The Multi-Store Model and Peterson & Peterson
The Multi-Store Model
Where does a memory go after you first notice it? The multi-store model of memory, put forward by Atkinson and Shiffrin (1968), proposes that information travels through three distinct stores: the sensory register, short-term memory, and long-term memory. Each store differs in how it codes information, how much it can hold (its capacity), and how long it keeps hold of that information (its duration).
The multi-store model has three stores: the sensory register (large capacity, extremely brief duration), short-term memory (limited capacity of about 7 items, up to around 30 seconds without rehearsal), and long-term memory (potentially unlimited capacity, potentially permanent duration).
The Sensory Register
The sensory register receives information from all of your senses at once. It has a very large capacity — it can hold a huge amount of incoming sensory data — but its duration is extremely brief, lasting only fractions of a second. Almost all of this information is never processed further and simply fades away. Only information that is attended to moves on to short-term memory.
Coding in the sensory register matches the sense that detected it: visual information is coded visually (iconic memory) and auditory information is coded acoustically (echoic memory). The brief afterimage you see when a camera flash goes off is your iconic sensory register at work — a flicker of visual information that is gone almost instantly.
Short-Term Memory
Short-term memory (STM) holds information currently in use — for example, a phone number you have just read, or the words in the sentence you are reading right now. STM has a limited capacity of around 7 items, plus or minus 2. Its duration is also limited: without rehearsal, information in STM lasts only around 15 to 30 seconds before it fades. Coding in STM is mainly acoustic — we tend to hold information as sounds in our head, even when reading silently.
Peterson and Peterson (1959) is the named study for STM duration. Aim — to investigate how long STM lasts when participants cannot rehearse. Procedure — participants were shown a trigram of three consonants (for example, CHJ) and immediately given a three-digit number. They had to count backwards from this number in threes, which prevented rehearsal of the trigram. Recall was tested after retention intervals ranging from a few seconds up to eighteen seconds. Results — recall accuracy fell sharply as the interval increased, from a high level of accuracy at the shortest interval to only around 10% correct at the longest interval. Conclusion — without rehearsal, information decays from STM very quickly, giving STM a duration of roughly 18–30 seconds.
Rehearsal is the key mechanism for moving information from STM into long-term memory. By repeating or thinking about information, you keep it active in STM and eventually strengthen it enough for long-term storage. Without rehearsal, the trace in STM decays and is lost — exactly what Peterson and Peterson's counting-backwards task was designed to force.
Long-Term Memory
Long-term memory (LTM) is the relatively permanent store for information that has passed through STM. Its capacity is thought to be essentially unlimited, and its duration can last a lifetime. Coding in LTM is mainly semantic — information is stored in terms of its meaning, rather than its sound or appearance.
STM codes mainly acoustically, holds about 7 items (plus or minus 2), and lasts up to around 30 seconds without rehearsal. LTM codes mainly semantically, has potentially unlimited capacity, and can last a lifetime.
A strength of the multi-store model is that it is supported by controlled research. Peterson and Peterson's carefully controlled laboratory procedure — using a standardised trigram and a fixed counting-backwards task — produced reliable results showing STM decays rapidly without rehearsal, exactly as the model predicts.
A weakness is that the model may be oversimplified. It treats STM and LTM as single, unitary stores, but later research suggests both are more complex than this. There is also a question over the validity of Peterson and Peterson's study: recalling meaningless consonant trigrams is quite different from remembering meaningful information in everyday life, which limits how far the findings can be generalised beyond the artificial laboratory task.
How Information Moves Between Stores
Understanding the flow of information through the multi-store model matters as much as knowing the features of each store. Information first reaches the sensory register automatically from all active senses. At any moment, your sensory register holds a vast amount of information — every sound in the room, every colour in view — and nearly all of it is never consciously noticed, fading within fractions of a second.
The gateway between the sensory register and STM is attention. When you selectively attend to a piece of incoming sensory information — for example, hearing your name mentioned in a noisy room — it transfers into STM and becomes available for conscious processing. Everything else in the sensory register is lost without ever being consciously registered.
Once information is in STM, rehearsal determines whether it moves on into LTM. Simply repeating information keeps it circulating in STM; with enough repetition, the trace strengthens enough to be stored in LTM. This explains why Peterson and Peterson's counting task worked so well as a way of blocking rehearsal — without it, participants' recall would not have dropped so sharply as the retention interval grew longer.