videos we are going to look at two models of  memory, explore different types of long term  

memory, find out the reasons we forget, and  look at problems with eyewitness testimony  

and how police can improve eyewitness testimony.  Many of the topics in the memory unit come from an  

area of psychology called the cognitive approach.  You may want to watch my video on the cognitive  

approach either before or after these videos. In this first video we will cover the  

multi-store model, the stores it is made  from, features of each of those stores  

and how information flows through it. And  of course some evaluations of the model.

Features of the Multi-store Model

Created by researchers Atkinson and Shiffrin  in 1968 this is a multi store model, its an  

information processing model. Its liner, meaning  information is moved through in one direction  

and the stores are passive, meaning  they just hold on to information. 

The stores are, The sensory register, short term  memory and long term memory. I will describe them  

in detail in a moment, but before I start,  as we go through them keep in mind we need  

to remember three features of each of these  stores, their coding, capacity and duration. 

Coding is the format the information  is stored in. In a similar way that  

your phone's memory will store image  files and music files differently,  

the brain is thought to have different ways  of coding information. Capacity means how  

much can that store hold, and duration is  how long the information can be held for. 

Ok now I’ve made those terms clear, We can  describe the model. starting with sensory  

information. If you consider the five main senses  the information they collect needs to be first  

detected by sensory neurons and received by the  brain. The store that directly receives sensory  

information is called the sensory register. Now as the five sensors are coming from  

very different sensory organs we would say the  coding is different for each sense impression.  

We could give the names for each form of coding,  Iconic, Auditory, Haptic, Gustatory, olfactory,  

but in the exam we can just say the coding  of the sensory register is modality specific,  

it depends on the sense organ. The capacity of the sensory register  

is very large, possibly unlimited. It's hard to  put an exact number on how big it is, but think  

of it this way, your brain needs to detect all of  the sense information you receive in each moment,  

all of that visual information coming into your  eyes and all the sound you hear. Even all the  

touch sensations from all across your body. The  taste in your mouth and smells are detected by the  

sensory register. I'm not saying you are always  consciously aware of those things, but the brain  

is receiving those sense impressions constantly. The sensory register has a very short duration,  

again this varies between the different stores but  we can use the figure of 250 Milliseconds, so a  

quarter of a second, so the sensory register has  a very short duration before information is lost. 

Now of course as I said you're not consciously  aware of that massive amount of information coming  

in. For that we need to pay attention, maybe  to a particular sound or object in our vision. 

So information moves from the sensory  register to the next store in short term  

memory through attention. Any information  that isn't paid attention to is lost. 

We say short term memory coding is acoustic,  so information in the form of sound. And it  

feels like that, if you need to remember a  list you may repeat it over and over again  

in your head with an inner voice. The duration of short term memory is  

between 18 and 30 seconds, meaning information  that isn't maintained by subvocal repetition  

is lost from short term memory after that  much time has passed. The capacity of short  

term memory is 7 items plus or minus two, so  some people have a slightly worse short term  

memory, some people a slightly better one. Information can be passed to the final store  

by rehearsal, now this can be either maintenance  rehearsal which is keeping it in short term memory  

by repeating it again and again in a rehearsal  loop until it passes into long term memory,  

or information can be passed to long term  memory in a process called elaborative rehearsal  

and this means linking the new information to  knowledge we already have in long term memory. 

Information from short term memory that isn't  passed to long term memory, is of course lost.  

This could be due to displacement, because of  the small capacity new information comes in and  

replaces what used to be in short term memory. Or  it could be because of decay, if the information  

is not rehearsed it simply disappears over time. The coding in long term memory then is  

semantic coding, the memory is stored as part of  a set of meaningful connections to other stored  

information. so as an example we have a semantic  connection between the word wave and the idea of a  

wave of water, or waving your hand and recently  you may have developed a semantic connection  

between the word wave and an increase in illness. Now I’m going to say both the capacity and  

duration of long term memory is very large,  potentially unlimited. That is because we  

haven't found a limit. The oldest people can  still recall facts from their childhood, and we  

don't see a point in which brains get completely  full up with information and can't fit more in. 

But we do forget things, so we could argue that  information is lost. But as we will see in a  

future video, maybe the information isn't truly  lost, but maybe we just lose the ability to access  

it and could recall it if we had the right cue. So that's the multi-store model,  

we need to describe the model, the process  of information passing through the model  

but also the coding capacity and duration  of each store. Here they all are in a table,  

you may want to copy both the model and this table  down and make sure you can recall it from memory. 

In the evaluations I want to  cover the research that revealed  

those features of the multi-store model of  memory, and how we could evaluate that research

Hey Sofia, when I was doing my research for the  video there was a lot of conflicting research.  

Much of it to do with a lack of gender bias in  samples. Too much to get into in such a short  

vid and at the level of A-level. I don't think you  are wrong to take either argument as there will be  

research to support your position on this one! Research evidence for the multi-store model. 

To start with, just a simple piece of research  by Glanzer (1966) showing the primacy-recency  

effect. when participants were asked to recall a  list of words participants tended to remember the  

first words in the list and the last words  in the list, but struggled to remember the  

middle words. This suggests long term and  short term memory are separate processes  

with the first words in the list having  moved to long term memory, and the last  

words in the list displacing the middle words,  and taking their place in short term memory. 

The first feature of the stores that I'm going to  evaluate is the capacity of the sensory register,  

this was a study by Sperling. Spearling flashed a grid of 20 letters  

onto a screen for a 20th of a second, then  got the participants to immediately recall  

one row of letters, because it would take  too long to say the middle row or the top  

row he used different tones for each  row (pause). Let's have a go. (pause)  

So what were the letters? (pause) I imagine that was a little too fast,  

but spearling found that with some practice  the participants were very accurate.  

Now I'm going to argue that short term memory  only has a capacity of between 5 and 9 items.  

To be able to accurately recall a full row  sensory memory would have to hold all 20 letters  

in that moment. This suggests sensory memory has  a much larger capacity than short term memory,  

and we are only considering here  visual or iconic sensory memory. 

Moving on from sensory memory, this is  a study that demonstrates the coding  

of both long term and short term memory. It’s  a little complicated but stay with me because  

we will be able to evaluate the coding in both  long term and short term memory with this one.  

Baddeley gave participants word lists  to recall, these world lists were either  

acoustically similar, acoustically dissimilar,  semantically similar or semantically dissimilar. 

Baddely had his participants either recall  immediately testing short term memory, or  

after 20 minutes testing long term memory. Now the  results showed the worst recall in the immediate  

recall condition was for the acoustically similar  words. Baddely argues that this shows short term  

memory is coded acoustically, the acoustic  similarity of the words causing confusion. 

Now students often struggle with the  conclusion to this study, it’s a little  

counterintuitive to say because short term  memory is acoustically coded words that are  

acoustically similar are harder to recall. Well to help you try to remember, I will give  

you a metaphor of a filing cabinet. If it's  coded, or organised alphabetically then if we  

have a bunch of files that are stored across the  filing system, when we want to retrieve one it's  

easy. But if all the files are similar it can be  more difficult to go in to get the file we want. 

Now let me be clear, I'm not saying we have  individual files to store our memories.  

It's just to help you remember how  if a storage system is organised, or  

coded in a certain way, if all the items are  similar it could be more difficult to recall  

those individual items. List A: man, cab,  

can, cad, cap, mad, max, mat, cat, map List B: pit, few, cow, pen, sup, bar,  

day, hot, rig, bun List C: great, large, big,  

huge, broad, long, tall, fat, wide, high List D: good, huge, hot, safe, thin,  

deep, strong, foul, old, late

Moving on to the capacity of short term  memory we have very old research by Jacobs  

that showed when asked to recall lists  of numbers people could recall around  

9, or when tested with letters 7 was the  average. Now there was a bit of a range in  

what could be recalled, some people are  a little better, some a little worse.  

So what we can say is the capacity of short  term memory is seven items plus or minus two. 

This research really shows that the capacity  of short term memory is small. However  

Miller points out that we can increase the  capacity of short term memory by chunking,  

this is making small sets or groups of items.  Effectively reducing the number of items. 

Finally for short term memory we have duration.  To test this Peterson and Peterson asked  

participants to recall trigrams, these are sets  of three letters that don't have meaning, so  

HFD, or TKU. But Peterson and Peterson used an  interference task to stop the participants using  

maintenance rehearsal, repeating the trigram to  keep it in their head, this was by asking the  

participants to count backwards. It was found in  this study that the duration of short term memory  

was between eighteen and thirty seconds. Moving on to long term memory, we can  

use the same Baddeley study again, and we  find that recall after twenty minutes, so  

when testing long term memory is worse for the  list of semantically similar words. The semantic  

similarity is causing confusion, suggesting  that long term memory is coded semantically. 

Now capacity and duration is  potentially unlimited, and there is no  

way to actually show that, but we can look  at two studies that show the capacity and  

duration of long term memory is at least very  large. Firsty a researcher called wagnaar  

created a diary over the course of six years,  over that time making more than 2400 entries.  

When tested on events there was a 75%  recall after 1 year and 45% after 5 years  

since making the entry. This suggests the  capacity of long term memory is very large. 

Bahrick showed old photographs of school  friends to participants between the ages of 17  

and seventy four. After fifteen years recall  was 90%, and even after 48 years recall was  

80%. This does suggest that the duration of long  term memory is very large, potentially unlimited. 

Additional evaluations of the multi-store model. Before we finish, we could be expected to evaluate  

research on the multi-store model of memory and we  want to be able to give overall evaluations of the  

multi-store model, and we have a few options. One of the most obvious evaluations  

is much of the research I have outlined seems  like a very artificial way of testing memory,  

it's not really like how we use memory in our  day to day life. There are a few aspects to this,  

the lab environment is unusual, not reflecting  the places we recall, like school and work, if we  

can't apply the findings to those environments we  would say the findings lack ecological validity. 

The tasks the participants are asked to complete  are often not like the tasks people do in day to  

day life, so we can say the tasks lack mundane  realism. What I've done here is question the  

external validity of this research, suggesting  we may not be able to apply what we have found  

about memory under experimental conditions  in situations outside of the experiment. 

But in the defence of psychologists experiments  have to be designed to make measurements possible,  

it may be the only way of scientifically  testing the limits of memory. 

This is because we are not able  to directly observe memory,  

psychologists have to make inferences, educated  guesses about the underlying model based on the  

behaviour observed in an experiment. This is a  problem because those inferences could be wrong,  

it could be a completely different process  that produces the observed behaviour. 

An interesting point about the large capacity  and short duration of the sensory register,  

its supported by evolutionary theory, in the wild  we would have needed quick reactions for survival,  

taking in as much information as possible but only  keeping and processing the important information,  

holding on to and then processing too much  information would slow down reaction speed. 

A criticism of the multi-store model is it's a  little simplistic. It suggests short term memory  

and long term memory are unitary and passive  stores, meaning they are just one thing and  

they are just holding on to information. But  from research we will see in future videos  

there are clearly multiple types of long term  memory and short term memory is not unitary,  

it has a number of components that are  active, they work on information together  

passing and processing information between them. And even some basic assumptions of the multi-store  

model lack face validity, we clearly have  long term memories that are tastes and smells,  

and the capacity of short term memory isn't fixed,  it changes over a lifetime and with practice.

have a go at this a real exam question on the  multi store model. if you are a psych boost  

patron at the neuron level and above you  can access a tutorial on psych boost .com.  

in it I will talk you through a model answer for  this question and general tips.for everyone else  

don't forget to subscribe so you don't miss  the videos released right up to your exams,  

and I will see you in the next psych  boost video, types of long term memory.