The multi-store memory as proposed by Atkinson and Shiffrin is composed of 3 main stores, all varying in capacity, duration, encoding and how they can be forgotten. These components are known as the sensory memory, short term memory and Long term memory, information flows in a linear direction.
The sensory memory begins with stimuli coming into the memory (this can be visual, auditory or tactile). It has a duration of about a fraction of a second before is spontaneously decays due to loss of physical trace of the stimuli. Evidence that support the sensory memory was composed by Sperling (1960) where participants were asked to recall a 3 rows of letters immediately after the 50 millisecond presentation, the data showed that participants only recalled 4 out of 12 (30%) of the whole array however under a different condition when participants were only asked to recall a single row, they recalled 3 out of 4 (75%).
After the sensory memory has been given attention, it is then passed into the STM, where it also has a limited capacity and duration. Miller suggested that the capacity of the STM is based on chunking, the more meaningful the chucks the more likely it will be better recalled. He concluded that the capacity to chunking of the STM Is around 7+-2. It is usually forgotten due to a lack of rehearsal or displacement where memories are pushed out before entering LTM to make room for new information.
After being rehearsed the information from STM will then be passed into LTM were it has unlimited capacity and a duration of a lifetime. In the case of being forgotten, it is usually by interference where LTM Is confused with memory.
An experiment that distinguishes STM from LTM is research composed by Glazer and Cunitz, with their primacy and recency effect. After asking participants to recall a list of words with immediate recall, they found that participants recalled the first few words (primacy effect) due to the words being stored into the LTM after rehearsal, and the last few words (recency effect) due to it