How limited is working memory capacity, and how can we optimise its use?
Working memory capacity is crucially important for thinking and learning, and it varies between people (although it is relatively fixed within each of us). Those with greater capacity are able to process more complicated ideas, while those with less capacity may find this more challenging.
Our capacity for working memory increases with age over the course of childhood—an adult’s capacity is more than twice that of a four year-old. Within an average primary school class of 7-8 year-old children, around 10% are likely to have the same capacity of a 4 year-old, while 10% are likely to have similar to that of an average 11 year-old.
So, what is considered to be ‘normal’ working memory capacity? You may have heard about the magic number 7 (which comes from George Miller’s classic paper of the same name, published 1956), and indeed many of us can remember around 7 random digits, letters or words 50% of the time, but not much more. The capacity was famously thought to be 7±2 (i.e. normal range is between 5 and 9 ‘items’). However, later studies showed that it is actually more like 3-4 items when material is unfamiliar.
But perhaps the more interesting questions, then, becomes:
- What is an ‘item’?; and
- Under what circumstances can we group a few items together into one ‘chunk’?
ITEM: is a small piece of information.
CHUNKING: is where a number of items can be combined; in so doing, the combined items seem to behave as a single item (in terms of how they’re represented in working memory). Chunking can occur when we have information in our prior knowledge that allows us to group a few items together. We usually do this based on meaning (for instance words associated with feelings, like ‘love’, ‘hate’ and ‘fear’) or frequent repetition of these things together, for example, the repetition of individual numbers in a telephone dialling code (like 0191 for Durham, UK), a series of actions involved in making a hot drink (fill the kettle, put the kettle on, find a tea bag, a cup…). Crucially, chunking based on meaning occurs when we have previously learned something that allows us to group the items together.
Consider the following alphanumeric sequence:
M I 6 0 0 7 K G B C I A
Could you quickly memorise it if you needed to? Maybe, but it would probably be a lot easier if you broke the string up into more meaningful “chunks”:
MI6 007 KGB CIA
What can we do to increase the capacity of our working memory?
‘Can we increase our working memory capacity through working memory training? It would be great if there were a way to increase it, for instance by recalling lists of increasing lengths of unfamiliar items, over and over again. So do such mental exercises work?
The short answer is, sadly, no.
As far as we know, working memory capacity cannot be increased by training. There are many commercial companies that promise such improvement, but research has shown that the increase is specific to the trained task, and not a general one. So mere training doesn’t work.
This is not to say that there aren’t things we can do to make better use of the capacity we have. However, at present, we have no evidence that there are any ways of actually increasing working memory capacity itself. So, it is important to acknowledge two key features of working memory:
- It is content-limited: we can only process a relatively small number of individual items or chunks at the same time; and
- It is time-limited: we can only hold information in working memory for a very short amount of time before it begins to decay (around 2-3 seconds, in some circumstances). Once the information decays, it’s gone until we are given the information again.
Gain a robust understanding of the basic cognitive faculties: attention, working memory, and long-term memory, and how these systems interact as a student progresses from novice to advanced stages with our Science of Learning Programme.
-
[…] manipuleren van informatie. Onderzoek toont aan dat deze werkruimte beperkt is, alhoewel de exacte limiet afhangt van hoe dit bestudeerd wordt. Omdat ruimte beperkt is, moeten onze hersenen een mechanisme hebben om te filteren welke […]
Leave a Comment
I was checking up and in fact there has been evidence that working memory can be improved.
https://www.researchgate.net/publication/257349535_Working_Memory_Training_May_Increase_Working_Memory_Capacity_but_Not_Fluid_Intelligence&sa=U&ved=2ahUKEwjnxuXxnuz3AhWYw4sBHeO6AGU4ChAWegQIARAC&usg=AOvVaw0EaK60BDxGLLRufl07zxI_
https://www.sciencedirect.com/science/article/pii/S1877042813008434/pdf?crasolve=1&r=70df1328795785b7&ts=1652986819926&rtype=https&vrr=UKN&redir=UKN&redir_fr=UKN&redir_arc=UKN&vhash=UKN&host=d3d3LnNjaWVuY2VkaXJlY3QuY29t&iv=db98daa3eef8c12d8e92f06755c3f585&token=31663466643766323537613332663330633831623732333337303039303564646164623664656539346635656532333535396462303765323838363236616433383062373661623931646134323830363563383436613236303435343a343131376335633039666532653336303063353634633938&text=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&original=3f6d64353d3737636261383432643232643333393465333634333836613866373132313233267069643d312d73322e302d53313837373034323831333030383433342d6d61696e2e706466
This article makes the following claim:
“There are many commercial companies that promise such improvement, but research has shown that the increase is specific to the trained task, and not a general one.”
The problem is that while people can practice and train for a specific task and see their performance improve, that doesn’t mean that their working memory capacity improves. It just means that they get better at that task. An actual increase in working memory capacity would mean an increase in performance in all tasks. Unfortunately, it’s unclear whether that’s possible.
Let’s look at the articles you shared. The first research paper you cited demonstrates that participants, after training, improved their performance in similar tasks, but the benefits didn’t transfer to substantially different tasks. They say this in the conclusion:
“For both the complex- and simple-span-training groups, there was some evidence for moderate transfer to the secondary memory component of free-recall and keep-track tasks. These findings imply improvement to WMC at the construct level, but the lack of transfer to
other tasks (e.g., the simple span tasks) leads us to be cautious about endorsing this conclusion.”
They then give two different interpretations of their results, with one being: “The second possible explanation of our data is that subjects developed strategies that were applicable to certain transfer tasks but not to others.” This is a common effect of WMC training — one gets better at the task, but doesn’t improve in general.
In the second study, the participants did the same test before and after training, which might suggest that the participants who trained just learned how to do better at that specific test. Again, it says in their conclusion: “The conclusions we can draw from this pilot study are limited by the fact that we did not include any measures of transfer.”
I really want WMC to be malleable and trainable as that would have extraordinary benefits to students and people in general. But sadly, I’ve yet to be convinced that it is possible!