume the limited space. If a student is using working memory to remember what each math symbol means (a low-level skill they have not automated), they have no space left for the actual math problem (a higher-level skill that requires thinking).
The higher-level skills cannot run if working memory is full of low-level ones. A student who fluently recognizes math symbols (in long-term memory) has working memory free for problem-solving. A student who is still working out what symbols mean (in working memory) cannot also do problem-solving at the same time.
Implication 3: Organizing into related chunks expands capacity. The 5-to-9 limit is for unrelated items. If items are connected by relationships, multiple items can be held as one chunk.
Example: 9 unrelated nouns is at the limit. But if those 9 nouns can be grouped into 3 categories (animals, plants, objects), the working memory now holds 3 items (the categories) with sub-items underneath. Effective capacity grows.
A teacher who organizes content into related chunks gives students more usable space. A teacher who delivers 9 unrelated facts uses up working memory without leaving room for thinking.
A practical example: low-level skills crowding out high-level thinking
Here is an example. A student is solving a multiplication word problem.
If the student does not have multiplication facts in long-term memory:
- They use working memory to figure out 7 × 8 (slow).
- They use working memory to figure out 6 × 4 (slow).
- They use working memory to figure out 7 × 6 (slow).
- By the time they finish the calculations, working memory is exhausted.
- They cannot use working memory to think about the word problem itself.
- They give up or get the wrong answer.
If the student has multiplication facts in long-term memory:
- They retrieve 7 × 8 = 56 automatically (no working memory).
- They retrieve 6 × 4 = 24 automatically (no working memory).
- They retrieve 7 × 6 = 42 automatically (no working memory).
- Working memory is free to think about the word problem.
- They solve the problem.
Both students “know multiplication”. The difference is whether the knowledge is in working memory (slow, exhausting) or long-term memory (automatic, frees up thinking).
The teacher’s job for the multiplication facts is to drill them into long-term memory rather than teach them and move on. Automation is the goal.
Automatic facts free working memory for higher-order thinking
A student calculating 7 × 8 in working memory uses up that limited space.
A student who retrieves 7 × 8 = 56 automatically from long-term memory has working memory free for the word problem itself.
The teacher’s goal is automation: skills running from long-term memory without consuming working memory.
Images vs words
Another working memory fact: images take less space than words. A picture that contains the same information as a long paragraph occupies less working memory.
Example: directions to a place 200 km away. One way to give directions: in writing, with all turns and landmarks named. The receiver gets two minutes to absorb this. The result is usually confusion - too much detail, too many separate items, working memory overwhelmed.
Another way: a small map showing the route. The receiver looks at the map for two minutes. The map’s spatial relationships fit into working memory more efficiently. The receiver understands the route.
Both delivery methods contain the same information. The map uses working memory more efficiently. This is why “a picture is worth a thousand words” has practical truth.
For teaching, this means: when a concept can be presented visually, the visual form often communicates more efficiently. A diagram of the water cycle communicates more efficiently than a paragraph describing it. A flowchart of a procedure communicates more efficiently than a numbered list.
The next two articles cover specific techniques that exploit these memory facts: elaboration techniques (article 2) and visual tools (article 3).
What teachers must do
To work with working memory’s limits and get information into long-term memory, teachers should:
- Limit unrelated content per lesson. Stay within 5-9 unrelated items.
- Organize content into related chunks. Categories, hierarchies, relationships.
- Drill foundational skills to automation. Multiplication facts, vocabulary, common procedures should run from long-term memory.
- Use elaboration, not just maintenance rehearsal. Help students connect new content to existing knowledge.
- Use visual tools. Images often carry information more efficiently than words.
- Give time for consolidation. Long-term memory builds over time. Cramming does not produce long-term storage.
A teacher who does these has students learning rather than receiving information temporarily.
Sensory register, working memory, long-term memory
Sensory register (short-term): brief reception of input from the senses.
Working memory: conscious memory where active thinking happens. Holds only a few unrelated items at once (Miller’s 7 ± 2; newer estimates ~3-5).
Long-term memory: permanent storage with effectively unlimited capacity.
Learning means moving information from working memory into long-term memory.